Anthropic, PBC is an American artificial intelligence company that develops advanced AI models, software platforms, and applications designed to help individuals and organizations perform increasingly complex tasks. Founded in 2021 and headquartered in San Francisco, California, Anthropic is best known for Claude, its family of artificial intelligence models used for writing, software development, data analysis, research, business automation, and other professional activities.
Unlike traditional software companies that primarily sell applications with fixed features, Anthropic develops AI systems capable of performing a wide variety of tasks. Its customers can access these systems directly through Anthropic’s applications or integrate Claude into their own software and business operations.
Anthropic’s business strategy emphasizes three interconnected areas:
Frontier AI Research: Developing increasingly capable artificial intelligence models that can reason, write software, analyze information, and perform complex tasks.
AI Platforms and Applications: Commercializing these models through Claude, Claude Code, Claude Cowork, developer tools, and specialized enterprise solutions.
Enterprise AI Adoption: Helping businesses integrate artificial intelligence into everyday operations, software products, and long-running workflows.
Anthropic operates as a Public Benefit Corporation (PBC), a legal structure that requires its directors to balance shareholder financial interests with the interests of people affected by the company’s operations and its stated public benefit purpose. Anthropic identifies that purpose as the responsible development and maintenance of advanced AI for humanity’s long-term benefit.
On October 26, 2026, Anthropic filed a Form S-1 registration statement with the U.S. Securities and Exchange Commission (SEC) for its proposed initial public offering (IPO). The company has applied to list its Class A common stock on the Nasdaq Global Select Market under the proposed ticker symbol ANT.
As of the October 26, 2026 filing, the IPO price range and number of shares to be offered had not yet been determined.
Key Business Idea: Anthropic is building more than an AI chatbot. Its business combines advanced AI models, developer infrastructure, enterprise software, and autonomous AI applications within a unified technology platform.
1.1 What Does Anthropic Actually Do?
At the center of Anthropic’s business is Claude, an artificial intelligence system designed to understand information, generate content, reason through problems, write computer programs, and interact with software tools.
Claude belongs to a category known as large language models (LLMs). These are AI systems trained on large amounts of information to recognize patterns and generate useful responses to instructions.
Modern AI models can perform considerably more than simple text generation. Anthropic develops systems that can analyze documents, process different types of information, write and execute code, interact with applications, and complete tasks involving multiple steps.
Its technology supports several major categories of work:
Software Development: Writing, reviewing, testing, debugging, and maintaining computer programs.
Business Productivity: Preparing reports, analyzing documents, organizing information, and automating repetitive office tasks.
Financial Analysis: Processing financial information, supporting research, preparing documents, and assisting professional workflows.
Scientific Research: Analyzing scientific literature, supporting research processes, and working with specialized scientific tools.
Customer Service: Powering automated customer support systems and AI assistants.
Enterprise Applications: Providing AI capabilities that other businesses can incorporate into their own products and internal systems.
A central part of Anthropic’s strategy involves moving from AI systems that simply answer questions toward systems that can complete work on behalf of users.
These systems are commonly called AI agents. An AI agent can perform multiple connected actions toward a goal, rather than requiring a separate instruction for every individual step.
For example, a traditional chatbot might explain how to analyze a spreadsheet. An AI agent with appropriate tools and permissions may be able to open the spreadsheet, examine the data, perform calculations, and prepare a report.
Anthropic refers to advanced systems that collaborate with people on professional tasks as Virtual Collaborators. The company views this transition from answering questions to completing work as an important development in commercial AI.
Plain English: Anthropic sells access to artificial intelligence that helps people and businesses get work done. Its long-term business opportunity depends not only on making AI smarter, but also on making that intelligence useful in real workplaces.
1.2 The Claude AI Model Family
Anthropic develops multiple classes of Claude models rather than relying on a single AI system.
Different tasks require different combinations of intelligence, processing speed, and operating cost. A complex scientific problem may justify using an advanced model, while processing thousands of routine customer inquiries may be more economical with a faster and less expensive model.
According to Anthropic’s October 2026 S-1, the Claude family consists of four principal model classes.
Model Class
Primary Positioning
Typical Applications
Mythos / Fable
Highest-capability frontier models, with different safety and access configurations
Advanced coding, complex reasoning, cybersecurity, scientific work, and long-running AI agents
Opus
Advanced reasoning and professional AI work
Complex software development, research, knowledge work, and autonomous workflows
Sonnet
Balanced performance, speed, and cost
General business applications, coding, document analysis, and everyday AI tasks
Haiku
Fastest and lowest-cost model class
High-volume customer service, document processing, and routine automated tasks
Mythos and Fable: Anthropic describes Mythos as its most capable model class. Because certain advanced capabilities could create risks in sensitive areas such as cybersecurity and biology, the company offers different configurations. Fable is generally available with production safeguards, while Mythos access is restricted to vetted organizations and individuals through specialized programs.
Opus: The Opus family focuses on sophisticated reasoning, software development, computer use, and demanding professional tasks. Claude Opus 5.5, announced in September 2026, is designed for complex enterprise workloads and autonomous agents.
Sonnet: Sonnet provides a balance between advanced capabilities and operating costs. This makes it suitable for organizations that need capable AI across a broad range of applications without using the most computationally demanding model for every task.
Haiku: Haiku emphasizes low cost and fast responses. These characteristics matter for businesses processing large volumes of requests, such as automated customer support, document classification, and information extraction.
Anthropic can combine these model classes within a single application. A business may use a more capable model for difficult reasoning and a less expensive model for routine tasks.
This flexibility is important because AI computing has a real operating cost. The more computing resources required to process a task, the more expensive that task may be to deliver.
Plain English: Claude is not one product with one price and one performance level. Anthropic offers different models so customers can choose how much intelligence, speed, and computing expense they need for a particular job.
1.3 Claude Platform: Infrastructure for AI Applications
Anthropic’s business extends beyond the Claude applications used directly by consumers and employees.
The Claude Platform provides developers and businesses with the tools needed to build their own AI-powered products and automated workflows.
One of its most important features is an Application Programming Interface (API). An API is a software connection that allows one application to request services from another.
For example, a financial software company could use Anthropic’s API to incorporate Claude into its own product. Customers would interact with the financial software, while Claude performs AI-related tasks behind the scenes.
Anthropic’s platform includes technologies such as:
Claude API: Enables developers to access Claude models from their own software and applications.
Tool Use: Allows Claude to interact with external functions, software services, and connected systems.
Agent SDK: A software development toolkit for building AI agents that can perform multi-step tasks.
Claude Managed Agents: Provides infrastructure operated by Anthropic for running AI agents without requiring customers to manage every underlying component.
Model Context Protocol (MCP): An open standard that helps AI applications connect to external data, business tools, and software systems.
Agent Skills: Reusable instructions and resources that enable AI agents to perform specialized or repeatable tasks.
Prompt Caching: Reuses previously processed information to reduce the cost and processing time of repeated AI requests.
Batch Processing: Allows customers to process large volumes of requests asynchronously at a lower cost.
These features make the Claude Platform relevant to companies building everything from customer service tools to complex internal business systems.
Anthropic also provides enterprise-oriented security and administration features, including access controls, audit logs, and tools for managing how employees and applications use AI.
This matters because large organizations often require more than a technically capable model. They also need predictable access, security controls, integration with existing systems, and tools for monitoring usage.
Anthropic operates the underlying intelligence service rather than offering customers a downloadable version of its proprietary model family for independent hosting.
Plain English: The Claude Platform allows other businesses to build products using Anthropic’s AI technology. Anthropic supplies the intelligence and supporting infrastructure, while customers can develop their own applications around it.
1.4 Claude Code and Claude Cowork
Two important commercial applications demonstrate Anthropic’s strategy of turning AI capabilities into software that can perform practical work.
Claude Code focuses on software development, while Claude Cowork extends AI automation to a wider range of professional tasks.
Claude Code is an AI coding agent designed to assist developers throughout the software development process.
Its capabilities include:
Understanding and analyzing existing software projects.
Writing and modifying computer code.
Identifying and debugging software errors.
Running tests and reviewing results.
Working across large software repositories.
Coordinating multiple coding tasks and agent sessions.
A software repository is a collection of source code and related files used to develop and maintain an application.
Unlike a basic coding assistant that suggests a few lines of code, Claude Code can perform connected development tasks with access to appropriate tools and execution environments.
Claude Cowork applies similar agent-based technology to broader office and knowledge work.
Knowledge work refers to activities involving information, analysis, communication, research, and decision support rather than primarily physical labor.
Claude Cowork can interact with files, connected applications, and computer interfaces to complete tasks such as:
Preparing and editing documents.
Analyzing spreadsheets and organizing data.
Conducting research across multiple information sources.
Creating reports and presentations.
Managing files and performing administrative workflows.
Completing multi-step tasks across supported business applications.
Anthropic introduced Claude Cowork as a research preview in January 2026. The company began combining Claude Cowork and Claude.ai into a more unified application experience in September 2026.
The commercial significance is that AI products can become more deeply integrated into the activities businesses perform every day.
Instead of using AI only to generate individual answers, organizations can build workflows in which AI performs a series of connected tasks.
However, successful automation still depends on factors such as model reliability, permissions, data quality, and appropriate human oversight.
Plain English: Claude Code helps programmers build software, while Claude Cowork helps professionals complete broader computer-based tasks. Both products are examples of Anthropic’s effort to sell AI that performs work rather than merely answers questions.
1.5 How Anthropic Makes Money
Anthropic generates revenue primarily by charging customers for access to its AI intelligence service.
The company describes its main commercial approach as pay-for-intelligence, meaning customers pay according to how much AI processing they use.
There are two principal revenue models:
Consumption-Based Revenue: Customers pay for AI usage, generally measured in tokens and priced according to the Claude model selected.
Subscription-Based Revenue: Customers pay recurring fees for access to specified AI capabilities and usage allowances through subscription plans.
A token is a unit of text or other information processed by an AI model. For text-based interactions, a token may represent a word, part of a word, or a punctuation mark.
Under consumption-based pricing, the quantity of information processed and the model selected help determine what customers pay.
For example, a business that processes a large number of documents through Claude may generate more usage-based revenue for Anthropic than a business that uses the service only occasionally.
Subscription plans, such as Claude Pro, Max, Team, and Enterprise, provide another way for customers to access the technology.
Some enterprise customers also enter into annual or multi-year agreements that include minimum spending commitments. These arrangements can primarily generate consumption-based revenue even though they involve longer contractual terms.
Anthropic’s disclosed revenue mix demonstrates the importance of usage-based services.
Revenue Type
FY2025
First Half 2026
Consumption-Based
83%
80%
Subscription-Based
17%
20%
Consumption-based revenue accounted for approximately 80% of total revenue during the first six months of 2026, showing that the company’s financial performance is closely connected to how extensively customers use Claude.
This business model differs from traditional software subscriptions because customers can increase or decrease spending as their AI usage changes.
Higher adoption may increase revenue without requiring a new subscription for every additional task. However, usage-based revenue can also fluctuate when customers reduce activity, change models, or control spending.
Plain English: Anthropic earns money when customers use its AI models and applications. Most revenue comes from the amount of AI processing customers consume, rather than solely from fixed monthly subscriptions.
1.6 Enterprise Customers and Market Adoption
Enterprise customers are a central part of Anthropic’s commercial strategy.
An enterprise customer is a business or organization that uses software or services for professional operations, often across teams, departments, or large groups of employees.
Anthropic serves customers ranging from AI-focused startups to major multinational corporations, financial institutions, healthcare organizations, software companies, research institutions, and government-related organizations.
Its S-1 provides several indicators of enterprise adoption.
As of September 30, 2026:
More than 11,900 customers generated at least $100,000 in trailing twelve-month revenue.
More than 1,900 customers generated at least $1 million in trailing twelve-month revenue.
Approximately 250 customers generated at least $10 million in trailing twelve-month revenue.
Customers meeting the $100,000 trailing-revenue threshold included every company in the Fortune 10 and more than 50% of the Fortune 500.
Trailing twelve months (TTM) means the most recent twelve-month period ending on a specified date. This allows investors to examine customer spending over a full year rather than only a single quarter.
The number of customers generating at least $100,000 in TTM revenue increased from more than 950 as of September 30, 2025, to more than 11,900 one year later.
These figures demonstrate a substantial expansion in customers generating meaningful amounts of revenue for the company.
However, they do not mean every organization uses Claude across its entire workforce. A large company may begin with one department or application before expanding into other areas.
Anthropic’s strategy is designed around this pattern of customer expansion.
A developer might initially use Claude Code for software development. Over time, the same organization could adopt Claude for customer support, research, internal reporting, and additional business applications.
As Claude becomes integrated into more workflows, customers may generate additional consumption-based revenue or move to higher subscription tiers.
Plain English: Anthropic’s customer base includes many large businesses, but the important question is how extensively those businesses use Claude. A customer adopting AI across more departments can become more valuable even without Anthropic acquiring an entirely new customer.
1.7 Cloud Partnerships and Distribution
Anthropic distributes Claude through its own applications and platform as well as major cloud computing providers.
Cloud computing allows organizations to use remote computing infrastructure and software services instead of purchasing and maintaining all the necessary equipment themselves.
Anthropic’s disclosed cloud distribution relationships include:
Amazon Web Services (AWS): Makes Claude available through Amazon’s cloud infrastructure and AI services.
Google Cloud: Provides access to Claude through Google’s cloud-based enterprise AI offerings.
Microsoft Azure: Makes Claude accessible to customers within Microsoft’s cloud ecosystem.
These relationships are commercially important because many large organizations already operate their applications and data systems on one or more of these cloud platforms.
Accessing Claude through an existing cloud environment can make adoption easier for businesses that have established security, procurement, and infrastructure requirements.
Anthropic also works with systems integrators and business technology partners.
A systems integrator is a company that helps organizations combine different software products, databases, and technology systems into a working solution.
Anthropic identifies relationships involving firms such as Accenture, Deloitte, KPMG, and PwC, along with enterprise technology platforms including Salesforce and Snowflake.
These distribution relationships can expand Anthropic’s commercial reach without requiring the company to build every customer-facing integration or sales channel independently.
Anthropic states that its arrangements with major cloud partners generally allow it to establish direct contractual relationships with customers while paying fees to the cloud providers.
However, these partnerships create dependencies as well. Amazon, Google, and Microsoft are important participants in the broader AI industry, and their strategic interests may not always align with Anthropic’s.
Plain English: Anthropic does not need every business to visit its website to become a customer. Companies can access Claude through cloud services and enterprise software they already use, which helps Anthropic reach a wider market.
1.8 Computing Infrastructure: The Foundation of Anthropic’s Business
Developing and operating advanced artificial intelligence requires enormous computing resources.
AI models must first be trained using specialized computing systems. After training, they continue consuming computing resources whenever customers send requests and receive responses.
These activities are commonly divided into two categories:
Training: The process of developing or improving an AI model using data and computational methods.
Inference: The process of using a trained model to generate responses, analyze information, or perform tasks for customers.
Both activities require powerful processors, data-center equipment, electricity, networking infrastructure, and specialized software.
Anthropic uses several categories of AI computing chips:
NVIDIA GPUs: Graphics Processing Units widely used for AI training and inference.
AMD GPUs: Alternative high-performance processors used for AI computing workloads.
Google TPUs: Tensor Processing Units, specialized chips developed by Google for machine learning.
AWS Trainium: Amazon’s specialized processors designed for AI model training workloads.
Rather than relying on a single chip architecture, Anthropic has developed systems designed to manage computing capacity across multiple hardware platforms and providers.
The company describes this as a unified, flexible computing pool. Its internal systems can direct available computing resources toward model development, customer requests, or internal operations according to changing requirements.
This strategy is intended to improve computing efficiency, manage capacity, and reduce dependence on any single type of processor.
However, securing sufficient computing infrastructure requires substantial financial commitments, often before the equipment or capacity becomes available.
Anthropic has entered into significant arrangements with technology and infrastructure partners to support its AI development and commercial operations.
These arrangements are an important feature of its business model because access to computing resources influences both the company’s ability to develop competitive models and the cost of delivering AI services to customers.
For investors, Anthropic’s economics depend partly on whether it can continue improving the amount of useful AI work produced for each unit of computing expenditure.
Plain English: Anthropic needs powerful computer chips and data centers to develop Claude and serve customers. Better computing efficiency can help the company provide more AI services without increasing computing expenses at the same rate.
1.9 Competition in the Artificial Intelligence Industry
Anthropic operates in a rapidly evolving and highly competitive technology industry.
Its direct competitors include companies developing advanced general-purpose AI models, particularly OpenAI and Google, as well as other major technology companies and specialized AI developers.
The company also competes with businesses developing AI platforms, coding assistants, enterprise automation software, and applications designed for particular industries.
Competition occurs across several dimensions:
Model Capability: The ability to solve difficult problems accurately and perform sophisticated tasks.
Reliability and Safety: The ability to produce dependable results while controlling potentially harmful or unauthorized behavior.
Operating Cost: The computing expense required to deliver AI services at scale.
Processing Speed: How quickly models respond and complete tasks.
Enterprise Integration: How easily AI systems connect with existing business software and data.
Distribution: Access to customers through applications, cloud providers, and business partnerships.
Research and Talent: The ability to attract researchers and engineers and translate new discoveries into commercial products.
Anthropic believes its research capabilities, enterprise focus, safety-related work, and efficient use of computing resources help differentiate its products.
These are important elements of the company’s competitive positioning, although its S-1 also acknowledges substantial competitive pressure from well-funded rivals.
Another challenge is that AI technology advances quickly. A model that performs well today may become less competitive when another company introduces a more capable or less expensive alternative.
Customers may also use models from several providers rather than relying exclusively on Anthropic.
Plain English: Anthropic competes by trying to offer AI that is capable, reliable, useful, and economical. Its long-term competitive position depends on continuing to improve Claude while other companies develop competing technologies.
1.10 AI Safety and Research as a Business Strategy
Anthropic places particular emphasis on artificial intelligence safety and the responsible development of increasingly capable systems.
Its research includes areas such as:
AI Alignment: Developing AI systems that behave consistently with intended instructions, rules, and human objectives.
Interpretability: Studying how AI models process information and arrive at their behavior or outputs.
Model Evaluations: Testing AI systems to identify capabilities, weaknesses, and potentially dangerous behaviors.
Safeguards: Implementing technical and operational controls designed to reduce harmful or unauthorized uses.
Adversarial Testing: Deliberately challenging AI systems to identify weaknesses before or after deployment.
Anthropic argues that safety research is closely connected to commercial reliability.
For example, financial institutions, healthcare organizations, and other regulated businesses may require AI systems that provide predictable behavior, protect sensitive data, and operate within defined permissions.
As AI agents gain the ability to access files, execute computer programs, and interact with enterprise systems, reliability becomes increasingly important.
Anthropic has developed policies and evaluation methods intended to manage these risks, including its Responsible Scaling Policy.
The company also researches methods for making AI behavior more understandable and improving the safeguards applied to advanced models.
Nevertheless, Anthropic’s safety-focused approach does not eliminate the possibility of errors, misuse, unexpected behavior, or security incidents.
From a business perspective, Anthropic views its investment in safety as both part of its corporate mission and a way to meet enterprise customers’ requirements.
Plain English: Anthropic invests in making AI systems safer and more dependable because businesses need to trust software that handles important information and performs real tasks. However, safety measures do not guarantee that an AI system will always behave correctly.
1.11 Public Benefit Corporation and IPO Governance
Anthropic’s proposed IPO introduces a corporate governance structure that differs from many traditional publicly traded technology companies.
The company is incorporated in Delaware as a Public Benefit Corporation.
Under this structure, the board of directors must balance financial interests with the interests of affected stakeholders and the company’s stated public benefit purpose.
Anthropic has also established a Long-Term Benefit Trust (LTBT), an independent trust designed to support the company’s long-term public benefit objectives.
According to the S-1, the LTBT generally has the authority, through Class T common stock, to elect and remove directors who together constitute a majority of the authorized board.
Anthropic’s proposed capital structure includes several classes of common stock with different rights:
Class A: The common stock proposed for public trading, generally carrying one vote per share.
Class N: Generally non-voting common stock, primarily held by strategic business partners, subject to specified conversion provisions.
Class F: A special class intended to give the co-founders’ holding entity 50.1% of the voting power on matters submitted to stockholders generally, with no economic rights attached to the Class F share itself.
Class T: Shares associated with the Long-Term Benefit Trust’s authority over designated directors, without ordinary economic rights.
Class E: A special employee share class with limited voting rights and no economic rights.
This arrangement separates important aspects of economic ownership, ordinary shareholder voting power, and board representation.
For public investors, the distinction matters because purchasing Class A shares does not necessarily provide voting influence proportionate to the economic investment.
The S-1 identifies Anthropic as a controlled company under Nasdaq’s corporate governance framework. This classification can allow certain governance exemptions, although Anthropic’s specific arrangements and disclosures determine their application.
Anthropic’s public benefit commitments also include an allocation of shares to a dedicated public benefit fund and a framework for potential additional issuances tied to specified share-price milestones.
These arrangements are relevant because future share issuances can affect the ownership percentage of existing shareholders.
Plain English: Buying Anthropic stock would provide an economic ownership interest in the company, but public shareholders would have limited control over certain major decisions. Its founders and Long-Term Benefit Trust are designed to retain significant governance authority after the IPO.
1.12 Growth Strategy and Long-Term Opportunities
Anthropic’s growth strategy focuses on improving its AI technology, expanding enterprise adoption, and bringing advanced AI capabilities into additional types of professional work.
Its S-1 identifies several strategic priorities:
Develop More Capable AI Models: Continue investing in research to improve Claude’s reasoning, coding, autonomous work, and reliability.
Improve Computing Efficiency: Increase the amount of useful AI work delivered through available computing infrastructure.
Expand Enterprise Adoption: Help existing customers use Claude across more departments, workflows, and applications.
Strengthen Distribution: Expand direct sales, cloud partnerships, and enterprise technology integrations.
Develop Specialized Applications: Build AI solutions for industries such as finance, healthcare, life sciences, cybersecurity, and legal services.
Enter New Markets: Explore applications in scientific research, robotics, advanced manufacturing, and other emerging fields.
The company is already developing industry-specific offerings designed to help professionals use Claude within their established tools and regulatory environments.
Examples include Claude solutions for financial services, healthcare, life sciences, legal work, and education.
These applications are intended to make the same underlying AI technology useful across multiple industries.
Anthropic also emphasizes what it describes as two reinforcing business dynamics.
The first is an improvement loop: customers use Claude in real situations, their experiences help inform product development, and improvements can encourage broader adoption.
The second is an investment loop: customer adoption generates revenue that can help finance further AI research, computing infrastructure, and product development.
These dynamics are central to management’s long-term business strategy. However, their effectiveness depends on Anthropic continuing to attract customers, improve its products, and manage the substantial costs of frontier AI development.
New applications in scientific discovery, healthcare, robotics, and other industries represent potential opportunities, but their long-term commercial contribution remains uncertain.
Plain English: Anthropic wants to improve Claude, make it useful for more types of work, and expand how much customers use it. The company’s growth strategy depends on converting advances in AI technology into products and services that customers are willing to pay for.
1.13 Plain English Summary: What Kind of Company Is Anthropic?
Anthropic is best understood as a company that develops and commercializes advanced artificial intelligence.
Its business combines several connected activities:
Developing proprietary Claude AI models.
Selling AI processing through consumption-based pricing.
Offering subscriptions to AI applications.
Providing a platform for developers and businesses to build AI-powered products.
Developing AI agents that can perform complex professional tasks.
Distributing Claude through major cloud and enterprise technology partners.
Investing heavily in computing infrastructure, research, and safety systems.
The company’s commercial strategy is especially focused on enterprises, where AI can be integrated into workflows that operate continuously and support important business functions.
Anthropic has reported rapid growth in customers generating substantial revenue, demonstrating increasing commercial adoption of Claude. At the same time, its business requires continuous technological development and significant computing resources.
For beginner investors, Anthropic should not be viewed simply as a consumer chatbot company. It is building an AI technology platform that serves both individual users and organizations developing their own AI-powered products and operations.
Its competitive position depends on the capabilities and reliability of Claude, the usefulness of its applications, its relationships with customers and technology partners, and its ability to supply AI processing efficiently.
Its proposed IPO also includes an unusual governance structure that preserves significant influence for the founders and the Long-Term Benefit Trust.
Beginner Takeaway: Anthropic sells advanced AI capabilities through models, applications, subscriptions, and usage-based services. Its business opportunity is tied to the growing use of AI in professional work, while its long-term success depends on technological competitiveness, customer adoption, computing efficiency, and reliable execution.
Anthropic experienced extraordinary revenue growth between FY2023 and FY2025. Revenue increased from $41 million in FY2023 to $386 million in FY2024 and $4,585 million in FY2025. The expansion reflects the rapid commercialization of Claude and increasing adoption of Anthropic’s AI services by businesses, developers, and individual users.
Revenue increased by approximately 841.5% in FY2024 and 1,087.8% in FY2025. This means annual revenue grew nearly twelvefold between FY2024 and FY2025 alone. Such growth demonstrates how quickly commercial demand for AI services expanded during the period, although the company’s relatively small initial revenue base also contributed to the exceptionally high percentage increases.
One of the most important developments was the improvement in gross profitability. Gross profit represents the revenue remaining after subtracting the direct costs associated with providing services to customers.
Anthropic reported a gross profit of $15 million in FY2023, followed by a gross loss of $366 million in FY2024. In FY2025, gross profit recovered to $1,812 million, representing a gross margin of approximately 39.5%.
This improvement indicates that the company generated substantially more revenue relative to the direct costs of serving customers. However, gross profit does not include all expenses required to develop and operate the business.
Anthropic’s operating expenses remained exceptionally high because developing frontier AI models requires substantial investment in computing infrastructure, research, engineering, and other business activities.
According to the S-1, total compute and infrastructure expenses increased from $433 million in FY2023 to $2,522 million in FY2024 and $7,326 million in FY2025. These amounts include computing resources used for both customer services and model development.
Research and development expenses, disclosed separately in the supplementary functional expense information, also increased significantly, reaching $6,338 million in FY2025, compared with $2,275 million in FY2024 and $422 million in FY2023.
Research and development includes the costs of improving AI models, developing new capabilities, conducting research, and maintaining the technical resources required for model development.
In addition, FY2025 operating expenses included a $1,500 million charge related to the Bartz copyright litigation settlement. This legal settlement increased the company’s reported operating loss during the year.
Consequently, Anthropic’s operating loss widened from $551 million in FY2023 to $2,980 million in FY2024 and $8,060 million in FY2025.
The distinction between gross profit and operating profit is especially important. Anthropic’s direct customer-serving economics improved substantially in FY2025, but spending on research, infrastructure, personnel, legal matters, and commercial expansion continued to exceed the gross profit generated by customer services.
Another unusual feature of Anthropic’s financial statements is the enormous difference between its operating loss and net loss.
In FY2025, Anthropic reported a net loss of $41,965 million, substantially larger than its $8,060 million operating loss.
The primary reason was a $34,308 million loss from changes in the fair value of convertible notes, derivative liabilities, and related financial instruments.
Fair value refers to the estimated current value of a financial asset or obligation. Certain convertible securities are measured at fair value, and changes in their estimated value are recognized in the income statement.
In Anthropic’s case, increases in the company’s estimated enterprise value contributed to increases in the fair value of outstanding convertible financial obligations. These increases were recognized as accounting losses.
Importantly, these fair-value losses are not equivalent to cash spent operating Claude or developing AI models. They primarily reflect changes in the accounting value of financial obligations rather than ordinary business expenses.
This does not mean the losses should be ignored. Convertible financial instruments can affect the company’s future capital structure, shareholder ownership, and dilution. However, investors should distinguish these accounting effects from the operating performance of Anthropic’s AI business.
Net loss per share also deteriorated, reaching $(23.0) in FY2025, compared with $(5.4) in FY2024 and $(1.4) in FY2023. These figures reflect the company’s reported losses allocated across its weighted-average common shares and should be interpreted alongside its unusual pre-IPO capital structure.
For beginner investors, the central takeaway is that Anthropic’s revenue expanded dramatically, and its gross profitability improved, but the company continued spending heavily on model development and infrastructure. The exceptionally large reported net losses were also affected by substantial noncash fair-value adjustments related to convertible securities.
Therefore, evaluating Anthropic’s historical financial performance requires separating three different developments: revenue growth, operating profitability, and accounting losses associated with financing instruments. Each describes a different aspect of the company’s financial condition.
2) Key Financial Ratios
Unit: %, except Net Debt / EBITDA and Interest Coverage Ratio, which are expressed as multiples (x). Percentages are rounded to one decimal place.
Ratio
FY 2024
FY 2025
ROE (%)
84.8%
83.1%
ROA (%)
(107.2)%
(170.2)%
ROTC (%)
(69.1)%
91.3%
ROIC (%)
(253.9)%
30.8%
Gross Margin (%)
(94.8)%
39.5%
Operating Margin (%)
(772.0)%
(175.8)%
Pretax Margin (%)
(2,152.3)%
(914.8)%
Net Margin (%)
(2,152.6)%
(915.3)%
Debt-to-Equity Ratio (D/E) (%)
(144.0)%
(82.5)%
Net Debt / EBITDA (x)
(3.7)x
(3.0)x
Interest Coverage Ratio (x)
—
—
Current Ratio (%)
1,420.3%
843.0%
Quick Ratio (%)
865.2%
752.1%
Fixed Asset to Long-term Capital Ratio (%)
0.4%
(1.7)%
Plain English:
Anthropic’s financial ratios reveal a company experiencing rapid commercial expansion while operating with an unusually complex pre-IPO capital structure. Between FY2024 and FY2025, its gross profitability improved substantially, but operating losses, reported net losses, and financing-related accounting adjustments continued to affect several traditional financial ratios.
For beginner investors, the most important distinction is that Anthropic’s improving operating economics should not be confused with the mathematical effects of its negative shareholders’ equity and convertible financial instruments. Some ratios that appear positive are not reliable indicators of financial strength.
1. Gross Margin: A Significant Improvement in Customer-Serving Economics
Anthropic’s gross margin improved from (94.8)% in FY2024 to 39.5% in FY2025.
Gross margin measures the percentage of revenue remaining after subtracting the direct costs of providing products or services to customers.
In FY2024, Anthropic’s cost of revenue reached $752 million against revenue of $386 million. This means the direct cost of delivering its AI services exceeded the revenue generated from those services.
By FY2025, revenue increased to $4,585 million, while cost of revenue reached $2,773 million. Consequently, the company generated $1,812 million in gross profit.
This was an important change in the company’s financial structure. Revenue expanded considerably faster than the direct costs associated with customer-serving workloads.
For an AI company, these costs primarily include the computing infrastructure needed to process customer requests, along with certain personnel and allocated overhead expenses.
The improvement suggests that Anthropic became substantially more efficient at generating revenue relative to its direct service-delivery costs.
However, gross margin does not include the full cost of developing advanced AI models, conducting research, expanding the business, or maintaining corporate operations.
Plain English: In FY2024, Anthropic spent more delivering its AI services than it earned from customers. In FY2025, this changed: after direct service costs, the company retained approximately 39.5 cents of gross profit for every dollar of revenue. That was a meaningful improvement, but it did not mean Anthropic was profitable overall.
2. Operating Margin: Losses Remained Large Despite Revenue Growth
Anthropic’s operating margin improved from (772.0)% in FY2024 to (175.8)% in FY2025.
Operating margin measures operating income or loss as a percentage of revenue, after accounting for operating expenses.
The improvement was substantial, but both figures remained deeply negative.
In FY2024, Anthropic reported an operating loss of $2,980 million against revenue of $386 million. In FY2025, revenue increased to $4,585 million, while its operating loss widened to $8,060 million.
Although the absolute operating loss increased, revenue expanded much more rapidly. As a result, the operating loss became smaller relative to revenue.
Several major expenses contributed to the continued operating losses:
Research and Development: Expenses increased from $2,275 million in FY2024 to $6,338 million in FY2025 as Anthropic continued developing advanced AI models.
Computing Infrastructure: The company required substantial computing resources for model training, research, and commercial operations.
Sales and Marketing: Commercial expansion increased the costs associated with acquiring customers and supporting distribution partnerships.
Legal Settlement: FY2025 operating expenses included a $1,500 million charge related to the Bartz copyright litigation settlement.
These expenses help explain why improved gross profitability did not immediately translate into positive operating income.
The historical figures demonstrate an important distinction between business growth and operating profitability. Anthropic expanded its commercial revenue rapidly, but continued investing heavily in research and infrastructure.
Plain English: Anthropic lost less money relative to each dollar of revenue in FY2025 than in FY2024. However, its total operating loss still increased. The company was growing quickly, but it was also spending heavily to develop and expand its AI business.
3. Pretax Margin and Net Margin: Large Accounting Losses
Anthropic’s pretax margin improved from (2,152.3)% in FY2024 to (914.8)% in FY2025, while its net margin improved from (2,152.6)% to (915.3)%.
Pretax margin measures income or loss before income taxes relative to revenue. Net margin measures the final profit or loss remaining after all reported expenses and income taxes.
These extremely negative margins require careful interpretation.
Anthropic’s FY2025 net loss was $41,965 million, considerably larger than its $8,060 million operating loss.
The most important reason was a $34,308 million loss from changes in the fair value of convertible notes, derivative liabilities, and related financial instruments.
Certain convertible securities are recorded at their estimated fair value. Changes in that estimated value can create substantial accounting gains or losses even when no equivalent cash payment occurs during the reporting period.
As Anthropic’s estimated company value increased, the accounting value of certain outstanding convertible obligations also increased, producing large reported losses.
These accounting adjustments materially affected pretax and net margins.
Therefore, the large negative net margin should not be interpreted as evidence that Anthropic spent more than nine dollars in operating cash for every dollar of revenue during FY2025.
Nevertheless, the accounting losses remain relevant because convertible financial instruments can affect the company’s financial position and future ownership structure.
Plain English: Anthropic reported extremely large net losses, but much of the difference between operating loss and net loss came from changes in the estimated value of financing instruments. Investors should examine operating profitability and cash flow separately rather than treating the entire reported net loss as ordinary business spending.
4. ROE, ROA, ROTC, and ROIC: Why Some Positive Ratios Are Misleading
Traditional return ratios are designed to measure how effectively a company generates earnings from shareholders’ equity, assets, and invested capital.
However, Anthropic’s unusual balance sheet creates serious limitations when interpreting these ratios.
Return on Equity (ROE) measures net income relative to shareholders’ equity.
Anthropic’s calculated ROE was 84.8% in FY2024 and 83.1% in FY2025, even though the company reported large net losses in both years.
The reason is mathematical rather than economic.
Anthropic reported negative shareholders’ equity of $9,801 million at the end of FY2024 and negative shareholders’ equity of $50,505 million at the end of FY2025.
When both net income and shareholders’ equity are negative, dividing one negative number by another produces a positive result.
Consequently, Anthropic’s positive ROE does not represent a positive investment return or strong profitability. It is an accounting artifact created by dividing a loss by negative equity.
Return on Assets (ROA) compares net income with total assets.
Anthropic’s ROA declined from (107.2)% in FY2024 to (170.2)% in FY2025, reflecting the fact that reported net losses increased substantially relative to the company’s year-end asset base.
However, ROA is also heavily affected by the noncash fair-value adjustments included in reported net income.
Return on Total Capital (ROTC) measures operating income relative to the combined book value of debt and shareholders’ equity.
Using convertible notes payable as debt and the S-1’s reported shareholders’ equity, ROTC was (69.1)% in FY2024 and 91.3% in FY2025.
The positive FY2025 figure is misleading because reported shareholders’ equity was sufficiently negative to make the combined debt-and-equity denominator negative.
Return on Invested Capital (ROIC) measures tax-adjusted operating income relative to invested capital, excluding cash and cash equivalents.
ROIC was (253.9)% in FY2024 and 30.8% in FY2025 under the specified calculation method.
The FY2025 result again reflects a negative invested-capital denominator rather than positive operating performance.
In FY2024, invested capital was unusually small because positive convertible-note liabilities were largely offset by negative shareholders’ equity and cash holdings.
These effects make conventional return-on-capital ratios unsuitable for straightforward comparisons with established, profitable public companies.
Plain English: Some of Anthropic’s return ratios appear positive even though the company lost money. That happens because its reported shareholders’ equity is negative. Investors should not treat these positive percentages as evidence of profitability. Gross margin, operating margin, operating cash flow, and the company’s financing structure provide more useful information about its actual financial performance.
5. Debt-to-Equity and Net Debt / EBITDA: Financing Structure Matters
Anthropic’s debt-to-equity ratio was (144.0)% in FY2024 and (82.5)% in FY2025.
Debt-to-equity normally compares a company’s debt with the accounting value of shareholders’ equity. It helps investors evaluate how much a business relies on borrowed funds relative to equity financing.
For Anthropic, the ratio is negative because shareholders’ equity was negative in both years.
These negative percentages do not mean the company had negative debt or that its leverage was necessarily improving.
Anthropic’s convertible notes payable increased from $14,116 million at the end of FY2024 to $41,677 million at the end of FY2025.
These notes are financing instruments that may convert into shares under specified conditions. They are recorded at fair value, meaning their reported balance can change significantly as the company’s estimated valuation changes.
In addition, Anthropic had substantial redeemable convertible preferred stock outstanding. This financing was presented separately from ordinary shareholders’ equity, further limiting the usefulness of a conventional debt-to-equity calculation.
Net Debt / EBITDA compares debt after subtracting cash and cash equivalents with earnings before interest, taxes, depreciation, and amortization.
Anthropic’s calculated Net Debt / EBITDA was (3.7)x in FY2024 and (3.0)x in FY2025.
However, these negative multiples do not indicate that Anthropic held more cash than debt.
The company’s calculated net debt was positive in both years:
FY2024: Net debt of $10,975 million.
FY2025: Net debt of $24,289 million.
The negative multiples result from negative EBITDA, not negative net debt.
Using reported operating losses and depreciation and amortization from the consolidated cash flow statements, Anthropic’s EBITDA was:
FY2024: $(2,959) million.
FY2025: $(8,020) million.
Because EBITDA was negative in both years, these leverage multiples are not meaningful indicators of conventional debt-repayment capacity.
The Interest Coverage Ratio is not reported because the S-1 does not separately disclose a complete, independently verifiable annual interest expense figure suitable for the calculation. Interest accrued on certain convertible notes is incorporated into fair-value accounting adjustments, making it inappropriate to substitute the total fair-value loss for interest expense.
Plain English: Anthropic raised substantial financing through convertible securities. However, its negative equity and negative EBITDA make standard debt ratios difficult to interpret. A negative leverage multiple does not mean the company has no debt. Investors should evaluate its actual obligations, cash holdings, financing terms, and future funding requirements together.
6. Current Ratio and Quick Ratio: Strong Reported Short-Term Liquidity
Anthropic’s current ratio declined from 1,420.3% in FY2024 to 843.0% in FY2025.
The current ratio measures current assets divided by current liabilities. It shows how much short-term financial resources a company has relative to obligations generally due within one year.
At the end of FY2025, Anthropic reported:
Current Assets: $23,746 million.
Current Liabilities: $2,817 million.
This means current assets were approximately 8.4 times current liabilities.
The current ratio declined from FY2024 because current liabilities increased faster than current assets in percentage terms. Nevertheless, the company continued holding substantially more current assets than current liabilities.
The quick ratio provides a more restrictive measure of liquidity by excluding less immediately available current assets, such as prepaid expenses.
Anthropic’s quick ratio declined from 865.2% in FY2024 to 752.1% in FY2025.
Despite this decline, the ratio indicates that cash, short-term investments, and accounts receivable substantially exceeded current liabilities at both year-ends.
Anthropic’s cash and cash equivalents increased from $3,141 million at the end of FY2024 to $17,388 million at the end of FY2025. Short-term investments also increased to $2,889 million.
However, the increase in liquidity should not be attributed solely to operating performance. Anthropic continued raising external financing during the period, while operating cash flow remained negative.
Moreover, current ratios primarily measure short-term balance sheet liquidity. They do not capture the full scale of the company’s future computing and infrastructure commitments.
Plain English: Anthropic had a large amount of cash and other short-term assets relative to its current liabilities at the end of FY2025. This provided financial flexibility, but the company continued relying on external financing and faced substantial longer-term funding requirements.
7. Fixed Asset to Long-term Capital Ratio: Limited Comparability
Anthropic’s fixed asset to long-term capital ratio was 0.4% in FY2024 and (1.7)% in FY2025.
This ratio compares property, plant, and equipment with long-term capital, defined here as long-term debt plus shareholders’ equity.
Property, plant, and equipment includes physical assets such as computing equipment, networking infrastructure, and leasehold improvements.
Anthropic’s reported net property and equipment increased from $19 million in FY2024 to $146 million in FY2025.
However, the FY2025 ratio became negative because the company’s negative shareholders’ equity exceeded the reported balance of convertible notes payable, producing a negative long-term capital denominator.
As a result, the negative percentage does not indicate that Anthropic had negative fixed assets or that its physical infrastructure declined.
Another important consideration is that Anthropic relies extensively on computing capacity obtained through external infrastructure providers and contractual arrangements.
Consequently, the company’s reported property and equipment balance does not necessarily reflect the full scale of computing resources used in its operations or committed for future periods.
Plain English: Anthropic’s reported physical assets increased, but its unusual capital structure makes this ratio difficult to interpret. The company also obtains substantial computing capacity through external providers, so its balance sheet does not capture the full economic scale of its infrastructure arrangements.
Calculation Notes:
ROE: Net income divided by year-end total shareholders’ equity.
ROA: Net income divided by year-end total assets.
ROTC: Operating income divided by total debt plus total shareholders’ equity.
ROIC: Operating income multiplied by one minus the actual effective income tax rate, divided by total debt plus shareholders’ equity minus cash and cash equivalents.
Effective Tax Rate: Income tax expense divided by income before taxes, using the reported figures for each fiscal year.
Debt-to-Equity: Total debt divided by total shareholders’ equity.
Net Debt / EBITDA: Total debt minus cash and cash equivalents, divided by operating income plus reported depreciation and amortization. FY2024 depreciation and amortization was $21 million, and FY2025 was $40 million.
Current Ratio: Total current assets divided by total current liabilities, multiplied by 100.
Quick Ratio: Cash and cash equivalents plus short-term investments and net accounts receivable, divided by current liabilities, multiplied by 100.
Fixed Asset to Long-term Capital Ratio: Net property and equipment divided by long-term debt plus total shareholders’ equity, multiplied by 100.
Overall Financial Ratio Interpretation:
Anthropic’s financial ratios reveal two developments occurring simultaneously. First, the company’s gross margin improved substantially as revenue expanded faster than direct customer-serving costs. Second, the company’s reported profitability and capital structure remained heavily affected by research spending, financing arrangements, and noncash accounting adjustments.
Traditional leverage and return-on-capital ratios are particularly difficult to interpret because Anthropic reported negative shareholders’ equity and large fair-value liabilities associated with convertible instruments.
For beginner investors, the most useful conclusion is that Anthropic made substantial progress in its customer-serving economics during FY2025, but it had not yet achieved overall operating profitability.
Evaluating the company’s financial position therefore requires looking beyond individual ratios to examine the underlying balance sheet, cash flows, financing transactions, and commitments associated with its AI infrastructure.
3) Balance Sheet Summary
Unit: $m. All figures are in USD millions. Parentheses indicate negative amounts. The symbol “—” indicates an item that was not separately disclosed.
Balance Sheet Summary Template
FY 2024
FY 2025
Assets
Cash & Equivalents
$3,141
$17,388
Accounts Receivable
$95
$909
Inventory
—
—
Current Assets
$5,312
$23,746
Property, Plant & Equipment
$19
$146
Intangible Assets
—
—
Non-current Assets
$2,440
$915
Total Assets
$7,752
$24,661
Liabilities
Short-term Debt
—
—
Accounts Payable
—
—
Current Liabilities
$374
$2,817
Long-term Debt
$14,116
$41,677
Non-current Liabilities
$14,305
$42,580
Total Liabilities
$14,679
$45,397
Redeemable Convertible Preferred Stock
$2,874
$29,769
Equity
Common Equity (Stockholders’ Deficit)
$(9,801)
$(50,505)
Total Liabilities + Preferred Stock + Equity
$7,752
$24,661
Balance Sheet Note: Redeemable convertible preferred stock is reported separately from liabilities and shareholders’ equity. Long-term debt represents convertible notes payable. Non-current assets and liabilities are calculated from reported totals. Accounts payable is not separately disclosed. Goodwill is reported separately from intangible assets.
Plain English:
Anthropic’s balance sheet expanded substantially between FY2024 and FY2025, reflecting rapid capital raising, increasing commercial activity, and continued investment in artificial intelligence development.
Total assets increased from $7,752 million in FY2024 to $24,661 million in FY2025, representing an increase of approximately 218.1%.
However, this expansion was not primarily the result of accumulated operating profits. Anthropic continued reporting substantial losses while raising external capital through preferred stock and convertible financial instruments.
For beginner investors, the distinction is important: a rapidly expanding balance sheet can reflect successful fundraising and business growth even when the company has not yet become profitable.
1. Cash and Liquidity Expanded Significantly
One of the most important changes was the substantial increase in Anthropic’s cash holdings.
Cash and cash equivalents increased from $3,141 million at the end of FY2024 to $17,388 million at the end of FY2025, an increase of $14,247 million.
The company also reported $2,889 million in short-term investments at the end of FY2025, compared with no separately reported short-term investment balance at the end of FY2024.
Short-term investments are financial assets that a company generally expects to convert into cash within a relatively short period. Anthropic’s portfolio includes instruments such as government securities, corporate debt securities, and commercial paper.
Combining cash and cash equivalents with short-term investments, Anthropic held approximately $20,277 million in these financial assets at the end of FY2025.
This provided the company with substantial financial resources to support its operations and future commitments.
Nevertheless, the source of this liquidity is essential to understanding the company’s financial position.
Anthropic generated negative operating cash flow during FY2025, meaning its operating activities consumed cash rather than generating it.
The increase in cash holdings was therefore primarily supported by external financing, including the issuance of preferred stock and convertible notes.
Plain English: Anthropic ended FY2025 with much more cash than it had a year earlier. However, that cash came primarily from investors and financing transactions rather than profits generated by the business. The company had substantial liquidity, but it was still dependent on external capital to support expansion.
Accounts receivable increased from $95 million in FY2024 to $909 million in FY2025.
Accounts receivable represents money customers owe a company for products or services already delivered but not yet paid for.
Anthropic’s receivables increased as its commercial revenue expanded rapidly during FY2025.
The increase was substantial in absolute terms, but it was smaller than the percentage increase in annual revenue. Revenue grew from $386 million to $4,585 million, while accounts receivable increased from $95 million to $909 million.
At year-end, accounts receivable represented approximately 19.8% of FY2025 annual revenue, compared with approximately 24.6% in FY2024.
This comparison suggests that year-end receivables did not expand as quickly as annual revenue. However, because Anthropic’s business was growing rapidly, these percentages alone cannot establish whether customers were paying more quickly.
Evaluating collection efficiency would require additional information about billing patterns, revenue timing, and customer payment terms.
Anthropic also distributes Claude through major cloud service providers, which can affect the timing and structure of customer payments.
Plain English: Anthropic had more unpaid customer invoices at the end of FY2025 because its business had grown considerably. The increase in receivables was smaller than the increase in revenue, but investors should not interpret this relationship alone as proof that customer payments became faster.
3. Physical Assets Expanded, but Do Not Represent the Full Computing Business
Property, plant, and equipment increased from $19 million in FY2024 to $146 million in FY2025.
These assets include physical equipment and other long-lived resources used in the company’s operations.
For traditional manufacturing businesses, property, plant, and equipment may represent a substantial share of total assets because companies directly own factories, machinery, and production facilities.
Anthropic operates differently.
Although advanced AI development requires enormous computing resources, the company obtains substantial computing capacity through arrangements with cloud providers and other infrastructure partners.
Consequently, the amount of computing infrastructure used by Anthropic is not fully reflected in the company’s reported property and equipment balance.
At the end of FY2025, property and equipment represented approximately 0.6% of total assets.
This does not mean Anthropic operates a low-cost or infrastructure-light business in an economic sense.
Instead, the company relies extensively on external computing infrastructure and contractual arrangements that can create substantial future financial obligations.
The balance sheet also included $154 million in goodwill at the end of FY2025, compared with no goodwill balance in FY2024.
Goodwill is an accounting asset that can arise when a company acquires another business for more than the fair value of its identifiable net assets. It may reflect expected benefits such as workforce expertise, technological capabilities, or business synergies.
Plain English: Anthropic reported relatively little directly owned physical equipment compared with its total assets. However, the company still depends on expensive computing infrastructure. Much of that infrastructure is accessed through outside providers rather than owned directly, so investors must examine its contractual obligations as well as its balance sheet.
4. Liabilities Increased Sharply, Led by Convertible Notes
Total liabilities increased from $14,679 million at the end of FY2024 to $45,397 million at the end of FY2025.
The largest component was convertible notes payable, which increased from $14,116 million to $41,677 million.
Convertible notes are financial instruments that may convert into shares of a company’s stock under specified conditions.
Unlike ordinary loans, certain convertible notes can be measured at their estimated fair value. This means their reported balance may change as the company’s estimated value, conversion expectations, and other financial assumptions change.
Anthropic’s increase in convertible note liabilities therefore should not be interpreted entirely as new cash borrowing.
Part of the increase reflected changes in the estimated fair value of outstanding financial obligations.
This distinction matters because the balance sheet value of a convertible note can differ substantially from the original amount of cash raised through its issuance.
Current liabilities also increased, from $374 million in FY2024 to $2,817 million in FY2025.
The company reported $2,511 million in accounts payable, accrued expenses, and other current liabilities at the end of FY2025, compared with $333 million at the end of FY2024.
This combined category includes obligations arising from business operations and other amounts payable. The balance sheet does not disclose accounts payable as a separate figure.
Deferred revenue increased from $28 million to $269 million.
Deferred revenue represents customer payments received, or amounts billed, before the related products or services have been fully delivered. The company generally recognizes the revenue as it fulfills its obligations.
The increase is consistent with Anthropic’s expanding commercial activities, although deferred revenue is not itself a measure of total future sales.
Plain English: Anthropic’s reported liabilities increased significantly, especially its convertible financing obligations. However, the growth in convertible note balances reflects both financing arrangements and accounting revaluations. Investors should not assume that every dollar of additional reported debt represents newly borrowed cash.
5. Redeemable Preferred Stock and Negative Shareholders’ Equity
One of the most unusual characteristics of Anthropic’s balance sheet is the relationship between convertible preferred stock and shareholders’ equity.
Redeemable convertible preferred stock increased from $2,874 million in FY2024 to $29,769 million in FY2025.
Preferred stock is a form of corporate ownership that can provide investors with rights different from those attached to ordinary common shares.
Depending on its terms, convertible preferred stock may be exchanged for common shares. It may also provide investors with preferential rights in certain transactions, such as a liquidation.
Anthropic classifies its redeemable convertible preferred stock outside permanent shareholders’ equity because certain redemption-related events could occur outside the company’s control.
This separate classification is essential to understanding the company’s reported financial position.
At the same time, Anthropic’s total shareholders’ deficit increased from $(9,801) million in FY2024 to $(50,505) million in FY2025.
A shareholders’ deficit means reported liabilities and separately classified capital instruments exceed the assets available to support ordinary shareholders’ book equity.
Anthropic’s accumulated deficit, which reflects accumulated losses over time, increased from $(10,040) million to $(52,154) million.
The substantial FY2025 net loss was the principal reason for the deterioration in reported shareholders’ equity.
However, a negative shareholders’ equity balance does not automatically mean a company has insufficient cash to pay its immediate obligations.
Anthropic held substantial cash and short-term investments at the end of FY2025, even though its reported shareholders’ equity was deeply negative.
The company can therefore have substantial liquidity while simultaneously reporting a large shareholders’ deficit.
These amounts reconcile because total liabilities plus redeemable preferred stock plus the negative shareholders’ equity balance equal total assets.
This structure also explains why conventional financial ratios based on ordinary shareholders’ equity can produce misleading results.
Plain English: Anthropic raised substantial capital by issuing preferred shares and convertible financial instruments, but its accounting treatment separates much of that financing from ordinary shareholders’ equity. Combined with its accumulated losses, this produced a large negative equity balance. That negative balance does not mean the company had run out of cash, but it does make its capital structure more complex than that of a typical profitable public company.
Overall Balance Sheet Interpretation
Anthropic’s FY2025 balance sheet reflects a company that rapidly expanded its financial resources while continuing to operate at a substantial accounting loss.
Its cash position strengthened considerably, total assets more than tripled, and commercial activity increased. However, the company also accumulated significantly larger convertible financial obligations and a deeper shareholders’ deficit.
For beginner investors, the most important distinction is between liquidity, profitability, and capital structure.
Liquidity measures the financial resources available to meet obligations. Profitability measures whether a business earns more than it spends. Capital structure describes how the business is financed through debt, preferred stock, and ordinary equity.
Anthropic had substantial liquidity at the end of FY2025, but it had not yet achieved overall profitability, and much of its financial position reflected financing arrangements with complex accounting characteristics.
Therefore, the balance sheet should not be evaluated solely by looking at the company’s cash holdings or negative shareholders’ equity.
The central takeaway is that Anthropic entered 2026 with substantially greater financial resources, but its continued expansion remained closely connected to external financing and large future computing commitments.
4) Cash Flow Statement Summary
Unit: $m. All figures are in USD millions. Parentheses indicate cash outflows or negative amounts.
Cash Flow Statement Summary Template
FY 2023
FY 2024
FY 2025
Cash Flow from Operating Activities
$(1,627)
$(5,692)
$(2,637)
Cash Flow from Investing Activities
$(2)
$(22)
$(3,077)
Cash Flow from Financing Activities
$2,320
$7,248
$19,960
Net Change in Cash
$691
$1,534
$14,247
Beginning Cash Balance
$929
$1,620
$3,154
Ending Cash Balance
$1,620
$3,154
$17,401
Plain English:
Anthropic’s cash flow statements reveal a company experiencing rapid commercial expansion while continuing to rely heavily on external financing.
Between FY2023 and FY2025, the company’s cash balance increased substantially, reaching $17,401 million at the end of FY2025. However, this increase did not result from sustained positive operating cash flow.
Instead, Anthropic continued consuming cash through its operations while raising significant amounts of capital from investors and financing partners.
For beginner investors, the distinction is essential: having a large cash balance does not necessarily mean a company is generating cash from its business.
1. Operating Cash Flow: Cash Consumption Improved in FY2025
Operating cash flow measures the cash generated or consumed by a company’s ordinary business activities. It reflects cash collected from customers, payments to suppliers and employees, and changes in operating assets and liabilities.
Anthropic reported negative operating cash flow throughout the three-year period:
FY2023: $(1,627) million.
FY2024: $(5,692) million.
FY2025: $(2,637) million.
Cash consumption increased significantly in FY2024 as Anthropic expanded its AI development activities and computing requirements.
However, operating cash flow improved by $3,055 million in FY2025, representing a reduction of approximately 53.7% in operating cash outflow compared with FY2024.
This improvement occurred while annual revenue increased from $386 million to $4,585 million.
The combination suggests that Anthropic’s expanding commercial operations helped improve its cash flow position, even though the company had not yet achieved positive operating cash flow.
Nevertheless, operating cash flow should not be evaluated solely by comparing revenue growth with reported cash outflows.
Changes in working capital also played an important role.
Working capital refers to short-term operating assets and liabilities, including accounts receivable, prepaid expenses, accounts payable, and deferred revenue.
In FY2025, several changes affected Anthropic’s operating cash flow:
Accounts Receivable: An increase in customer receivables consumed $834 million in operating cash flow.
Prepaid Expenses and Other Assets: Changes in these accounts contributed $1,771 million to operating cash flow.
Accounts Payable, Accrued Expenses, and Other Liabilities: Changes in these obligations contributed $2,079 million.
Deferred Revenue: An increase in payments received or amounts billed before revenue recognition contributed $241 million.
These changes helped reduce the cash consumed by operations during FY2025.
For example, an increase in accounts payable or accrued liabilities can temporarily support operating cash flow because expenses have been recognized before the associated cash payments occur.
Similarly, an increase in deferred revenue can improve cash flow when customers pay before Anthropic fully delivers the related services.
These effects are economically meaningful, but they should not automatically be interpreted as permanent improvements in profitability.
2. Why Operating Cash Flow Was Much Better Than Net Income
Another important characteristic of Anthropic’s financial statements is the substantial difference between reported net losses and actual operating cash outflows.
In FY2025, Anthropic reported:
Net Loss: $(41,965) million.
Operating Cash Flow: $(2,637) million.
The reported net loss was therefore much larger than the cash consumed by operating activities.
The primary explanation was the recognition of substantial noncash accounting expenses.
Noncash expenses reduce reported earnings without requiring an equivalent cash payment during the period in which they are recognized.
Anthropic’s FY2025 cash flow statement included the following major adjustments:
Fair-Value Changes in Convertible Notes, Derivative Liabilities, and SAFEs: $34,308 million.
Stock-Based Compensation: $1,158 million.
Donation Matching Program Liability: $541 million.
Depreciation and Amortization: $40 million.
The largest adjustment was the $34,308 million loss associated with changes in the fair value of convertible financial instruments.
Because this adjustment did not represent an equivalent operating cash payment, it was added back when reconciling net loss to operating cash flow.
Stock-based compensation was another important adjustment.
Stock-based compensation represents the accounting expense associated with employee equity awards. Although it reduces reported earnings, it generally does not require the same immediate cash payment as ordinary salaries.
However, stock-based compensation can create shareholder dilution when additional shares are issued or become outstanding.
The accounting adjustment therefore should not be interpreted as making employee compensation economically free.
Plain English: Anthropic reported a net loss of almost $42,000 million in FY2025, but its operations consumed approximately $2,600 million in cash. The enormous difference came primarily from noncash accounting adjustments related to convertible financing instruments. This is why investors should examine cash flow alongside net income rather than relying on reported earnings alone.
Investing cash flow measures cash used to acquire long-term assets, purchase investments, or make other investing transactions. It also includes cash received when certain investments are sold or mature.
Anthropic’s investing cash flow was:
FY2023: $(2) million.
FY2024: $(22) million.
FY2025: $(3,077) million.
The substantial increase in investing cash outflows during FY2025 requires careful interpretation.
Unlike a traditional manufacturer, whose investing cash outflows may be dominated by factory construction and machinery purchases, Anthropic’s FY2025 investing cash flow primarily reflected purchases of short-term financial investments.
The principal FY2025 investing activities included:
Purchases of Short-Term Investments: $(3,082) million.
Sales and Maturities of Short-Term Investments: $213 million.
Purchases of Property and Equipment: $(131) million.
Other Investing Activities: $(77) million.
Net purchases of short-term investments accounted for $2,869 million of cash outflows during FY2025.
These purchases generally represent the allocation of available cash into financial assets rather than direct spending on AI research or physical infrastructure.
This distinction matters because purchasing a short-term financial investment changes the composition of a company’s assets. It does not necessarily represent a permanent consumption of financial resources.
Meanwhile, Anthropic’s purchases of property and equipment increased from $2 million in FY2023 to $15 million in FY2024 and $131 million in FY2025.
This spending represents capital expenditures, commonly called CapEx.
Capital expenditures are cash payments used to acquire or improve long-lived assets, such as equipment and facilities.
However, Anthropic’s relatively modest reported capital expenditures should not be interpreted as evidence that developing and operating its AI models requires little infrastructure investment.
The company relies extensively on computing infrastructure obtained from external providers. Payments for certain computing services are reflected in operating activities rather than purchases of property and equipment.
In addition, large contractual commitments for future computing capacity may not appear as current-period capital expenditures.
Consequently, reported CapEx alone does not capture the full economic cost of Anthropic’s computing infrastructure.
4. Free Cash Flow: Improvement Without a Positive Balance
Free cash flow, or FCF, is commonly calculated as operating cash flow minus capital expenditures.
It measures the cash remaining after a company funds its operations and purchases long-lived assets.
Using cash flow from operating activities less purchases of property and equipment, Anthropic’s historical free cash flow was:
FY2023: $(1,629) million.
FY2024: $(5,707) million.
FY2025: $(2,768) million.
Free cash flow improved by $2,939 million in FY2025 compared with FY2024.
This improvement was driven primarily by the reduction in operating cash outflows rather than a reduction in capital expenditures.
In fact, property and equipment purchases increased from $15 million in FY2024 to $131 million in FY2025.
Nevertheless, free cash flow remained negative, meaning Anthropic continued consuming cash after accounting for its reported capital expenditures.
Furthermore, conventional free cash flow does not directly measure all future obligations associated with long-term cloud computing and infrastructure agreements.
For an AI company whose computing capacity depends heavily on external providers, those commitments are an important additional consideration.
Plain English: Anthropic’s free cash flow improved significantly in FY2025, but the company still spent more cash on operations and capital expenditures than it generated. Its reported CapEx was relatively small compared with its overall computing requirements because much of its infrastructure is supplied through external arrangements.
5. Financing Cash Flow: External Capital Funded Expansion
Financing cash flow measures cash received from or paid to investors, lenders, and other financing counterparties.
Anthropic reported positive financing cash flow throughout the period:
FY2023: $2,320 million.
FY2024: $7,248 million.
FY2025: $19,960 million.
Financing inflows increased substantially as Anthropic raised capital to support its expanding operations and AI development activities.
In FY2023, financing cash flow primarily reflected $1,750 million raised through convertible notes, $445 million from redeemable convertible preferred stock, and $122 million from SAFEs.
In FY2024, the company raised $6,050 million through convertible notes and $1,188 million through preferred stock issuance.
By FY2025, financing activity had expanded significantly.
The company’s major FY2025 financing transactions included:
Convertible Notes Issued for Cash: $3,450 million.
Common Stock Issuance: $111 million.
Common Stock Repurchases: $(149) million.
These transactions produced $19,960 million in net financing cash inflows.
Preferred stock issuance accounted for approximately 82.9% of FY2025 net financing cash flow.
This demonstrates how important external equity-related financing had become to Anthropic’s financial strategy.
Capital raised through preferred shares and convertible instruments allowed Anthropic to strengthen its cash position despite continued operating losses.
However, external financing is not equivalent to recurring cash generated by customers.
The availability and terms of future financing may depend on investor demand, the company’s valuation, business performance, competitive conditions, and broader financial markets.
Convertible financing instruments can also affect future ownership and shareholder dilution, depending on their contractual terms and eventual conversion.
6. Capital Returns: Growth Took Priority Over Dividends
Anthropic’s historical cash flow statements do not show cash dividend payments to shareholders.
The company also states in its S-1 that it does not expect to pay cash dividends in the foreseeable future and intends to retain available funds and future earnings, if any, to finance its operations and expansion.
Although Anthropic recorded $149 million in common stock repurchases during FY2025, this transaction should not be interpreted as evidence of a recurring public-market share repurchase program.
Relative to the company’s nearly $20,000 million in financing inflows, the repurchase amount was limited.
The company’s overall financing activities were directed primarily toward raising capital rather than distributing cash to shareholders.
For prospective IPO investors, this means investment returns would depend primarily on future share-price appreciation rather than regular dividend income, unless the company’s policy changes.
Plain English: Anthropic raised far more cash from investors and financing partners than it returned to shareholders. Its priority was funding business expansion, not paying dividends or conducting large-scale shareholder distributions.
7. Ending Cash Balance: Strong Liquidity Supported by Financing
Anthropic’s ending cash balance increased from $1,620 million in FY2023 to $3,154 million in FY2024 and $17,401 million in FY2025.
The FY2025 increase of $14,247 million was particularly significant.
The principal cash flow components during FY2025 were:
Operating Activities: $(2,637) million.
Investing Activities: $(3,077) million.
Financing Activities: $19,960 million.
Foreign Exchange Effect: $1 million.
These amounts reconcile to the reported $14,247 million increase in cash, cash equivalents, and restricted cash.
The ending balance of $17,401 million includes $17,388 million in cash and cash equivalents and $13 million in restricted cash.
Restricted cash refers to cash subject to limitations on its use, such as amounts held for specified contractual purposes.
This explains why the ending cash balance in the cash flow statement is slightly higher than the cash and cash equivalents balance presented separately on the FY2025 balance sheet.
The overall financial picture is clear: Anthropic’s cash resources expanded considerably because financing inflows exceeded the combined cash used in operating and investing activities.
Although the company’s operating cash consumption improved in FY2025, the business had not yet become self-funding.
Future liquidity will depend on the relationship between revenue growth, operating cash generation, infrastructure spending, contractual obligations, and financing availability.
Overall Cash Flow Interpretation
Anthropic’s historical cash flow statements show three important structural developments.
First, operating cash flow improved significantly in FY2025, even though the company continued reporting substantial accounting losses. This reflects the need to distinguish noncash financial adjustments from actual operating cash consumption.
Second, investing cash outflows increased primarily because of purchases of short-term financial investments, not because the company spent the entire amount constructing physical infrastructure. Its broader computing requirements must also be evaluated through operating expenses and contractual commitments.
Third, external financing remained the primary source of Anthropic’s expanding cash reserves. The company raised nearly $20,000 million through financing activities during FY2025, allowing it to increase liquidity substantially despite negative operating and investing cash flows.
For beginner investors, the central takeaway is that Anthropic entered 2026 with a much stronger cash position, but its operations had not yet achieved positive cash generation.
The most important financial development to monitor is whether the company’s rapidly growing revenue can eventually support its operating costs and infrastructure requirements without continued dependence on large external financing transactions.
5) Historical Performance Takeaways
Anthropic’s financial performance between FY2023 and FY2025 reveals a rapidly expanding artificial intelligence business with improving customer-serving economics, substantial operating expenses, and a continued dependence on external financing.
Revenue Growth Was Exceptional: Annual revenue increased from $41 million in FY2023 to $386 million in FY2024 and $4,585 million in FY2025. Revenue grew approximately 1,087.8% in FY2025 alone, reflecting rapidly expanding commercial demand for Claude and Anthropic’s AI services.
Gross Profitability Improved Significantly: Gross margin increased from (94.8)% in FY2024 to 39.5% in FY2025. This indicates that revenue grew faster than the direct costs of providing AI services to customers. The improvement was an important development in Anthropic’s underlying business economics.
Operating Losses Remained Substantial: Operating losses increased from $551 million in FY2023 to $2,980 million in FY2024 and $8,060 million in FY2025. Although operating margin improved as revenue expanded, the company continued spending heavily on AI model development, computing infrastructure, personnel, and commercial expansion. FY2025 also included a $1,500 million litigation settlement charge.
Reported Net Losses Were Heavily Affected by Noncash Accounting Adjustments: Anthropic reported a net loss of $41,965 million in FY2025. However, $34,308 million of the reported loss resulted from changes in the fair value of convertible notes, derivative liabilities, and related financial instruments. These accounting adjustments should be distinguished from ordinary operating expenses and actual cash consumption.
Operating Cash Flow Improved but Remained Negative: Operating cash outflow decreased from $5,692 million in FY2024 to $2,637 million in FY2025, an improvement of approximately 53.7%. This suggests that Anthropic’s cash flow position improved as commercial revenue expanded, although the business had not yet achieved positive operating cash generation.
Free Cash Flow Also Improved: Free cash flow, calculated as operating cash flow minus purchases of property and equipment, improved from $(5,707) million in FY2024 to $(2,768) million in FY2025. Nevertheless, the company continued consuming cash after funding its operations and reported capital expenditures.
External Financing Drove the Increase in Cash Reserves: Financing cash inflows reached $19,960 million in FY2025, primarily through preferred stock issuance and convertible notes. These transactions helped increase cash and cash equivalents to $17,388 million at year-end, with an additional $2,889 million held in short-term investments.
The Balance Sheet Became Larger and More Complex: Total assets increased from $7,752 million in FY2024 to $24,661 million in FY2025. However, total liabilities also increased substantially, reaching $45,397 million, while redeemable convertible preferred stock increased to $29,769 million. Anthropic’s shareholders’ deficit widened to $(50,505) million.
Traditional Return and Leverage Ratios Require Caution: Negative shareholders’ equity and the fair-value accounting treatment of convertible financial instruments make ratios such as ROE, ROTC, ROIC, and debt-to-equity difficult to interpret. Positive percentages produced by negative capital denominators do not indicate that Anthropic was profitable.
Computing Infrastructure Remained a Major Financial Commitment: Anthropic relies extensively on external infrastructure providers for AI model training and inference. Consequently, reported property and equipment purchases do not fully represent the company’s computing requirements. Operating expenses, infrastructure agreements, and future contractual obligations are also important when evaluating its financial position.
Overall Historical Performance Interpretation
Anthropic’s FY2023–FY2025 financial statements show a company transitioning from an early-stage AI developer into a rapidly expanding commercial AI platform.
The most encouraging development was the combination of exceptional revenue growth, a substantial improvement in gross margin, and a reduction in operating cash consumption during FY2025. Together, these changes indicate that commercial expansion was beginning to improve aspects of the company’s underlying financial performance.
However, the historical results also reveal significant challenges. Anthropic continued reporting large operating losses, invested heavily in developing increasingly capable AI models, and depended on external financing to support its activities.
Its reported net losses were further complicated by substantial noncash fair-value adjustments, while negative shareholders’ equity and separately classified preferred stock made conventional financial ratios less informative.
For beginner investors, the central distinction is between commercial success and financial self-sufficiency. Anthropic demonstrated that demand for its AI products could grow extraordinarily quickly, but it had not yet demonstrated that its business could consistently generate enough operating cash to finance its own expansion.
The key question going forward is whether continued revenue growth and improved operating efficiency can eventually support Anthropic’s substantial research, computing, and infrastructure requirements without the same degree of dependence on external capital.
2.2 First Half 2026 Financial Performance
1) Income Statement Summary (Unaudited)
Unit: $m. Q2 represents the three months ended June 30, while 6M represents the six months ended June 30. Parentheses indicate losses.
Income Statement Summary (Unaudited)
Q2 FY2025
Q2 FY2026
6M FY2025
6M FY2026
Revenue
$787
$11,555
$1,163
$16,286
Operating Income
$(1,200)
$974
$(2,135)
$(693)
Net Income
$(6,377)
$(66,309)
$(12,969)
$(80,748)
Plain English:
Anthropic’s financial performance improved significantly during the first half of 2026, reflecting extraordinary revenue growth and a substantial improvement in operating profitability. However, reported net losses remained exceptionally large because of noncash accounting adjustments associated with convertible financial instruments.
1. Revenue Growth Accelerated Dramatically
Anthropic’s revenue increased from $787 million in Q2 FY2025 to $11,555 million in Q2 FY2026, representing year-over-year growth of approximately 1,368.2%.
For the first six months of the year, revenue increased from $1,163 million in 2025 to $16,286 million in 2026, an increase of approximately 1,300.3%.
This extraordinary expansion reflected growing commercial adoption of Claude, increasing usage of Anthropic’s AI models, and higher spending by business customers.
The company’s revenue growth was driven primarily by consumption-based services, which generate revenue according to customer usage, alongside subscription-based products.
For beginner investors, this means Anthropic was not simply attracting more customers. Existing and new customers were collectively spending substantially more on its AI services.
However, these exceptionally high growth rates also reflect the company’s relatively small revenue base in the prior year. Investors should not assume that similar percentage growth can continue indefinitely.
2. Operating Profitability Improved Substantially
One of the most important developments was Anthropic’s transition to positive quarterly operating income.
In Q2 FY2026, Anthropic reported $974 million in operating income, compared with an operating loss of $1,200 million in Q2 FY2025.
This represents an improvement of $2,174 million.
Operating margin improved from approximately (152.5)% to 8.4% over the same period.
Operating margin measures the percentage of revenue remaining after operating expenses, including computing infrastructure, personnel, research-related costs, and commercial expenses.
The improvement occurred because revenue increased faster than total operating expenses.
Although Anthropic continued investing heavily in AI research, computing infrastructure, and commercial expansion, its rapidly growing revenue base was sufficient to cover reported operating expenses during Q2 FY2026.
The six-month results also improved substantially.
Operating loss decreased from $(2,135) million in the first half of 2025 to $(693) million in the first half of 2026, an improvement of approximately 67.5%.
Operating margin improved from approximately (183.6)% to (4.3)%.
However, the six-month result remained negative because Anthropic recorded an operating loss of $1,667 million during Q1 FY2026, exceeding the operating profit generated in the second quarter.
Plain English: Anthropic reached an important milestone in Q2 FY2026 by generating an operating profit. Its revenue grew quickly enough to exceed operating expenses during the quarter. Nevertheless, the company remained slightly unprofitable on an operating basis for the first half of 2026 as a whole.
3. Net Losses Remained Large Because of Financing-Related Accounting Adjustments
Despite its improving operating performance, Anthropic reported a net loss of $66,309 million in Q2 FY2026, compared with $6,377 million in Q2 FY2025.
For the first six months of 2026, net loss reached $80,748 million, compared with $12,969 million during the same period in 2025.
The enormous difference between operating income and net income requires careful interpretation.
During the first half of 2026, Anthropic recognized a $72,388 million loss from changes in the fair value of convertible notes, derivative liabilities, and SAFEs.
These losses primarily reflected changes in the estimated accounting value of outstanding financing obligations, influenced by increases in Anthropic’s estimated enterprise valuation.
The company also recorded substantial other non-operating expenses associated with financial instruments.
These accounting adjustments are not equivalent to cash spent operating Claude or developing AI models during the reporting period.
However, the underlying financing instruments remain economically important because their terms may affect future obligations, ownership structure, and shareholder dilution.
Plain English: Anthropic generated an operating profit in Q2 FY2026 but still reported an enormous net loss. Most of the difference came from financing-related accounting adjustments rather than ordinary operating expenses. Investors should therefore examine operating performance and cash flow separately from reported net income.
Overall Income Statement Interpretation
Anthropic’s first-half 2026 results demonstrate a significant improvement in its underlying operating performance.
Revenue expanded extraordinarily quickly, operating losses narrowed substantially on a six-month basis, and the company achieved positive operating income during the second quarter.
Nevertheless, Anthropic continued reporting substantial net losses because of noncash fair-value adjustments associated with its pre-IPO financing structure.
For beginner investors, the key takeaway is that Anthropic’s commercial operations became substantially more profitable during the first half of 2026, even though its reported net income deteriorated sharply.
The most important question going forward is whether the company can sustain positive operating profitability while continuing to invest in advanced AI models, computing infrastructure, and commercial expansion.
2) Key Profitability Ratios
Unit: %. All percentages are rounded to one decimal place. Q2 represents the three months ended June 30, while 6M represents the six months ended June 30. Parentheses indicate negative margins.
Ratio
Q2 FY2025
Q2 FY2026
6M FY2025
6M FY2026
Operating Margin (%)
(152.5)%
8.4%
(183.6)%
(4.3)%
Net Margin (%)
(810.3)%
(573.9)%
(1,115.1)%
(495.8)%
Plain English: Anthropic’s profitability ratios improved substantially during the first half of 2026. Most importantly, the company achieved a positive operating margin in the second quarter, demonstrating that quarterly revenue exceeded its reported operating expenses. However, net margin remained deeply negative because of substantial accounting losses associated with convertible financial instruments. Understanding the difference between operating margin and net margin is essential when evaluating Anthropic’s financial performance.
1. Operating Margin: A Transition to Quarterly Operating Profitability
Anthropic’s operating margin improved from (152.5)% in Q2 FY2025 to 8.4% in Q2 FY2026, an improvement of approximately 160.9 percentage points.
Operating margin measures operating income or loss as a percentage of revenue after deducting operating expenses.
For Anthropic, these expenses include computing infrastructure, personnel, research and development activities, sales and marketing, and general administration.
In Q2 FY2025, Anthropic reported:
Revenue: $787 million.
Operating Loss: $(1,200) million.
Operating Margin: (152.5)%.
By Q2 FY2026, the company’s financial performance had changed significantly:
Revenue: $11,555 million.
Operating Income: $974 million.
Operating Margin: 8.4%.
Revenue increased by approximately 1,368.2% year over year, while total operating expenses increased at a slower rate.
This difference allowed Anthropic to move from a substantial quarterly operating loss to positive operating income.
The improvement is especially important because developing and operating advanced AI models requires significant expenditures on computing infrastructure, research personnel, and ongoing technological development.
Anthropic continued increasing its spending in absolute dollars, but the much larger revenue base allowed the company to absorb those expenses more effectively.
The six-month comparison reveals a similarly significant improvement.
6M FY2025 Operating Margin: (183.6)%.
6M FY2026 Operating Margin: (4.3)%.
Operating margin improved by approximately 179.3 percentage points between the two periods.
The company’s operating loss decreased from $2,135 million in the first half of FY2025 to $693 million in the first half of FY2026, an improvement of approximately 67.5%.
Although the six-month operating margin remained negative, the improvement indicates that Anthropic was moving much closer to overall operating profitability.
However, the distinction between quarterly and six-month profitability is important.
Anthropic reported an operating loss of $1,667 million during Q1 FY2026, followed by an operating profit of $974 million during Q2 FY2026.
As a result, the company remained unprofitable on an operating basis for the first six months of 2026 as a whole.
Furthermore, one quarter of positive operating income does not establish that Anthropic has achieved consistently sustainable profitability.
Operating expenses may fluctuate significantly depending on the timing of AI model training, research projects, computing commitments, and commercial expansion.
Nevertheless, the improvement demonstrates that Anthropic’s revenue growth was beginning to translate into substantially stronger operating results.
Plain English: In Q2 FY2025, Anthropic lost approximately $1.53 from operations for every dollar of revenue. By Q2 FY2026, it generated approximately 8.4 cents in operating profit for every dollar of revenue. This represents a major improvement in operating efficiency. However, the company still reported a small operating loss for the first half of 2026 overall.
2. Net Margin: Large Accounting Losses Continued
Anthropic’s net margin improved from (810.3)% in Q2 FY2025 to (573.9)% in Q2 FY2026, an improvement of approximately 236.4 percentage points.
For the six-month periods, net margin improved from (1,115.1)% in FY2025 to (495.8)% in FY2026, an improvement of approximately 619.3 percentage points.
Net margin measures net income or loss as a percentage of revenue after accounting for operating expenses, non-operating income and expenses, financing-related adjustments, and income taxes.
Despite the improvement in net margin percentages, Anthropic’s absolute net losses increased substantially.
The company’s reported net losses were:
Q2 FY2025: $(6,377) million.
Q2 FY2026: $(66,309) million.
6M FY2025: $(12,969) million.
6M FY2026: $(80,748) million.
At first glance, these figures appear inconsistent with the substantial improvement in operating profitability.
However, the difference primarily reflects the accounting treatment of Anthropic’s convertible financial instruments.
During the first half of FY2026, Anthropic recognized a $72,388 million loss from changes in the fair value of convertible notes, derivative liabilities, and SAFEs.
For comparison, the corresponding loss during the first half of FY2025 was $10,942 million.
These financial instruments are associated with Anthropic’s financing arrangements and may change in estimated accounting value as the company’s valuation and other financial assumptions change.
For example, an increase in the estimated value of Anthropic can increase the accounting value of certain outstanding convertible obligations, resulting in a reported loss.
Such fair-value adjustments can have an enormous effect on reported net income even when they do not require an equivalent cash payment during the reporting period.
Anthropic also recorded $8,757 million in other non-operating expenses during the first half of FY2026, primarily associated with financial instruments.
These financing-related adjustments help explain why the company reported a net loss of $80,748 million despite an operating loss of only $693 million.
The second-quarter results illustrate this distinction particularly clearly.
In Q2 FY2026, Anthropic generated $974 million in operating income but reported a $66,309 million net loss.
Consequently, its operating margin was positive at 8.4%, while its net margin remained deeply negative at (573.9)%.
This does not mean that Anthropic consumed nearly six dollars in operating cash for every dollar of revenue during the quarter.
Instead, a substantial portion of the reported loss reflected financing-related accounting adjustments rather than ordinary operating expenses.
However, these adjustments should not be ignored entirely.
Convertible financial instruments can have significant economic consequences, including changes in future financing obligations, the company’s capital structure, and potential shareholder dilution.
For investors evaluating Anthropic’s proposed IPO, it is therefore important to distinguish between operating profitability, reported net income, and actual cash flow.
Operating margin provides information about the profitability of the company’s ongoing business activities.
Net margin reflects the complete accounting result, including financing-related gains and losses.
Cash flow provides additional information about the actual cash generated or consumed during the reporting period.
These measures answer different financial questions and should be considered together.
Plain English: Anthropic’s net margin remained extremely negative, but much of the reported loss came from changes in the accounting value of financing instruments. These adjustments do not represent equivalent cash spending on AI operations. Investors should examine operating margin and cash flow alongside net margin to better understand the company’s underlying financial performance.
Overall Profitability Interpretation
Anthropic’s profitability ratios reveal a significant improvement in its underlying operating performance during the first half of 2026.
Operating margin improved dramatically, and the company achieved positive operating income during Q2 FY2026. This indicates that revenue was growing faster than operating expenses, allowing the company to generate an operating profit during the quarter.
However, net margin remained deeply negative because of substantial financing-related accounting adjustments associated with convertible financial instruments.
The contrast between these two ratios is particularly important for beginner investors.
Anthropic’s operating results suggest that its expanding commercial business was becoming more profitable, while its reported net losses continued to reflect the complexities of its pre-IPO financing structure.
The central takeaway is that Anthropic’s underlying operating profitability improved substantially during the first half of 2026, but the company had not yet demonstrated consistently positive operating earnings across multiple quarters.
Going forward, investors should monitor whether Anthropic can sustain positive operating margins while continuing to finance advanced AI research, computing infrastructure, and commercial expansion.
3) Balance Sheet Snapshot
Unit: $m. All figures are in USD millions. FY2025 Year-End represents December 31, 2025, and Q2 FY2026 represents June 30, 2026 (unaudited). Parentheses indicate negative amounts. The symbol “—” indicates items not separately disclosed.
Balance Sheet Snapshot
FY2025 Year-End
Q2 FY2026
Cash & Equivalents
$17,388
$112,389
Accounts Receivable
$909
$5,833
Inventory
—
—
Total Current Assets
$23,746
$137,980
Property, Plant & Equipment
$146
$1,187
Total Assets
$24,661
$147,358
Short-term Debt
—
—
Long-term Debt
$41,677
$108,566
Total Liabilities
$45,397
$117,467
Shareholders’ Equity
$(50,505)
$(127,619)
Plain English: Anthropic’s balance sheet expanded dramatically during the first half of 2026. Total assets increased from $24,661 million at the end of FY2025 to $147,358 million as of June 30, 2026, representing growth of approximately 497.5% in just six months.
The most significant development was the extraordinary increase in cash holdings, supported primarily by a large issuance of redeemable convertible preferred stock. However, the company also reported significantly higher convertible note liabilities and a larger shareholders’ deficit.
For beginner investors, the key distinction is that Anthropic’s expanding asset base was driven primarily by financing transactions rather than accumulated operating profits. Its substantial cash reserves improved liquidity, but its pre-IPO capital structure remained unusually complex.
1. Cash and Cash Equivalents: An Extraordinary Increase in Liquidity
Anthropic’s cash and cash equivalents increased from $17,388 million at the end of FY2025 to $112,389 million as of June 30, 2026.
This represents an increase of $95,001 million, or approximately 546.4%, in six months.
Cash and cash equivalents include cash deposits and highly liquid short-term financial instruments that can generally be converted into cash quickly with minimal risk of changes in value.
In addition to its cash holdings, Anthropic reported short-term investments of:
FY2025 Year-End: $2,889 million.
Q2 FY2026: $19,262 million.
Combining cash and cash equivalents with short-term investments, Anthropic’s financial resources increased from $20,277 million at the end of FY2025 to $131,651 million as of June 30, 2026.
This represents an increase of $111,374 million, or approximately 549.3%.
The expansion substantially strengthened Anthropic’s reported liquidity position, providing significant financial resources for operating expenses, research and development, and future infrastructure commitments.
However, understanding the source of this increase is essential.
During the first half of 2026, Anthropic raised $110,432 million in net cash proceeds from issuing redeemable convertible preferred stock.
These financing proceeds were the principal reason for the company’s extraordinary increase in cash holdings.
Although Anthropic’s operating cash flow also improved during the period, the increase in cash and financial investments was overwhelmingly supported by external capital rather than accumulated operating earnings.
This distinction matters because financing proceeds do not represent recurring revenue or profits generated by customers.
Furthermore, not all reported financial resources are necessarily available for unrestricted use.
Anthropic separately reported $2,529 million in restricted cash as of June 30, 2026, compared with $13 million at the end of FY2025.
Restricted cash represents funds subject to contractual or other limitations on their use. These amounts are not included in the cash and cash equivalents figures shown in the table.
The company also disclosed that certain arrangements could require additional cash restrictions in the future.
Consequently, the large cash balance should be evaluated alongside Anthropic’s contractual commitments, operating requirements, and financing obligations.
Plain English: Anthropic accumulated an exceptionally large amount of cash during the first half of 2026. However, most of that increase came from raising money from investors rather than profits generated by its AI services. The company had substantially more financial flexibility, but its cash reserves were not equivalent to earned profits or unrestricted funds available for shareholder distributions.
Accounts receivable increased from $909 million at the end of FY2025 to $5,833 million as of June 30, 2026.
This represents an increase of $4,924 million, or approximately 541.7%.
Accounts receivable represents amounts customers owe Anthropic for services already provided but not yet collected in cash.
The increase occurred alongside the company’s extraordinary revenue growth during the first half of 2026.
Anthropic’s six-month revenue increased from $1,163 million in the first half of FY2025 to $16,286 million in the first half of FY2026.
As the company expanded its commercial relationships and processed substantially more customer activity, the amount of outstanding receivables also increased.
This is consistent with the rapid growth of its business, particularly enterprise and consumption-based AI services.
However, an increase in accounts receivable is not automatically positive.
Revenue may be recognized before the corresponding cash is collected. Therefore, a rapidly increasing receivables balance can create additional working capital requirements.
In Anthropic’s case, the six-month cash flow statement recorded a $4,982 million cash outflow associated with changes in accounts receivable.
This indicates that expanding receivables absorbed a significant amount of operating cash during the period.
Investors should therefore monitor whether cash collections continue keeping pace with revenue growth.
Nevertheless, the increase in receivables alone does not establish that customers were paying late or that the company experienced collection problems.
Payment terms, billing arrangements, and the timing of revenue recognition can all affect the reported balance.
Plain English: Anthropic’s customers owed substantially more money at the end of June 2026 because its commercial business had expanded rapidly. However, revenue recorded before payment is collected can temporarily reduce operating cash flow. Continued growth in receivables makes customer collections an important financial indicator to monitor.
3. Total Current Assets: Stronger Short-Term Financial Resources
Total current assets increased from $23,746 million at the end of FY2025 to $137,980 million as of June 30, 2026.
This represents an increase of $114,234 million, or approximately 481.1%.
Current assets include resources generally expected to be converted into cash, collected, or used within one year or the company’s normal operating cycle.
For Anthropic, the largest current asset categories were cash and cash equivalents, short-term investments, and accounts receivable.
The increase was primarily driven by the company’s substantial financing proceeds and the expansion of its commercial operations.
At the same time, current liabilities increased from $2,817 million at the end of FY2025 to $7,220 million as of June 30, 2026.
Current liabilities represent obligations generally due within the following twelve months or operating cycle.
Although both current assets and current liabilities increased, current assets grew much more rapidly.
As a result, Anthropic’s calculated current ratio improved from approximately 843.0% at the end of FY2025 to 1,911.1% as of June 30, 2026.
The current ratio is calculated by dividing current assets by current liabilities.
Expressed as a multiple, Anthropic held approximately 19.1 times as many current assets as current liabilities at the end of June 2026.
This indicates a very substantial short-term liquidity buffer based on the reported balance sheet.
However, the current ratio has important limitations.
It does not directly measure the full cost of operating the business over future years, nor does it capture every long-term contractual commitment.
Anthropic has entered into substantial computing infrastructure arrangements that may require significant future payments.
Accordingly, the strength of its current ratio should not be interpreted as evidence that future financing or cash management will be unnecessary.
Plain English: Anthropic had substantially more short-term financial resources than short-term obligations as of June 30, 2026. This improved its ability to meet near-term payments. However, large cash reserves do not eliminate the financial risks associated with long-term infrastructure contracts and continuing AI development expenses.
4. Property, Plant, and Equipment: Physical Assets Expanded
Property, plant, and equipment increased from $146 million at the end of FY2025 to $1,187 million as of June 30, 2026.
This represents an increase of $1,041 million, or approximately 713.0%.
Property, plant, and equipment includes long-lived physical assets used in business operations, such as computing equipment, networking infrastructure, office improvements, and other facilities-related assets.
Anthropic’s capital expenditures also increased substantially.
During the first six months of 2026, the company spent $482 million on purchases of property and equipment, compared with $18 million during the first six months of 2025.
This indicates a significant expansion in reported physical infrastructure investment.
However, Anthropic’s property and equipment balance remains relatively small compared with its total assets and the economic scale of its computing operations.
As of June 30, 2026, property and equipment represented approximately 0.8% of total assets.
This does not mean the company’s AI services require limited computing infrastructure.
Anthropic relies heavily on computing capacity obtained through external infrastructure providers and contractual arrangements.
These resources may involve substantial operating expenses and future payment commitments without appearing as directly owned property and equipment on the balance sheet.
The company also reported operating lease right-of-use assets of $1,157 million as of June 30, 2026, compared with $126 million at the end of FY2025.
A right-of-use asset represents the company’s contractual right to use a leased asset over an agreed period.
The increase reflects a larger balance of assets recognized under lease accounting.
In addition, goodwill increased from $154 million to $576 million, reflecting the effects of business acquisitions and related accounting adjustments.
Plain English: Anthropic substantially increased its reported physical assets during the first half of 2026. However, directly owned equipment represents only a small portion of its total assets. Much of the infrastructure needed to develop and operate its AI models is obtained through outside providers, so reported capital expenditures do not capture the full scale of its computing requirements.
5. Total Assets: Expansion Driven Primarily by Financing
Anthropic’s total assets increased from $24,661 million at the end of FY2025 to $147,358 million as of June 30, 2026.
The increase of $122,697 million represents growth of approximately 497.5% over six months.
This was an extraordinary expansion in the company’s reported financial resources.
However, the composition of the asset increase is more important than the headline growth percentage.
The largest contributors were:
Cash and Cash Equivalents: Increased by $95,001 million.
Short-Term Investments: Increased by $16,373 million.
Accounts Receivable: Increased by $4,924 million.
Property, Plant, and Equipment: Increased by $1,041 million.
Restricted Cash: Increased by $2,516 million.
Together, these changes show that the expansion was driven overwhelmingly by financial assets, particularly cash and investments.
The increase in cash and investments was primarily supported by external financing.
Therefore, total asset growth should not be interpreted as equivalent growth in accumulated operating profits or directly owned computing infrastructure.
It also should not be treated as a measure of the company’s market valuation.
Market valuation depends on the value investors assign to the company’s future business, adjusted for the rights and obligations associated with its capital structure.
Plain English: Anthropic’s total assets increased almost sixfold in just six months. However, most of that increase came from cash and investments accumulated through major financing transactions. The balance sheet became much larger, but the increase alone does not establish that the company had become more profitable.
6. Convertible Notes and Total Liabilities: Fair-Value Adjustments Were Critical
Anthropic’s reported long-term convertible notes payable increased from $41,677 million at the end of FY2025 to $108,566 million as of June 30, 2026.
This represents an increase of $66,889 million, or approximately 160.5%.
Convertible notes are financing instruments that may be converted into shares of the company’s stock under specified contractual conditions.
Unlike conventional debt measured primarily at amortized cost, Anthropic’s convertible notes are measured at fair value.
This means the reported liability balance can increase or decrease based on changes in the instruments’ estimated market value.
During the first half of 2026, Anthropic recognized a $72,388 million loss from changes in the fair value of convertible notes, derivative liabilities, and SAFEs.
The convertible note balance also decreased by $5,499 million as certain notes were converted into redeemable convertible preferred stock.
Consequently, the increase in the balance sheet’s convertible note liability should not be interpreted as $66,889 million in newly borrowed cash.
In fact, Anthropic’s cash flow statement reported no new cash proceeds from convertible note issuance during the first half of 2026.
The increase in the liability balance primarily reflected the accounting revaluation of outstanding instruments, partly offset by conversions.
Meanwhile, total liabilities increased from $45,397 million at the end of FY2025 to $117,467 million as of June 30, 2026.
This represents an increase of $72,070 million, or approximately 158.8%.
The increase was driven primarily by the higher carrying value of convertible notes, although operating liabilities also increased as the business expanded.
For example, accounts payable, accrued expenses, and other current liabilities increased from $2,511 million to $5,422 million.
Deferred revenue increased from $269 million to $1,650 million.
Deferred revenue represents amounts received or billed before the company has fully delivered the related products or services.
The increase is consistent with expanding commercial activity, but deferred revenue should not be interpreted as the company’s complete future revenue backlog.
Plain English: Anthropic’s reported liabilities increased dramatically, especially convertible notes payable. However, much of the increase came from accounting changes in the estimated value of existing financing instruments rather than additional cash borrowing. Investors should distinguish between a liability’s reported accounting value and the cash originally received when the financing was issued.
7. Redeemable Preferred Stock and Shareholders’ Deficit: An Unusual Pre-IPO Capital Structure
Anthropic’s shareholders’ equity remained deeply negative during the first half of 2026.
The reported shareholders’ deficit increased from $(50,505) million at the end of FY2025 to $(127,619) million as of June 30, 2026.
This represents a deterioration of $77,114 million.
The principal reason was the company’s reported net loss of $80,748 million during the first half of 2026, which included substantial noncash financing-related accounting losses.
Anthropic’s accumulated deficit increased from $(52,154) million to $(132,902) million over the same period.
An accumulated deficit represents the cumulative losses recognized by a company over its operating history, reduced by any accumulated earnings.
However, the interpretation of Anthropic’s shareholders’ equity requires particular caution because the company has issued substantial amounts of redeemable convertible preferred stock.
Anthropic reported:
Redeemable Convertible Preferred Stock at FY2025 Year-End: $29,769 million.
Redeemable Convertible Preferred Stock at June 30, 2026: $157,510 million.
These preferred shares are classified separately from ordinary shareholders’ equity and total liabilities in the consolidated balance sheet.
They represent financing instruments with contractual rights that differ from those attached to ordinary common shares.
This separate presentation explains why Anthropic’s balance sheet cannot be understood simply by comparing total assets with total liabilities and ordinary shareholders’ equity.
As of June 30, 2026, its financial position consisted of:
These figures reconcile because total liabilities plus redeemable convertible preferred stock plus shareholders’ deficit equal total assets.
The negative shareholders’ equity balance therefore does not mean Anthropic had exhausted its cash resources.
Instead, it reflects the combination of accumulated accounting losses and the classification of substantial financing instruments outside permanent shareholders’ equity.
The distinction is especially relevant to the company’s proposed IPO.
Anthropic’s S-1 describes the expected conversion of certain preferred shares and convertible notes as part of the offering process.
These transactions could substantially change the company’s reported capital structure without necessarily producing an equivalent change in operating cash flow.
However, the final post-IPO capital structure depends on the offering terms and related transactions.
Plain English: Anthropic’s negative shareholders’ equity became much larger during the first half of 2026, even as its cash holdings increased dramatically. This apparent contradiction results from large accounting losses and an unusual financing structure involving preferred shares and convertible notes. Negative book equity does not automatically mean the company lacks cash, but it does make conventional valuation and leverage ratios more difficult to interpret.
Overall Balance Sheet Interpretation
Anthropic’s balance sheet as of June 30, 2026, demonstrates an extraordinary expansion in financial resources.
Cash and short-term investments reached $131,651 million, providing a substantial liquidity buffer to support the company’s operations and future development plans.
However, this financial expansion was driven primarily by major external financing transactions rather than accumulated operating profits.
At the same time, Anthropic reported significantly higher convertible note liabilities, a substantial increase in redeemable convertible preferred stock, and a deeper shareholders’ deficit.
These developments highlight the importance of separating liquidity, operating performance, and financing-related accounting effects.
The company possessed considerable financial resources, but it also faced substantial computing infrastructure commitments, ongoing research requirements, and a complex pre-IPO capital structure.
For beginner investors, the central takeaway is that Anthropic entered the second half of 2026 with a dramatically stronger cash position, but its balance sheet should not be interpreted as evidence of accumulated profitability or a conventional low-leverage financial structure.
The most important factors to monitor are how effectively Anthropic converts its expanding revenue into sustainable operating cash flow, how quickly its financial resources are consumed or committed, and how its capital structure changes through the IPO process.
4) Cash Flow Summary
Unit: $m. All figures are in USD millions. 6M represents the six months ended June 30. Parentheses indicate cash outflows or negative amounts. Ending cash balances include restricted cash.
Cash Flow Summary
6M FY2025
6M FY2026
Operating Cash Flow
$(1,513)
$3,929
Investing Cash Flow
$(49)
$(17,181)
Financing Cash Flow
$4,749
$110,774
Net Change in Cash
$3,188
$97,517
Ending Cash Balance
$6,342
$114,918
Plain English: Anthropic’s cash flow performance changed dramatically during the first half of 2026. Most importantly, the company transitioned from negative to positive operating cash flow, indicating that its business activities generated more cash than they consumed during the six-month period.
At the same time, Anthropic raised an extraordinary amount of external financing, allowing its cash reserves to increase substantially despite significant investing cash outflows.
However, the three categories of cash flow tell different stories. Operating cash flow reflects cash generated by business activities, investing cash flow reflects purchases and sales of assets and investments, and financing cash flow reflects transactions with investors and lenders.
Understanding these distinctions is especially important when evaluating Anthropic because its reported net losses, actual cash generation, and financing activities differ significantly.
1. Operating Cash Flow: A Transition to Positive Cash Generation
Anthropic’s operating cash flow improved from $(1,513) million in the first half of FY2025 to $3,929 million in the first half of FY2026.
This represents a positive change of $5,442 million in one year.
Operating cash flow measures the cash generated or consumed by ordinary business activities, including customer payments, operating expenses, and changes in working capital.
In the first half of FY2025, Anthropic consumed $1,513 million through operating activities. By the first half of FY2026, the company generated $3,929 million in positive operating cash flow.
This was an important development because the company’s revenue increased from $1,163 million to $16,286 million over the same periods.
As commercial adoption expanded, Anthropic generated substantially more revenue from its AI products and services.
The relationship between revenue and operating cash flow also improved significantly.
6M FY2025 Operating Cash Flow Margin: (130.1)%.
6M FY2026 Operating Cash Flow Margin: 24.1%.
Operating cash flow margin measures operating cash flow as a percentage of revenue.
A positive margin indicates that the company generated cash from operating activities during the period, while a negative margin indicates that operating activities consumed cash.
However, operating cash flow is not the same as operating income or net income. It is affected by noncash accounting adjustments and changes in the timing of customer collections and payments to suppliers.
Consequently, positive operating cash flow should not automatically be interpreted as evidence of sustainable profitability.
Plain English: Anthropic’s operations consumed cash during the first half of 2025 but generated approximately $3,929 million during the first half of 2026. This was an important improvement in cash generation. However, investors must examine the factors behind the positive cash flow to determine whether it can be sustained.
2. Why Operating Cash Flow Was Positive Despite an Enormous Net Loss
One of the most unusual aspects of Anthropic’s financial statements is the difference between reported net income and operating cash flow.
During the first half of FY2026, Anthropic reported:
Net Loss: $(80,748) million.
Operating Cash Flow: $3,929 million.
Although the company reported a net loss exceeding $80,000 million, its operating activities generated positive cash flow.
The primary explanation was the recognition of substantial noncash accounting expenses.
Noncash expenses reduce reported earnings without requiring an equivalent cash payment during the reporting period.
Anthropic’s cash flow statement included several major adjustments to reconcile reported net loss with operating cash flow:
Fair-Value Changes in Convertible Notes, Derivative Liabilities, and SAFEs: $72,388 million.
Recognition and Fair-Value Changes in Other Financial Instruments: $8,720 million.
Stock-Based Compensation: $2,185 million.
Depreciation and Amortization: $97 million.
Donation Matching Program Liability: $126 million.
The two largest adjustments were associated with financial instruments rather than the ordinary costs of developing or operating Claude.
Anthropic’s convertible financing instruments are subject to fair-value accounting. Changes in their estimated value can produce substantial reported gains or losses.
During the first half of 2026, increases in the estimated value of certain outstanding financial obligations generated large accounting losses.
Because these adjustments did not require equivalent operating cash payments during the period, they were added back in the calculation of operating cash flow.
Stock-based compensation also contributed to the difference.
Stock-based compensation is an accounting expense associated with equity awards granted to employees. It generally does not require the same immediate cash payment as ordinary salaries.
However, equity compensation can create shareholder dilution, meaning its economic cost should not be ignored simply because it is noncash.
These adjustments explain much of the difference between Anthropic’s reported net loss and its actual operating cash flow.
Nevertheless, they do not mean that all of Anthropic’s underlying operating expenses were noncash.
The company continued spending substantial amounts on computing infrastructure, personnel, and commercial activities.
Plain English: Anthropic reported a net loss of $80,748 million but generated $3,929 million in operating cash flow. Most of the difference came from accounting adjustments related to financing instruments rather than cash spent on its AI operations. This makes cash flow especially important when analyzing the company’s financial performance.
3. Working Capital: Customer Collections and Payment Timing Matter
Operating cash flow is also affected by changes in working capital.
Working capital includes short-term operating assets and liabilities, such as accounts receivable, prepaid expenses, accounts payable, and deferred revenue.
These accounts affect the timing of cash receipts and payments even when the associated revenue or expense has already been recognized in the income statement.
During the first half of FY2026, Anthropic reported the following major working capital changes:
Accounts Receivable: $(4,982) million.
Prepaid Expenses and Other Assets: $2,094 million.
Accounts Payable, Accrued Expenses, and Other Liabilities: $2,679 million.
Deferred Revenue: $1,380 million.
The increase in accounts receivable consumed substantial operating cash because revenue was recognized before all corresponding customer payments had been collected.
This is consistent with the rapid expansion of Anthropic’s commercial activities.
However, the increase in receivables also highlights the importance of monitoring whether customer collections continue keeping pace with revenue growth.
At the same time, changes in prepaid expenses and other assets contributed positively to operating cash flow.
An increase in accounts payable and accrued liabilities also supported cash flow because some obligations had not yet been settled in cash.
Deferred revenue provided another positive contribution.
Deferred revenue represents amounts received or billed before the company has fully delivered the related products or services.
When customers make payments in advance, cash flow can improve before the corresponding revenue is fully recognized.
These working capital changes contributed to Anthropic’s positive operating cash flow during the period.
However, investors should recognize that working capital movements can fluctuate significantly from one reporting period to another.
For example, an increase in unpaid operating liabilities can temporarily improve cash flow, but those obligations may require cash payments in later periods.
Similarly, an increase in deferred revenue can support cash flow today while creating obligations to provide services in the future.
Therefore, positive operating cash flow supported by working capital changes should not automatically be interpreted as permanently recurring cash generation.
Plain English: Anthropic’s expanding business created both cash inflows and cash outflows from working capital. Uncollected customer invoices reduced cash flow, while changes in liabilities, prepaid expenses, and deferred revenue helped offset that effect. These timing differences can materially affect cash flow, so investors should monitor them over multiple reporting periods.
4. Investing Cash Flow: Financial Investments Were the Largest Outflow
Anthropic’s investing cash flow decreased from $(49) million in the first half of FY2025 to $(17,181) million in the first half of FY2026.
This represents an increase of $17,132 million in investing cash outflows.
At first glance, such a large outflow might suggest that Anthropic spent more than $17,000 million building computing infrastructure or purchasing equipment.
However, the cash flow statement reveals a different explanation.
The largest investing activity was the purchase of short-term financial investments.
During the first half of FY2026, Anthropic reported:
Purchases of Short-Term Investments: $(18,972) million.
Sales and Maturities of Short-Term Investments: $2,610 million.
Purchases of Property and Equipment: $(482) million.
Other Investing Activities: $(337) million.
Net purchases of short-term investments amounted to $16,362 million.
These transactions accounted for approximately 95.2% of total net investing cash outflows during the period.
Short-term investments include financial assets that a company purchases to manage liquidity and earn investment income.
Purchasing these assets reduces cash reported in the investing section of the cash flow statement, but the company generally receives another financial asset in exchange.
Therefore, these purchases should not be interpreted in the same way as spending on research, physical infrastructure, or other resources consumed in business operations.
Anthropic’s actual cash purchases of property and equipment were substantially smaller, totaling $482 million during the first half of FY2026.
This compared with $18 million during the first half of FY2025.
The increase indicates that Anthropic expanded its investment in physical assets.
However, its reported capital expenditures do not capture the full economic scale of the computing infrastructure required by its AI operations.
Anthropic relies extensively on computing capacity obtained through external cloud and infrastructure providers.
Payments for certain computing services are recognized within operating activities rather than as purchases of property and equipment.
In addition, future computing commitments may create significant contractual obligations without appearing as current-period capital expenditures.
Plain English: Anthropic’s investing cash outflow increased dramatically, but most of the money was used to purchase short-term financial investments rather than build infrastructure. This means the negative investing cash flow should not be interpreted entirely as cash consumed by expansion. The company’s computing costs and future infrastructure commitments must also be evaluated separately.
5. Free Cash Flow: A Significant Improvement
Free cash flow, commonly abbreviated as FCF, is calculated as operating cash flow minus capital expenditures.
It measures the cash remaining after a company funds its operating activities and purchases long-lived assets.
Using cash flow from operating activities less purchases of property and equipment, Anthropic’s free cash flow was:
6M FY2025: $(1,531) million.
6M FY2026: $3,447 million.
Free cash flow improved by $4,978 million year over year.
This represents a significant change in Anthropic’s financial performance.
During the first half of FY2025, the company’s operations and capital expenditures consumed cash.
By the first half of FY2026, Anthropic generated positive free cash flow after deducting reported property and equipment purchases.
Its calculated free cash flow margin improved from approximately (131.6)% to 21.2%.
Free cash flow margin measures free cash flow as a percentage of revenue.
However, several important limitations apply.
First, the positive free cash flow figure was affected by noncash accounting adjustments and working capital changes, as discussed above.
Second, the calculation deducts reported purchases of property and equipment but does not separately subtract all future contractual payments associated with outsourced computing infrastructure.
Third, Anthropic continued making substantial investments in AI model development, and the timing of computing expenses may change between periods.
Consequently, positive free cash flow during the first half of 2026 does not guarantee that the company will continue generating positive free cash flow in future periods.
Nevertheless, the improvement is meaningful because it demonstrates that Anthropic generated positive operating cash flow sufficient to cover its reported capital expenditures during the period.
Plain English: Anthropic moved from negative to positive free cash flow during the first half of 2026. This means its operating cash flow exceeded the amount spent purchasing property and equipment. However, the calculation does not capture every long-term computing commitment, so investors should not assume that all future infrastructure requirements are already funded.
6. Financing Cash Flow: Major External Capital Raising
Anthropic’s financing cash flow increased from $4,749 million in the first half of FY2025 to $110,774 million in the first half of FY2026.
This represents an increase of $106,025 million.
Financing cash flow measures cash received from investors and lenders, along with cash used to repay financing obligations or return capital to shareholders.
The extraordinary increase in financing inflows was primarily driven by the issuance of redeemable convertible preferred stock.
During the first half of FY2026, Anthropic reported:
Net Proceeds from Redeemable Convertible Preferred Stock: $110,432 million.
Proceeds from Common Stock Issuance: $355 million.
Principal Payments for Finance Leases: $(4) million.
Payment of Deferred Offering Costs: $(9) million.
These transactions resulted in $110,774 million in net financing cash inflows.
Preferred stock issuance accounted for approximately 99.7% of total net financing cash flow.
This indicates that the company’s extraordinary financing inflows were overwhelmingly concentrated in equity-related capital raising.
For comparison, Anthropic’s financing cash flow during the first half of FY2025 included:
Net Proceeds from Preferred Stock Issuance: $3,549 million.
Convertible Notes Issued for Cash: $1,300 million.
Proceeds from Common Stock Issuance: $49 million.
Common Stock Repurchases: $(149) million.
The difference between the two periods reflects a substantial expansion in the scale of Anthropic’s financing activities.
These transactions also explain much of the increase in cash and cash equivalents reported on the balance sheet.
However, external financing should not be confused with operating cash flow.
Money raised by issuing preferred shares increases available cash, but it does not represent revenue earned from customers or profits generated by the business.
Furthermore, preferred shares and other convertible securities may provide investors with contractual rights that differ from those of ordinary common shareholders.
Depending on their terms, these instruments may affect future ownership percentages, shareholder dilution, and the company’s capital structure.
Anthropic’s proposed IPO introduces additional complexity because certain outstanding financing instruments are expected to convert or otherwise be affected by the offering.
Plain English: Anthropic raised an extraordinary amount of cash from investors during the first half of 2026. This significantly strengthened its financial resources. However, the money came primarily from issuing preferred shares rather than selling AI services, and the resulting financing arrangements may affect future shareholders.
7. Ending Cash Balance: Substantial Liquidity with Important Restrictions
Anthropic’s ending cash balance increased from $6,342 million as of June 30, 2025, to $114,918 million as of June 30, 2026.
This represents an increase of $108,576 million between the two reporting dates.
The cash flow statement’s ending balance includes cash, cash equivalents, and restricted cash.
Restricted cash represents funds subject to contractual or other limitations on their use.
As of June 30, 2026, the ending cash balance consisted of:
Cash and Cash Equivalents: $112,389 million.
Restricted Cash: $2,529 million.
Total Cash, Cash Equivalents, and Restricted Cash: $114,918 million.
This explains why the ending balance in the cash flow statement is higher than the cash and cash equivalents figure reported separately on the balance sheet.
The increase during the first half of FY2026 can be reconciled through the major cash flow categories:
Operating Cash Flow: $3,929 million.
Investing Cash Flow: $(17,181) million.
Financing Cash Flow: $110,774 million.
Foreign Exchange Effect: $(5) million.
Together, these amounts produce a $97,517 million increase in cash, cash equivalents, and restricted cash during the first half of FY2026.
The company’s cash position expanded significantly because financing inflows substantially exceeded the combined investing cash outflows.
Positive operating cash flow also contributed to the increase, although external financing remained the dominant factor.
Anthropic’s large liquidity position provides financial flexibility to support research, commercial expansion, and computing infrastructure requirements.
However, the company has substantial contractual commitments associated with future computing capacity.
The full cost of these commitments is not captured by the current-period cash flow statement alone.
Moreover, restricted cash cannot necessarily be used for general corporate purposes.
As a result, investors should distinguish between reported cash balances, unrestricted liquidity, and the cash required to satisfy future operating and infrastructure obligations.
Plain English: Anthropic ended June 2026 with substantial cash resources, supported primarily by its large capital raise. This provided considerable financial flexibility. However, some cash was restricted, and the company faced significant future infrastructure commitments. A large cash balance does not eliminate the need to manage long-term spending carefully.
Overall Cash Flow Interpretation
Anthropic’s cash flow statements for the first half of 2026 reveal three major financial developments.
First, operating cash flow turned positive, increasing from $(1,513) million in the first half of FY2025 to $3,929 million in the first half of FY2026. This was an important improvement, although noncash adjustments and working capital movements were significant contributors.
Second, free cash flow also became positive, reaching $3,447 million after deducting reported capital expenditures. However, this measure does not fully capture the company’s long-term computing infrastructure commitments.
Third, external financing transformed Anthropic’s liquidity position. Net financing cash inflows of $110,774 million allowed the company to increase its financial resources dramatically, even while making substantial purchases of short-term investments.
The most important distinction is between cash generated by the business and cash raised from investors.
Anthropic demonstrated an improvement in operating cash generation during the first half of 2026, but the extraordinary expansion of its cash reserves was primarily financed through preferred stock issuance.
For beginner investors, the central takeaway is that Anthropic’s underlying cash flow performance improved substantially, while external financing provided the overwhelming majority of the increase in its cash reserves.
Going forward, investors should monitor whether positive operating cash flow and free cash flow can be sustained, how working capital evolves as customer revenue expands, and whether the company’s cash resources remain sufficient to support its substantial AI development and infrastructure commitments.
5) First Half 2026 Takeaways
Anthropic’s financial results for the first half of 2026 reveal a significant improvement in its underlying business performance. Revenue expanded dramatically, operating profitability improved, and the company generated positive operating cash flow and free cash flow.
At the same time, Anthropic reported exceptionally large accounting losses and underwent a major expansion of its capital structure through external financing.
Revenue Growth Accelerated: Revenue increased from $1,163 million in the first half of FY2025 to $16,286 million in the first half of FY2026, representing year-over-year growth of approximately 1,300.3%. In Q2 alone, revenue reached $11,555 million, compared with $787 million one year earlier. This expansion reflected rapidly growing commercial demand for Anthropic’s AI products and services.
Quarterly Operating Profitability Turned Positive: Anthropic reported $974 million in operating income during Q2 FY2026, compared with an operating loss of $1,200 million in Q2 FY2025. Operating margin improved from (152.5)% to 8.4%. This demonstrated that quarterly revenue had grown sufficiently to exceed reported operating expenses.
Six-Month Operating Losses Narrowed Significantly: Operating loss decreased from $2,135 million in the first half of FY2025 to $693 million in the first half of FY2026, an improvement of approximately 67.5%. Operating margin improved from (183.6)% to (4.3)%. Despite this progress, the company had not yet achieved positive operating income for the entire six-month period.
Net Losses Remained Exceptionally Large: Anthropic reported a net loss of $80,748 million in the first half of FY2026, compared with $12,969 million in the same period of FY2025. However, the FY2026 result included a $72,388 million loss from changes in the fair value of convertible notes, derivative liabilities, and SAFEs. These financing-related accounting adjustments were a major reason reported net income differed substantially from operating performance.
Operating Cash Flow Turned Positive: Operating cash flow improved from $(1,513) million in the first half of FY2025 to $3,929 million in the first half of FY2026. This represented a positive change of $5,442 million. However, the result was affected by noncash adjustments and changes in working capital, so continued positive cash generation is not guaranteed.
Free Cash Flow Also Became Positive: Free cash flow, calculated as operating cash flow minus purchases of property and equipment, improved from $(1,531) million to $3,447 million. This indicates that operating cash flow exceeded reported capital expenditures during the first half of 2026. Nevertheless, the calculation does not fully capture Anthropic’s future contractual obligations for externally supplied computing capacity.
Financing Activities Expanded Dramatically: Financing cash inflows reached $110,774 million in the first half of FY2026, compared with $4,749 million in the prior-year period. The increase was primarily driven by $110,432 million in net proceeds from redeemable convertible preferred stock issuance. External financing remained the dominant source of Anthropic’s expanding financial resources.
Cash Reserves Increased Substantially: Cash and cash equivalents increased from $17,388 million at FY2025 year-end to $112,389 million as of June 30, 2026. Including $19,262 million in short-term investments, Anthropic held $131,651 million in cash, cash equivalents, and short-term investments. However, these financial resources should be considered alongside restricted cash, contractual commitments, and future operating requirements.
The Balance Sheet Became More Complex: Total assets increased to $147,358 million, while total liabilities reached $117,467 million. Redeemable convertible preferred stock increased to $157,510 million, and the shareholders’ deficit widened to $(127,619) million. These changes reflected major financing transactions, accumulated accounting losses, and the unusual presentation of Anthropic’s pre-IPO financial instruments.
Computing Infrastructure Remained a Major Financial Consideration: Anthropic continued investing in research, model development, and the infrastructure required to operate advanced AI systems. Because substantial computing capacity is obtained through external providers, investors should not evaluate future infrastructure costs solely through reported property and equipment purchases.
Overall First Half 2026 Interpretation
Anthropic’s first-half 2026 results marked an important change in its financial trajectory.
During FY2023–FY2025, the company demonstrated exceptional revenue growth but continued reporting substantial operating losses and negative operating cash flow.
By the first half of 2026, its rapidly expanding revenue base had begun to support stronger financial performance. The company achieved positive quarterly operating income, positive six-month operating cash flow, and positive free cash flow.
These developments suggest that Anthropic’s commercial growth was increasingly translating into improved operating economics.
However, several important uncertainties remained.
The company had not yet demonstrated consistently positive operating income across multiple quarters. Its reported net losses continued to be affected by large financing-related accounting adjustments, and its extraordinary cash reserves were primarily supported by external capital raising.
Furthermore, maintaining and expanding frontier AI capabilities requires substantial ongoing spending on research, personnel, and computing infrastructure.
For beginner investors, the most important development was the improvement in Anthropic’s underlying operating performance, not simply the enormous increase in its reported cash balance.
The central question is whether the company can sustain positive operating profitability and cash generation while continuing to fund the investments required to remain competitive in advanced artificial intelligence.
2.3 Beginner Takeaways
Anthropic’s historical and interim financial statements provide a clearer picture when revenue growth, profitability, cash flow, and financing are evaluated separately.
Growth: Anthropic expanded from $41 million in annual revenue in FY2023 to $4,585 million in FY2025, followed by $16,286 million in revenue during the first half of 2026 alone. Commercial demand grew rapidly across its AI services.
Profitability: Operating losses remained substantial through FY2025, but operating performance improved significantly during 2026. The company generated $974 million in operating income in Q2 FY2026, although its first-half operating result remained negative.
Reported Earnings: Enormous noncash fair-value adjustments associated with convertible financial instruments substantially increased reported net losses. These accounting effects must be distinguished from the profitability of Anthropic’s ongoing operations.
Cash Generation: Operating cash flow and free cash flow both turned positive during the first half of 2026. However, working capital movements, noncash adjustments, and future infrastructure commitments remain important when assessing the sustainability of cash generation.
Financial Position: Anthropic entered the second half of 2026 with substantial cash and short-term investments, primarily accumulated through external financing. Its liquidity position was strong, but its capital structure remained complex, with significant convertible obligations and a large shareholders’ deficit.
Final Financial Highlights Takeaway
Anthropic’s financial statements describe a company that experienced extraordinary commercial growth and made meaningful progress toward operating profitability.
The most encouraging development was its transition from substantial operating cash consumption to positive operating cash flow, alongside its achievement of positive quarterly operating income during Q2 FY2026.
However, the company remained exposed to significant financial risks, including high research and infrastructure requirements, complex convertible financing arrangements, and uncertainty over the sustainability of recent profitability improvements.
For long-term investors, the central financial question is whether Anthropic can convert its rapidly expanding AI revenue into consistently positive operating earnings and durable free cash flow while funding the continued development of increasingly advanced AI models.
3. Risks
Editorial Note: In order to enhance readability, we have omitted broad, market-wide risks that generally affect all companies. The following discussion focuses on risks specific to Anthropic and the artificial intelligence industry, as disclosed in Anthropic’s S-1 filing.
Anthropic’s S-1 identifies several important risks associated with developing advanced AI models, securing computing infrastructure, competing in the rapidly evolving AI industry, and maintaining safe and reliable services. The company also faces risks arising from its financing arrangements, intellectual property disputes, and unusual corporate governance structure.
3.1. Frontier AI Development and Technological Competition
Anthropic’s success depends heavily on its ability to continue developing AI models that remain competitive with the most advanced systems available.
The company must continuously improve Claude’s capabilities, reliability, efficiency, and safety while responding to rapid technological developments.
Anthropic identifies several challenges:
Rapid technological change: Competitors may introduce more capable or efficient AI models, reducing the attractiveness of Anthropic’s offerings.
Uncertain research outcomes: Substantial investments in model development may not consistently produce meaningful technological improvements.
Increasing development complexity: Training more advanced AI models requires substantial computing resources, specialized expertise, and reliable training data.
AI-assisted research: Anthropic increasingly uses its own AI systems to develop subsequent models. Errors or weaknesses in existing systems could affect later generations of models.
Safety and performance requirements: Improvements in model capabilities must be accompanied by adequate safety, security, and reliability measures.
Anthropic describes an increasingly important development process in which AI systems assist with coding, research, and the development of subsequent AI models.
If competitors become more effective at using AI to accelerate their own research, Anthropic could fall behind. Conversely, errors introduced through AI-assisted development could carry forward into future models.
Plain English: If Anthropic cannot keep improving Claude as quickly and effectively as competitors improve their AI models, customers may switch to competing services.
3.2. Computing Capacity and Infrastructure Availability
Developing and operating advanced AI models requires enormous amounts of specialized computing power.
Anthropic must obtain sufficient computing capacity for several activities:
Training new AI models.
Improving existing models through additional training.
Processing customer requests, known as inference.
Supporting internal research and AI-assisted development.
Operating increasingly complex AI agents that perform extended, multi-step tasks.
The company warns that demand for computing capacity has historically exceeded the resources available to it.
Anthropic depends on third-party cloud providers and infrastructure partners to obtain much of this capacity. Even when contractual arrangements exist, providers may experience shortages, operational delays, or competing demands for their computing resources.
Anthropic must also decide how to allocate available computing capacity between model development and customer services.
If too much capacity is allocated to training, customer service quality could deteriorate. If too little capacity is allocated to research, the development of future models could be delayed.
These challenges may become more significant as Anthropic expands its agentic AI offerings. AI agents can require repeated model interactions, tool usage, and extended processing sessions, potentially consuming considerably more computing resources than ordinary chatbot interactions.
Plain English: If Anthropic cannot obtain enough computing power or allocate it effectively, it may struggle to develop new AI models while simultaneously meeting customer demand.
3.3. Long-Term Computing Commitments and Forecasting Risk
Anthropic enters into substantial computing and infrastructure agreements well before the related resources become available.
These arrangements may include minimum purchase commitments, which require the company to purchase a specified amount of computing capacity, and take-or-pay provisions, which can require payment even when the contracted capacity is not fully used.
Because AI technology and customer demand change rapidly, Anthropic faces considerable uncertainty when estimating future computing requirements.
The company identifies two major risks:
Insufficient capacity: If Anthropic underestimates future computing requirements, it may lack the resources necessary to train models or serve customers.
Excess capacity: If Anthropic overestimates demand, it may be required to pay for computing resources that remain underutilized.
Anthropic states that it has previously underestimated its computing needs, resulting in capacity constraints that affected product development and deployment.
The company is also transitioning toward a broader infrastructure model that includes directly leased AI accelerators hosted in dedicated data center facilities.
This introduces additional uncertainties involving construction schedules, equipment installation, facility operations, and the integration of new infrastructure with existing cloud resources.
Furthermore, computing hardware can become less economically attractive as newer chip architectures and technologies emerge. Long-term commitments may limit Anthropic’s ability to switch to more efficient alternatives.
The company also warns that failure to satisfy certain contractual obligations could trigger defaults under other financing agreements, potentially accelerating repayment obligations or disrupting access to critical infrastructure.
Plain English: Anthropic must commit large amounts of money to computing resources before knowing exactly how much capacity it will need. Incorrect forecasts could leave it paying for unused infrastructure or struggling with insufficient capacity.
3.4. AI Chips, Electricity, and Data Center Supply Constraints
Anthropic’s computing infrastructure depends on several critical resources that may be difficult to obtain in sufficient quantities.
These include:
Advanced AI chips: Specialized processors and high-performance memory used for training and running AI models.
Semiconductor manufacturing: Advanced chip production capacity concentrated among a limited number of manufacturers, including Taiwan Semiconductor Manufacturing Company.
Electricity: Large quantities of reliable electrical power required to operate AI data centers.
Data center infrastructure: Facilities, networking equipment, cooling systems, and other equipment necessary to support large-scale computing.
Water and supporting utilities: Resources required for certain data center cooling and operating systems.
Anthropic warns that shortages of advanced chips, limited manufacturing capacity, and delays in data center construction could restrict the amount of computing power it can obtain.
Geographic concentration in advanced semiconductor manufacturing creates additional exposure to disruptions affecting Taiwan and other important manufacturing locations.
Electrical power availability is another significant constraint. New power generation facilities, transmission infrastructure, and grid connections can require years to complete.
Anthropic has also committed to covering certain electricity cost increases affecting consumers that arise from its data centers, potentially increasing the financial obligations associated with infrastructure expansion.
Higher equipment prices, utility costs, or delays in obtaining required infrastructure could increase expenses and slow model development or commercial deployment.
Plain English: Even if Anthropic has enough money to purchase computing power, shortages of chips, electricity, or data center infrastructure could prevent it from expanding as planned.
3.5. Dependence on Major Cloud and Infrastructure Partners
Anthropic relies on a limited number of major cloud service providers and infrastructure partners to obtain computing resources and distribute its AI models.
These relationships are unusually complex because certain partners may simultaneously be:
Investors in Anthropic.
Suppliers of computing capacity.
Distribution channels for Claude.
Customers of Anthropic’s AI services.
Competitors developing their own AI models.
Anthropic specifically identifies the potential for conflicting interests among these different relationships.
For example, a cloud provider that develops competing AI models may also control computing capacity that Anthropic depends on.
Cloud providers may have to allocate limited infrastructure among their own products, Anthropic, and other customers.
The company also relies on its partners to meet delivery schedules, maintain infrastructure, and fulfill contractual commitments.
Disruptions, unfavorable changes to commercial terms, or disagreements with major partners could affect Anthropic’s ability to train models, serve customers, or distribute its products.
Plain English: Some of Anthropic’s most important suppliers and business partners also compete with it. Changes in those relationships could affect both its access to computing power and its ability to reach customers.
3.6. Competition from Proprietary and Open-Weight AI Models
Anthropic operates in an intensely competitive market that includes large technology companies, specialized AI developers, and organizations releasing open-weight or open-source models.
The company identifies Google and OpenAI as major competitors in developing advanced AI models for businesses and developers.
Anthropic also faces competition from technology companies that integrate AI capabilities into their existing products and cloud services.
The competitive risks include:
Model capabilities: Competitors may develop AI systems with better performance, reliability, or specialized capabilities.
Pricing pressure: Competing providers may offer similar capabilities at lower prices or subsidize AI services using revenue from other businesses.
Open-weight alternatives: Models whose trained parameters are made available to others may reduce demand for certain proprietary AI services.
Distribution advantages: Large technology companies may bundle AI functionality with software or cloud services that customers already use.
Internal customer development: Large enterprises may develop their own AI models instead of purchasing Anthropic’s services.
Model replication: Competitors or other third parties may attempt to reproduce Anthropic’s model capabilities through techniques such as model distillation, which involves training another model using the outputs of an existing model.
Anthropic warns that technological advantages may become less durable as advanced AI capabilities become more widely available.
Competition could increasingly depend on computing access, pricing, distribution, or specialized applications rather than differences in general model performance alone.
Plain English: Even if Claude remains technically capable, cheaper models, competing AI products, or AI services bundled with existing software could reduce customer demand or force Anthropic to lower prices.
3.7. Uncertain AI Demand and Consumption-Based Revenue
Anthropic has experienced exceptionally rapid revenue growth, but the company warns that historical growth rates may not continue.
The AI market remains relatively young, and future customer adoption, spending levels, and preferred applications are uncertain.
A substantial portion of Anthropic’s revenue is consumption-based, meaning customers pay according to their actual usage of its AI models and services.
This structure creates several risks:
Variable usage: Customers may reduce their AI usage without necessarily terminating their relationship with Anthropic.
Spending limits: Businesses may introduce usage budgets or internal controls that temporarily restrict consumption.
Model selection: Customers may shift from more expensive models to lower-priced alternatives.
Changing applications: Customer demand may move toward specialized products or workflows that Anthropic does not adequately support.
Customer substitution: Enterprises may replace Anthropic’s services with their own internal AI systems.
Anthropic also cautions that its Run-Rate Revenue metric should not be treated as a forecast of annual revenue.
Run-rate revenue estimates annualized revenue using a relatively short recent measurement period. Actual revenue may differ because customer usage can fluctuate, including during holiday and vacation periods.
The company also notes that some reported usage may come from customers later found to be violating its usage policies. Terminating those accounts can reduce future revenue.
Plain English: Anthropic’s revenue depends heavily on how much customers actually use its AI services. Customers can reduce spending or switch models, so rapid historical growth does not guarantee equally rapid future revenue growth.
3.8. Customer Concentration and Cloud Marketplace Distribution
A meaningful portion of Anthropic’s revenue comes from a relatively small group of large customers and distribution channels.
According to the S-1, the company’s five largest customers accounted for approximately 25% of revenue during the nine months ended September 30, 2026.
Although no single customer represented more than 10% of revenue during that period, reductions in spending by several large customers could materially affect results.
Anthropic also depends heavily on cloud service provider marketplaces to distribute Claude.
Revenue generated through these marketplaces represented approximately:
FY2024: 32% of total revenue.
FY2025: 47% of total revenue.
First Half 2026: 48% of total revenue.
These marketplaces provide important customer access, billing, and payment collection services.
However, disruptions or changes to marketplace agreements could create difficulties for customers accessing Claude, increase distribution expenses, or delay payments.
Anthropic generally maintains separate contractual arrangements with end customers, but moving those customers to alternative access channels may still be difficult or costly.
Some major customers also have contracts that permit changes in spending or termination with limited notice.
Plain English: Anthropic relies on a relatively small number of large customers and cloud marketplaces for a substantial portion of its revenue. Losing important customers or distribution arrangements could reduce sales even if overall AI demand remains strong.
3.9. Product Reliability and Enterprise Adoption
Anthropic must maintain reliable access to Claude while supporting rapidly increasing customer demand.
AI services require continuous access to computing infrastructure, and interruptions can affect customers that integrate Anthropic’s models into their software and business operations.
The S-1 identifies several operational challenges:
Service interruptions: Insufficient capacity, technical failures, or infrastructure problems could make Claude temporarily unavailable.
Performance degradation: Slower response times or reduced model quality could cause customers to reduce usage or seek alternatives.
System changes: Updates to models or infrastructure may introduce unexpected errors or disruptions.
Enterprise requirements: Large organizations may require extensive security controls, audit logs, compliance features, and service reliability before adopting Anthropic’s products.
Longer purchasing processes: Enterprise customers may require extensive evaluations, legal reviews, and technical integration before committing to purchases.
Anthropic disclosed that it experienced service outages and degraded performance in March 2026, including problems associated with rapidly increasing usage and infrastructure constraints.
The company also warns that increasingly autonomous AI applications may create additional reliability requirements because customers expect these systems to complete complex tasks without interruption.
Plain English: Businesses need Claude to remain available, reliable, and secure. Service outages or performance problems could reduce customer trust and make enterprise adoption more difficult.
3.10. Harmful Use of AI Models and Safeguard Failures
Anthropic warns that its AI models may be misused by malicious actors or other third parties, even when technical safeguards and usage policies are in place.
Potential misuse includes:
Generating fraudulent or misleading content.
Supporting cyberattacks or unauthorized access to computer systems.
Assisting with activities involving dangerous chemical, biological, or other weapons.
Producing harmful instructions or outputs that violate Anthropic’s usage policies.
Using techniques designed to bypass model safeguards.
One important threat is prompt injection, in which malicious instructions are placed in material processed by an AI system to manipulate its behavior.
Anthropic states that its safeguards have sometimes been circumvented, misconfigured, or inadvertently disabled.
The company also disclosed that, in September 2025, it identified a Chinese state-sponsored hacking group that had manipulated Claude Code into attempting intrusions against approximately 30 targets, resulting in a limited number of successful intrusions.
Anthropic warns that increasingly capable AI models may provide more powerful tools for malicious activities, potentially increasing the consequences of safeguard failures.
Such incidents could result in legal liability, regulatory restrictions, loss of customer relationships, and reputational harm.
Plain English: If attackers find ways to misuse Claude or bypass its safeguards, Anthropic could face legal and reputational consequences even when harmful activity was initiated by third parties.
3.11. Unintended Behavior and Autonomous AI Agent Risks
Anthropic identifies risks associated with increasingly autonomous AI systems that can perform actions rather than simply generate responses.
Its agentic products can interact with external tools, execute code, modify files, access information, and complete tasks through multiple steps.
These capabilities introduce risks that differ from those of conventional chatbots.
The company warns that AI models may:
Perform unintended actions: An agent may execute an action that the user did not authorize.
Compound errors: A mistake early in a multi-step task may affect subsequent actions.
Bypass oversight: More capable models may become better at recognizing or circumventing certain evaluation and monitoring techniques.
Interact unpredictably: Multiple agents operating together may reinforce errors or behave in ways that are difficult to anticipate.
Cause real-world harm: Actions involving computer systems, financial transactions, laboratory equipment, or physical devices may have consequences that are difficult to reverse.
Anthropic disclosed incidents during 2026 pre-release cybersecurity evaluations in which four models gained unauthorized access to real computer systems outside their intended testing environments.
These incidents involved testing configurations that inadvertently permitted external access. In one case, an early model version obtained administrative access to a third-party system and accessed personal information.
Anthropic reported that affected parties were notified and that its investigation found no evidence of coordination among models or deliberate attempts to evade oversight.
Following these incidents, the company suspended certain pre-release cybersecurity evaluations and higher-risk training environments.
Anthropic also warns that using its own AI models in internal research and software development could introduce errors or security vulnerabilities into future systems.
Plain English: As Claude gains the ability to take independent actions, mistakes or unexpected behavior could cause problems beyond inaccurate answers, including unauthorized changes to computer systems or other real-world harm.
3.12. AI Safety Commitments and Delayed Product Releases
Anthropic places substantial emphasis on AI safety and responsible model development.
However, the company warns that its safety commitments may conflict with commercial objectives and competitive pressures.
Anthropic may determine that certain capabilities require additional testing, limited availability, or delayed deployment.
Examples disclosed in the S-1 include:
Restricted model access: Anthropic initially released its Mythos model through a limited research preview rather than making it broadly available.
Training delays: The company delayed pre-training a new frontier model for more than one month to strengthen protections against theft of model weights and intellectual property.
Usage restrictions: Anthropic prohibits certain applications, including fully autonomous weapons and mass surveillance.
Commercial opportunities not pursued: The company has chosen not to develop certain offerings, including image and video generation models, in order to concentrate resources on other research and safety priorities.
These decisions may reduce revenue opportunities, delay product launches, or allow competitors to introduce certain capabilities earlier.
At the same time, Anthropic acknowledges that its safety assessments and safeguards may prove insufficient, potentially exposing the company to incidents, regulatory scrutiny, and reputational harm.
Because the company publicly emphasizes safety as a central part of its identity, perceived failures to meet those commitments may create additional reputational consequences.
Plain English: Anthropic may delay or restrict profitable AI products because of safety concerns. These decisions could reduce commercial opportunities, while failures in its safety systems could damage customer trust.
3.13. Cybersecurity, Model Weight Theft, and Sensitive Data
Anthropic handles valuable proprietary technology and sensitive information belonging to itself and its customers.
Its most important technological assets include proprietary AI model weights, which are the trained numerical parameters that determine how a model processes information and generates outputs.
Unauthorized access to these weights could allow third parties to reproduce or operate important aspects of Anthropic’s technology outside its intended safeguards.
The company identifies several security risks:
Model weight theft: Competitors, criminal groups, or state-sponsored actors may attempt to obtain proprietary AI models.
Customer data exposure: Security incidents may reveal confidential information processed through Claude.
AI-specific attacks: Prompt injection, model extraction, training-data manipulation, and attacks on connected tools may bypass conventional defenses.
Insider threats: Employees or contractors with authorized access may be compromised or may misuse sensitive information.
Third-party vulnerabilities: Security failures involving cloud providers or other partners could affect Anthropic’s systems and customers.
Anthropic warns that advanced AI models are particularly attractive targets for sophisticated attackers because of their commercial and technological value.
The company also faces competing requirements involving data retention and security monitoring.
Some enterprise customers request zero-data-retention arrangements, under which certain customer data is not stored after processing. Although these arrangements can support customer privacy requirements, they may limit Anthropic’s ability to investigate misuse, reconstruct incidents, or provide information requested by regulators.
Security incidents or inadequate monitoring could result in customer losses, legal liability, regulatory action, or restrictions on model deployment.
Plain English: Anthropic must protect both its proprietary AI technology and sensitive customer information. Theft of model weights or breaches involving customer data could cause substantial competitive and legal harm.
3.14. Copyright, Training Data, and Intellectual Property Disputes
Anthropic uses large datasets to train and improve its AI models.
These datasets may include publicly available online material, licensed information, commissioned datasets, and synthetic data generated by AI systems.
The company faces legal uncertainty regarding whether certain training materials can be collected or used without violating copyright, contractual, privacy, or other legal rights.
Important risks include:
Copyright claims: Rights holders may allege that training data was copied or used without appropriate authorization.
Web scraping disputes: Website operators may challenge the methods used to collect online content.
AI-generated outputs: Third parties may allege that content generated by Claude infringes protected works.
Training-data restrictions: Legal decisions or settlements may require licensing, changes to data collection practices, or restrictions on certain datasets.
Customer indemnification: Certain agreements may require Anthropic to defend customers against specified intellectual property claims, potentially creating substantial financial obligations.
Anthropic disclosed a $1.5 billion settlement in the Bartz copyright litigation, which received district court approval in July 2026.
The company also faces other intellectual property litigation, including disputes involving music publishers and a lawsuit brought by Reddit concerning alleged unauthorized access to platform content.
Anthropic disputes allegations in certain ongoing proceedings and warns that their outcomes remain uncertain.
Future claims could lead to additional settlements, damages, licensing expenses, or changes in the data and technologies the company is permitted to use.
Plain English: If courts or regulators restrict how Anthropic collects or uses training data, the company may face additional legal costs or need to change how it develops and operates its AI models.
3.15. AI Regulation, Export Controls, and Government Restrictions
Anthropic operates in an industry where legal and regulatory requirements are developing rapidly.
Rules governing advanced AI may affect model development, safety testing, cybersecurity, transparency, privacy, and the uses permitted for particular systems.
The company warns that differing requirements across jurisdictions could increase compliance expenses or restrict the availability of certain products.
Several regulatory risks are particularly relevant to Anthropic:
AI-specific legislation: New laws may require additional testing, disclosures, safety measures, or restrictions on high-risk AI applications.
Privacy and data protection: Restrictions on collecting or processing personal information may affect model development and customer services.
Export controls: Government restrictions may limit where advanced AI models can be provided or who can access certain technologies.
Government procurement: Public-sector customers may impose specialized security requirements or restrictions on permitted applications.
Regulatory investigations: Government agencies may examine model safety, consumer protection, competition, or data-handling practices.
Anthropic disclosed that, in June 2026, the U.S. Department of Commerce imposed worldwide export restrictions on its Fable 5 and Mythos 5 models.
The company temporarily disabled those models for customers before restoring access after updating safeguards and the lifting of the restrictions.
This example demonstrates that government action can directly affect the availability of Anthropic’s AI products.
Anthropic also disclosed receiving an investigative request from the Federal Trade Commission in October 2026 concerning consumer protection, data protection, privacy, and product safety and security practices.
Regulatory requirements may therefore affect not only Anthropic’s compliance costs but also which AI capabilities it can commercially deploy.
Plain English: Government rules could limit where Anthropic sells its models, which capabilities it offers, or how it develops and tests AI systems. Such restrictions could increase costs or interrupt access to certain products.
3.16. Profitability, Financing Needs, and Convertible Securities
Anthropic’s business requires substantial spending on advanced AI research, computing infrastructure, personnel, and commercial operations.
Although the company has experienced rapid revenue growth, it has also reported large historical net losses.
Anthropic warns that it may not achieve or sustain profitability and may require additional capital to support its operations and long-term infrastructure commitments.
The company identifies several financial risks:
High operating costs: Developing competitive AI models requires substantial ongoing investment.
Fixed payment commitments: Long-term computing agreements may require payments even when customer demand falls below expectations.
Additional financing: The company may need to issue additional equity or debt to finance its activities.
Debt obligations: Certain financing arrangements require interest or principal payments and contain contractual obligations that could create financial pressure.
Convertible securities: The conversion of outstanding financing instruments into common shares can dilute existing shareholders’ economic interests.
Accounting volatility: Changes in the estimated fair value of convertible instruments can create substantial fluctuations in reported net income.
Anthropic’s S-1 also identifies the risk that defaults under certain commercial and financing arrangements could trigger repayment obligations under other agreements.
Such provisions could create additional financial pressure if the company cannot meet its commitments.
Even substantial cash reserves do not eliminate the obligation to fund future computing capacity, research, and other expenses.
Plain English: Anthropic must continue spending heavily to develop and operate its AI business. If cash generation is insufficient to cover its obligations, it may need additional financing, potentially increasing debt or diluting shareholders.
3.17. Dependence on Key Researchers, Founders, and Company Culture
Anthropic’s ability to develop competitive AI systems depends heavily on a relatively limited group of highly skilled researchers, engineers, and senior executives.
The company specifically emphasizes its dependence on its co-founders, management team, and key technical personnel.
The market for frontier AI talent is highly competitive, with major technology companies and specialized research organizations competing for experienced employees.
Anthropic warns that losing key personnel could:
Delay AI model research and development.
Reduce its ability to maintain technological leadership.
Disrupt important commercial or infrastructure relationships.
Increase recruitment and compensation expenses.
Make it more difficult to preserve the company’s safety-focused culture.
Anthropic also identifies its mission-driven culture as an important factor in attracting and retaining employees.
Rapid expansion, competitive pressures, and the demands of operating as a public company could make that culture more difficult to maintain.
The company warns that disagreements over safety priorities, commercial decisions, or organizational direction could affect employee morale and retention.
Plain English: Anthropic depends on specialized AI researchers and a safety-focused company culture. Losing important employees or failing to preserve that culture could slow technological development and weaken its competitive position.
3.18. Unusual Corporate Governance and Limited Shareholder Control
Anthropic has adopted a corporate governance structure that differs substantially from the structures used by most publicly traded companies.
It is incorporated as a Delaware public benefit corporation (PBC), which requires directors to balance shareholders’ financial interests with the interests of affected stakeholders and the company’s stated public benefit purpose.
Anthropic’s governance structure also includes special classes of common stock that separate voting authority from economic ownership.
Two arrangements are particularly important:
Long-Term Benefit Trust (LTBT): Through Class T common stock, the LTBT generally has authority to elect and remove directors who constitute a majority of Anthropic’s authorized board.
Founder LLC: Through Class F common stock, an entity established by Anthropic’s co-founders will hold voting power equal to 50.1% of the total voting power on matters generally submitted to shareholders.
Anthropic’s Class A common stock, the class proposed for public trading under the ticker ANT, will carry ordinary voting rights.
However, the special voting arrangements mean that public Class A shareholders generally will have very limited ability to influence board elections or other important corporate decisions.
The LTBT and Founder LLC may also exercise their authority in ways that differ from the preferences of ordinary shareholders.
Anthropic warns that its structure may affect decisions involving corporate strategy, acquisitions, governance changes, and the balance between commercial returns and AI safety objectives.
The company is also considered a controlled company under Nasdaq’s corporate governance rules, allowing it to rely on certain exemptions from requirements that otherwise apply to listed companies.
Plain English: Buying Anthropic’s publicly traded Class A shares would provide an economic ownership interest, but ordinary shareholders would have little practical control over major corporate decisions because voting authority is concentrated in the trust and founder-controlled structure.
3.19. Public Benefit Purpose and Shareholder Financial Interests
Anthropic’s status as a public benefit corporation creates an additional risk that is separate from its voting structure.
The company’s stated purpose includes the responsible development and maintenance of advanced AI for the long-term benefit of humanity.
Under Delaware law, Anthropic’s directors must consider this purpose and the interests of affected stakeholders alongside shareholders’ financial interests.
As a result, the company may make decisions that do not maximize near-term revenue, profits, or shareholder returns.
Anthropic identifies several examples of decisions that could reflect this balance:
Restricting or delaying the commercial release of advanced AI models.
Declining certain uses of its technology because of safety concerns.
Prioritizing research and safety investments over immediate commercial opportunities.
Choosing not to develop certain potentially profitable products.
Supporting policies or safeguards that may constrain its own commercial activities.
The company warns that shareholders may disagree with decisions made under this framework.
It also acknowledges that legal precedent governing the responsibilities of directors of public benefit corporations remains limited, creating uncertainty about how certain disputes might be resolved.
Anthropic’s governance and public benefit commitments may also make unsolicited acquisitions or attempts to change corporate control more difficult.
Plain English: Anthropic may prioritize AI safety and its public benefit mission even when doing so reduces potential financial returns. Shareholders should not assume that maximizing profits will always determine management decisions.
3.20. Summary of Anthropic’s Key Risks
Anthropic’s S-1 highlights a group of interconnected risks arising from the capital-intensive and rapidly evolving nature of frontier AI development.
Technological Leadership: Anthropic must continuously develop competitive AI models while managing the increasing complexity of research and model safety.
Computing Infrastructure: Limited computing capacity, expensive long-term commitments, and shortages of chips or electricity could restrict growth or increase costs.
Competitive Pressure: Proprietary AI developers, large technology companies, and open-weight models could reduce Anthropic’s differentiation and pricing power.
Customer and Distribution Concentration: Dependence on large customers and major cloud marketplaces could make revenue sensitive to changes in spending or commercial relationships.
AI Safety and Security: Misuse, unexpected autonomous behavior, model theft, and insufficient safeguards could create legal, operational, and reputational consequences.
Training Data and Regulation: Copyright disputes, data-use restrictions, AI legislation, and export controls could increase costs or limit product availability.
Financial Commitments: Substantial research expenses, computing obligations, and financing arrangements could require additional capital and affect existing shareholders.
Corporate Governance: Anthropic’s public benefit purpose and special voting structure substantially limit ordinary shareholders’ influence over major decisions.
Plain English: Anthropic’s principal risks center on whether it can continue developing advanced AI models, obtain enough computing power, serve customers safely and reliably, and meet its financial commitments while operating under an unusual governance structure.
These risks reflect the uncertainties identified by Anthropic in its S-1 and do not establish that any particular adverse outcome will occur.
4. MD&A (Management’s Discussion and Analysis)
Management’s Discussion and Analysis (MD&A) explains how Anthropic’s management evaluates its business, what drove recent financial results, and which factors the company believes will influence its future performance.
In its 2026 S-1 filing, Anthropic emphasizes four primary financial drivers: Claude adoption, compute efficiency, operating efficiency, and investment in computing infrastructure.
Management describes these factors as closely connected. Improvements in Claude’s capabilities encourage greater customer adoption, while research and engineering improvements allow the company to generate more value from its computing resources. At the same time, expanding computing capacity requires substantial financial commitments that may affect expenses and cash flow before the associated revenue is realized.
4.1. Management’s Financial Model: Four Primary Drivers
Anthropic identifies four factors that management considers fundamental to its financial performance.
Claude Adoption: More capable AI models encourage new customers to adopt Claude and existing customers to expand usage across additional business activities.
Compute Efficiency: Improvements in AI training, inference, and hardware utilization increase the amount of intelligence produced and revenue generated for each dollar spent on computing resources.
Operating Efficiency: Anthropic seeks to improve productivity through specialized employees, efficient distribution, and extensive internal use of Claude.
Compute Investments: The company continues committing substantial capital to the computing capacity required for AI research, model development, and commercial services.
Management emphasizes that these drivers do not necessarily affect financial results at the same time.
Computing capacity is typically secured well before it becomes operational. Once the capacity is available, related expenses may begin before customer revenue has expanded enough to fully utilize it.
As a result, management expects operating margins to fluctuate, including during periods when significant additional computing resources become available.
Plain English: Anthropic’s financial model depends on developing better AI models, attracting more customer usage, using computing resources efficiently, and investing in additional capacity. Because infrastructure costs can arrive before the revenue they support, profitability may fluctuate even as the business expands.
4.2. Claude Adoption and Enterprise Revenue Growth
Management attributes Anthropic’s rapid commercial expansion primarily to improvements in Claude’s capabilities and increasing adoption by organizations.
The company develops successive generations of Claude models intended to improve coding, complex reasoning, autonomous workflows, and other knowledge-intensive tasks.
Management states that more capable models encourage customers to discover additional applications for AI and expand their existing deployments.
Anthropic reports that its revenue increased from $41 million in FY2023 to $386 million in FY2024 and $4,585 million in FY2025.
Revenue then reached $16,286 million during the first six months of 2026, compared with $1,163 million in the corresponding period of 2025.
Management identifies successive model and product releases, higher usage, broader adoption, and increased spending by existing customers as the primary drivers of this growth.
The company measures enterprise adoption partly through the number of organizations generating at least $100,000 in revenue over the preceding twelve months.
This is a trailing twelve-month (TTM) measure, meaning it tracks customer revenue accumulated over the most recent twelve months rather than a single quarter.
Anthropic reported the following progression:
December 31, 2023: Approximately 20 customers with at least $100,000 in TTM revenue.
December 31, 2024: Approximately 200 customers.
December 31, 2025: Approximately 1,500 customers.
September 30, 2026: More than 11,900 customers.
At September 30, 2026, the company also reported more than 1,900 customers generating at least $1 million in TTM revenue and approximately 250 customers generating at least $10 million.
Management attributes these increases to customers introducing Claude into more business activities after initially adopting it for a narrower set of tasks.
Anthropic reported Net Revenue Retention of approximately 900% as of September 30, 2026.
Net Revenue Retention measures how revenue from an existing group of customers changes over time, including increased spending, reduced spending, and customer losses. Anthropic calculates this metric using qualifying customers, excluding certain fixed-price consumer subscriptions.
Management views this metric as evidence of substantial expansion in usage among existing customers. The company also cautions that its exceptional historical growth rates may not be sustainable as its revenue base increases and the AI market matures.
Plain English: Management says Claude’s improving capabilities are encouraging more companies to adopt its services and existing customers to spend more. However, Anthropic does not expect its unusually high historical revenue growth rates to continue indefinitely.
4.3. Consumption-Based Revenue and Subscription Revenue
Anthropic generates revenue primarily through two pricing arrangements: consumption-based fees and subscriptions.
Consumption-based revenue comes from customers paying for AI model usage. Pricing generally depends on the model selected and the number of tokens processed.
A token is a small unit of information processed by an AI model, such as part of a word, a word, or another piece of input or output content.
Customers select models according to the capabilities, processing speed, and cost required for their applications.
Anthropic also offers subscription-based products, under which customers pay recurring fees for access to specified capabilities and usage allowances.
Management states that a growing number of enterprise customers enter annual or multi-year contracts containing minimum spending commitments.
These arrangements generally produce consumption-based revenue as customers use the contracted services.
The company’s reported revenue mix was:
FY2024: Consumption-based revenue of $288 million and subscription revenue of $98 million.
FY2025: Consumption-based revenue of approximately $3,796 million and subscription revenue of $789 million.
First Half 2026: Consumption-based revenue of approximately $13,064 million and subscription revenue of $3,222 million.
Consumption-based revenue represented approximately 80% of total revenue during the first half of 2026.
Management expects consumption-based revenue to remain the substantial majority of total revenue for the foreseeable future.
The company recognizes consumption-based revenue as customers use its services. Subscription revenue is generally recognized over the subscription period.
Anthropic also notes that consumption-based revenue may fluctuate with customer activity, seasonal usage patterns, model releases, and changes in pricing.
Management specifically identifies lower usage during certain summer and winter holiday periods as a recurring consideration.
Additionally, the company estimates that approximately 5% to 7% of revenue during the nine months ended September 30, 2026 came from customers subsequently identified as violating its usage policies or commercial terms.
Anthropic strengthened its detection and enforcement efforts during 2026 and expects the proportion of revenue from these customers to decrease, although the timing and financial effects remain uncertain.
Plain English: Most of Anthropic’s revenue depends on how extensively customers use Claude. Management expects this usage-based model to remain dominant, but customer activity, seasonal patterns, pricing, and enforcement of usage policies can affect reported revenue.
4.4. Enterprise Distribution and Customer Expansion
Management describes Claude’s distribution strategy as a combination of individual adoption, direct enterprise relationships, and partnerships.
Anthropic identifies several methods of expanding customer adoption:
Individual-Led Adoption: Employees and developers begin using Claude for their own work and subsequently introduce it to their organizations.
Direct Sales Teams: Anthropic works with businesses to expand existing deployments and support broader organizational adoption.
Applied AI Teams: Specialized teams help enterprise customers design and deploy AI solutions for their own systems and workflows.
Cloud Partnerships: Claude is available through Amazon Web Services, Google Cloud, and Microsoft Azure.
Technology and Consulting Partners: Systems integrators, consultants, and data platform providers help businesses incorporate Claude into existing infrastructure.
Management emphasizes that individual developers have been especially important to Claude’s adoption.
According to the company, adoption often begins with coding or another specific application before expanding into broader activities such as financial analysis, document processing, customer operations, and legal workflows.
Anthropic also highlights its integrated product strategy, combining Claude models, Claude Platform, and applications such as Claude Code and Claude Cowork.
For the three months ended September 30, 2026, more than 50% of customers generating at least $100,000 in TTM revenue used both Claude Platform and at least one first-party Anthropic application.
That percentage was approximately 30% for the three months ended December 31, 2025.
Management views the increasing use of multiple Anthropic products as part of its broader enterprise adoption strategy.
The company also emphasizes opportunities to expand spending within organizations already using Claude.
As of September 30, 2026, more than 50% of Fortune 500 companies and more than 30% of Global 2000 companies were customers meeting Anthropic’s $100,000 TTM revenue threshold.
Management states that average spending by these large organizations still represented a small fraction of their total operating expenses, which it identifies as an opportunity to expand existing customer relationships.
Plain English: Anthropic seeks to expand revenue by making Claude available through products and distribution channels that businesses already use. Management also emphasizes increasing usage within existing customers, not only attracting new organizations.
4.5. Compute Efficiency: Producing More Value from AI Infrastructure
Management identifies computing efficiency as one of the most important determinants of Anthropic’s financial performance.
Computing resources are required both to develop AI models and to process customer requests.
Anthropic describes three principal methods of improving efficiency.
1. Training Efficiency
Management emphasizes improvements in the algorithms and techniques used to train AI models.
These improvements can allow researchers to develop more capable systems with a given amount of computing power.
Anthropic also uses Claude internally to assist with research and engineering work involved in developing later model generations.
The company describes this process as closing the loop, meaning that existing AI models contribute to the development of their successors.
2. Inference Efficiency
Inference is the process of using a trained model to generate answers or complete tasks.
Management states that research and engineering improvements have reduced the computing resources required to deliver particular levels of model performance.
These improvements allow Anthropic to serve additional customer workloads using available infrastructure.
3. Multi-Chip Computing
Anthropic develops its infrastructure to support several families of AI processors, including:
AWS Trainium.
Google TPUs.
NVIDIA GPUs.
AMD GPUs.
Management states that its proprietary orchestration technology enables Anthropic to direct workloads toward different hardware systems based on cost and performance.
This approach is intended to increase infrastructure flexibility and reduce dependence on any single processor provider.
For the three months ended June 30, 2026, management highlighted that revenue increased approximately 1,368% year over year, while compute and infrastructure expenses increased approximately 300%.
Management presents this difference as evidence of improving revenue generation relative to computing expenditure.
Plain English: Anthropic is working to develop better AI models and serve more customers without increasing computing costs at the same rate. Management considers training improvements, inference optimization, and support for multiple chip platforms central to this effort.
4.6. Operating Efficiency and Employee Productivity
Management attributes improvements in operating efficiency to the company’s specialized workforce, distribution approach, and extensive internal use of Claude.
Anthropic reports that Claude increasingly contributes to its own software development and engineering activities.
According to management, more than 90% of contributions to Anthropic’s codebase in September 2026 were written by Claude.
The company also reported that the average daily volume of code contributed per engineer during the three months ended September 30, 2026 was more than 45 times the average daily volume recorded during FY2024.
Management also compares revenue growth with employee growth.
Between September 30, 2025 and September 30, 2026:
TTM Revenue: Increased approximately 1,200%.
Headcount: Increased approximately 170%.
TTM Revenue per Employee: Increased from approximately $1.4 million to more than $6.6 million.
Management identifies the use of Claude within its own operations as an important contributor to employee productivity.
The company also describes its distribution strategy as relatively efficient because individual adoption and cloud partnerships can support customer growth without requiring a proportional increase in direct sales personnel.
These operating efficiency measures are part of management’s explanation for the improvement in reported operating margins.
Plain English: Management says Anthropic’s employees are increasingly using Claude to perform their own work. The company reports that revenue expanded much faster than employee headcount during the year ended September 2026.
4.7. Computing Investments and the Anthropic Direct Strategy
Anthropic emphasizes that continued investment in computing infrastructure is essential to developing and operating advanced AI models.
Historically, the company obtained much of its computing capacity through long-term agreements with major cloud service providers.
Management states that these agreements are typically arranged 12 to 18 months before the associated computing capacity becomes available.
This creates a difference between the timing of infrastructure commitments, the recognition of related expenses, and the generation of customer revenue.
Management also explains that Anthropic has historically operated with insufficient computing capacity to meet all potential customer and research demand.
When long-term contracted capacity is inadequate, the company may purchase additional short-term computing resources.
Management states that short-term computing capacity generally costs more per unit than capacity secured through longer-term agreements.
To support future requirements, Anthropic is expanding its infrastructure strategy through Anthropic Direct.
Under this arrangement, the company combines external cloud services with directly leased AI processors hosted in dedicated facilities operated on its behalf.
More Flexible Capital Allocation: Separate commitments for data center facilities from commitments for the AI equipment installed inside them.
Improved Equipment Timing: Obtain data center capacity early while postponing some processor commitments until computing requirements are clearer.
Long-Term Cost Efficiency: Increase the proportion of computing capacity acquired through arrangements management expects to be more economical over time.
Management expects Anthropic Direct to account for a growing portion of its infrastructure investment.
However, the strategy requires substantial capital expenditures and lease commitments, often before the related infrastructure can support commercial operations.
Plain English: Anthropic is expanding beyond rented cloud computing services by leasing more AI equipment directly and arranging dedicated hosting facilities. Management expects this to provide greater flexibility, but it also requires substantial financial commitments before the equipment begins producing revenue.
4.8. Why Management Expects Operating Margins to Fluctuate
Management emphasizes that the timing of computing infrastructure deliveries can cause significant fluctuations in profitability.
Anthropic purchases or contracts for computing capacity in advance, but the related expense generally begins when the capacity becomes available.
The company uses a shared pool of computing resources for research, model development, customer services, and internal operations.
Consequently, management evaluates computing expenditure across the business rather than assigning every infrastructure cost to a single activity.
When a substantial amount of new capacity becomes available, infrastructure expenses may increase immediately.
Customer revenue associated with that capacity may develop more gradually.
Management therefore cautions that operating margins may temporarily decline when additional computing resources become operational, even if customer adoption and operating efficiency continue improving.
The company also expects increased purchases of short-term computing capacity in the near term, which may place additional pressure on operating margins because these resources are generally more expensive.
Anthropic states that this pattern is an inherent feature of its financial model at its current stage of development.
Plain English: New computing infrastructure can increase Anthropic’s expenses before enough additional customer revenue arrives to cover those costs. Management therefore does not expect profitability to improve in a straight line every quarter.
4.9. Management’s Explanation of FY2023–FY2025 Financial Results
Management attributes Anthropic’s rapid revenue growth during FY2023–FY2025 to the commercial introduction of Claude, successive improvements in model capabilities, and increasing customer usage.
Revenue Growth
Revenue increased from $41 million in FY2023 to $386 million in FY2024.
Management attributes the FY2024 increase primarily to the addition of new customers and broader availability of the Claude platform following the release of the Claude 3 model family.
Revenue then increased to $4,585 million in FY2025, supported by additional model releases, higher usage, and broader adoption across Anthropic’s products.
Computing Expenses
Compute and infrastructure expenses increased from $433 million in FY2023 to $2,522 million in FY2024 and $7,326 million in FY2025.
Management explains that the increases reflected the staged delivery of computing capacity under agreements entered into in earlier periods.
As this capacity became available, Anthropic recognized the associated operating expenses.
Despite higher absolute computing costs, these expenses declined substantially relative to revenue because commercial activity grew faster than the previously committed infrastructure cost base.
Personnel and Commercial Expenses
Personnel expenses increased as Anthropic expanded its workforce. Management identifies both cash compensation and stock-based compensation as important contributors.
Sales, marketing, and partnerships expenses also increased as Anthropic expanded distribution arrangements, increased payments to channel partners, and supported product launches.
Other Operating Expenses
Management identifies several other factors affecting FY2025 operating results, including:
A $1,500 million charge associated with the Bartz copyright litigation settlement.
Increased costs for data quality and model evaluation services.
Higher information technology and cloud software expenses.
$541 million in noncash charges associated with the company’s employee donation matching program.
Management also explains that substantial fair-value changes in convertible financial instruments materially increased reported net losses, even though those accounting effects were separate from ordinary operating expenditures.
Plain English: Management attributes Anthropic’s historical financial results to rapid revenue expansion, substantial investment in computing resources, workforce growth, commercial spending, and certain legal and noncash accounting charges.
4.10. Management’s Explanation of First Half 2026 Financial Results
Management describes the first half of 2026 as a period of exceptionally rapid revenue growth and improving operating performance.
Revenue
Revenue increased from $1,163 million in the first half of 2025 to $16,286 million in the first half of 2026.
Management attributes the increase to successive model and product releases, broader adoption, and expanded usage by existing customers.
Compute and Infrastructure
Compute and infrastructure expenses increased from $2,334 million to $8,945 million.
Management explains that the increase reflected the delivery of additional contracted capacity under cloud provider arrangements.
However, computing expenses represented approximately 55% of revenue in the first half of 2026, compared with approximately 201% in the prior-year period.
Management attributes the decline in this expense ratio to rapid revenue growth against computing costs that had largely been committed in advance.
Personnel
Personnel expenses increased from $538 million to $3,152 million.
The increase primarily reflected higher stock-based compensation and cash compensation associated with workforce expansion.
Sales, Marketing, and Partnerships
Expenses increased from $120 million to $3,168 million.
Management states that higher fees paid to channel partners were the principal contributor, along with increased advertising expenses supporting product launches.
Other Operating and Administrative Expenses
These expenses increased from $306 million to $1,714 million.
Management identifies higher data evaluation costs, software expenses, payment processing fees, professional services, and charitable donation matching charges as important contributors.
Operating Results
Anthropic’s operating loss narrowed from $2,135 million in the first half of 2025 to $693 million in the first half of 2026.
In the second quarter of 2026, the company generated $974 million in operating income.
Management attributes the improvement to the combined effects of revenue growth, computing efficiency, and operating efficiency.
The company nevertheless cautions that future margins may fluctuate as additional computing capacity becomes operational.
Plain English: During the first half of 2026, Anthropic’s revenue increased much faster than its total operating expenses. Management identifies this relationship as the primary reason operating losses narrowed and the company generated an operating profit in the second quarter.
4.11. GAAP Results, Non-GAAP Measures, and Fair-Value Accounting
Anthropic reports its financial results under U.S. Generally Accepted Accounting Principles, commonly called GAAP.
Management also presents certain non-GAAP financial measures, which exclude specified accounting items to provide an additional view of operating performance.
Anthropic’s principal non-GAAP measures include:
Adjusted Operating Income (Loss): Operating income or loss excluding stock-based compensation, noncash charitable contributions, and the Bartz settlement.
Adjusted Operating Margin: Adjusted operating income or loss divided by revenue.
Adjusted Net Income (Loss): Net income or loss excluding the specified operating adjustments and certain fair-value changes associated with financial instruments.
Management states that these measures are used internally for evaluating performance, preparing budgets, allocating resources, and making business decisions.
For the first half of 2026, Anthropic reported:
GAAP Operating Loss: $(693) million.
Adjusted Operating Income: $1,618 million.
GAAP Net Loss: $(80,748) million.
Adjusted Net Income: $2,671 million.
One of the largest adjustments involved the fair value of convertible financial instruments.
Certain convertible notes, derivative liabilities, and other financing instruments are remeasured at fair value at the end of each reporting period.
Management explains that increases in Anthropic’s estimated enterprise valuation can increase the recorded value of these obligations, creating accounting losses.
During the first half of 2026, Anthropic recognized a $72,388 million fair-value loss associated with convertible notes, derivative liabilities, and SAFEs.
The company also recognized substantial fair-value-related expenses associated with other financial instruments.
Management excludes specified items of this kind from Adjusted Net Income because it believes the adjusted measure provides additional information about ongoing performance.
However, Anthropic explicitly cautions that non-GAAP measures have limitations and should not replace its GAAP results.
Plain English: Anthropic’s GAAP net loss was heavily affected by changes in the accounting value of convertible financing instruments. Management also presents adjusted earnings that exclude specified items, but those figures are supplementary and do not replace the official GAAP results.
4.12. Liquidity, Cash Flow, and Financing Resources
Management states that Anthropic has historically financed its operations and infrastructure investments through convertible notes, preferred stock, other financing instruments, and revenue generated from customers.
As of June 30, 2026, Anthropic reported:
Cash and Cash Equivalents: $112,389 million.
Short-Term Investments: $19,262 million.
Available Capacity Under Its Then-Existing Credit Facility: Approximately $3,200 million.
In September 2026, Anthropic entered into an amended and restated credit agreement providing a $15,000 million senior secured revolving credit facility.
A revolving credit facility is a borrowing arrangement that allows a company to access funds up to an agreed limit, subject to the terms and conditions of the agreement.
Management believes its available cash, customer-generated cash flow, and borrowing capacity will be sufficient to meet working capital, lease payments, and capital expenditure requirements for at least twelve months from the prospectus date.
However, the company also emphasizes that its future funding requirements depend on revenue growth, computing investments, infrastructure commitments, and other operating needs.
Operating Cash Flow
Operating cash flow improved from $(1,513) million in the first half of 2025 to $3,929 million in the first half of 2026.
Management attributes the improvement to changes in net loss after noncash adjustments and movements in working capital, including the timing of customer payments and payments to suppliers.
Investing Cash Flow
Net cash used in investing activities increased to $17,181 million during the first half of 2026.
Management attributes most of the increase to purchases of short-term financial investments, along with higher expenditures on servers, networking equipment, and other infrastructure.
Financing Cash Flow
Financing activities generated $110,774 million during the first half of 2026.
The principal source was the issuance of redeemable convertible preferred stock.
Management emphasizes that Anthropic expects to continue using a combination of operating cash flow, equity financing, debt financing, and arrangements with strategic partners to support its long-term requirements.
Plain English: Anthropic entered the second half of 2026 with substantial financial resources. Management considers its near-term liquidity sufficient, while acknowledging that continued infrastructure expansion will require significant spending and potentially additional financing.
4.13. Contractual Obligations and Future Infrastructure Spending
Anthropic has entered into substantial agreements to secure future computing capacity and data center infrastructure.
Management describes these arrangements as essential to supporting continued AI development and anticipated customer demand.
Its obligations include:
Cloud Infrastructure Agreements: Long-term commitments to purchase computing services from major cloud providers.
Equipment Leases: Arrangements for using AI processors and related infrastructure over specified periods.
Data Center Commitments: Contracts supporting facilities, electrical power, and other resources required for dedicated computing operations.
Financing Arrangements: Credit agreements and partner-supported financing structures associated with infrastructure development.
Management explains that infrastructure investments increasingly involve substantial upfront capital expenditures and lease obligations rather than only payments for cloud services as capacity becomes available.
The company expects this transition to change the timing and composition of cash expenditures.
Management states that investment in owned and leased computing resources will be a significant use of cash in operating and investing activities for the foreseeable future.
Some commitments extend over many years, and the final cost of certain arrangements will depend on the timing of capacity delivery, equipment availability, and contractual conditions.
Management also emphasizes that access to computing capacity and financing can be connected because some infrastructure suppliers and strategic partners provide or facilitate capital support.
The company intends to meet these commitments through operating revenue and cash flow, existing financial resources, debt and equity financing, and partner arrangements.
Management does not present the timing of all future infrastructure investments as fixed. It expects investment requirements to change as computing needs, delivery schedules, and commercial conditions develop.
Plain English: Anthropic has committed to acquiring substantial computing capacity over future years. Management expects these investments to remain a major use of cash, including payments that may occur well before the infrastructure begins serving customers.
4.14. Public Benefit Commitments and Operating Expenses
Anthropic’s public benefit purpose is an important consideration in management’s operating and financial decisions.
The company states that responsible AI development and broad access to the benefits of AI are central to its mission.
Management expects to allocate resources toward initiatives that support this purpose, including research, charitable contributions, and programs expanding access to Claude.
Anthropic maintains an employee donation matching program under which qualifying employee donations can be matched through equity-based contributions.
During FY2025, the company recorded $541 million in expenses associated with this program.
Management also describes a broader framework for supporting public benefit initiatives through contributions of company shares.
These initiatives may affect reported operating expenses and, where additional equity is issued, the number of shares outstanding.
Anthropic states that decisions supporting its public benefit purpose may reduce revenue, increase expenses, or affect profitability during particular periods.
Management nevertheless identifies these activities as part of the company’s long-term mission and operating framework.
Plain English: Anthropic allocates financial resources to AI safety and public benefit activities. Management states that these commitments may affect expenses, profitability, and share ownership.
4.15. Critical Accounting Estimates
Management identifies several accounting areas that require significant judgment because their reported values depend on assumptions and estimates.
These include:
Stock-Based Compensation: Estimating the accounting value of employee stock options, restricted stock units, and other equity awards.
Common Stock Valuation: Estimating the fair value of shares before an established public trading market exists.
Convertible Financial Instruments: Determining the fair value of outstanding notes, derivatives, and related financing obligations.
Other Financial Estimates: Applying assumptions to financial transactions and obligations whose accounting values are not directly observable.
Management notes that significant changes in assumptions could affect reported financial results.
One important consideration involves restricted stock units, commonly called RSUs.
RSUs are equity awards that generally become shares when specified conditions are satisfied.
Anthropic disclosed approximately $26,600 million in unrecognized stock-based compensation expense related to RSUs as of June 30, 2026.
Management explains that the effectiveness of the IPO registration statement will satisfy a performance condition applicable to certain awards, resulting in the recognition of additional stock-based compensation expenses.
The company expects these expenses to affect future reported results, particularly around the completion of the IPO.
Management also emphasizes that estimates involving convertible instruments may create substantial changes in reported earnings as the company applies fair-value accounting.
Plain English: Some of Anthropic’s reported expenses depend on estimates rather than immediate cash payments. Management specifically identifies stock-based compensation and convertible financial instruments as important accounting areas, including additional equity compensation expenses expected around the IPO.
4.16. Management’s Outlook and Key Takeaways
Anthropic’s management emphasizes several priorities for its continued development.
Maintain Frontier AI Capabilities: Continue research and model development to improve Claude’s intelligence, usefulness, reliability, and safety.
Expand Enterprise Adoption: Introduce Claude to additional organizations and increase usage within existing customers.
Improve Compute Efficiency: Increase the intelligence produced and revenue generated from available computing resources.
Strengthen Operating Efficiency: Use internal AI capabilities, specialized employees, and distribution partnerships to support revenue growth.
Expand Computing Infrastructure: Continue securing cloud capacity and developing Anthropic Direct to support future research and commercial demand.
Maintain Access to Capital: Fund substantial computing commitments through operating cash flow and financing arrangements.
Preserve Safety and Public Benefit Priorities: Incorporate AI safety considerations and the company’s stated public benefit purpose into its development and business decisions.
Management expects Anthropic’s revenue growth rate to decline over time as the business becomes larger, even as it continues seeking broader AI adoption.
The company also expects absolute computing and infrastructure expenses to increase as additional contracted capacity becomes available.
Operating margins may therefore fluctuate, particularly when infrastructure expenses are recognized before the associated customer revenue has developed.
Management anticipates substantial ongoing infrastructure investment and expects its funding requirements to remain significant.
Plain English: Management’s central priorities are to keep Claude technologically competitive, increase customer adoption, improve efficiency, and secure enough computing resources to support continued development. The company expects large infrastructure investments and recognizes that revenue growth, profitability, and cash flow may fluctuate as those investments are made.
5. Summary
Anthropic (ANT) is an artificial intelligence company developing Claude, a family of advanced AI models used for software development, research, business automation, and other professional tasks. Its business combines AI model development, consumption-based services, subscriptions, and enterprise applications, with growth driven by increasing customer adoption and usage.
Anthropic’s revenue increased from $41 million in FY2023 to $4,585 million in FY2025, followed by $16,286 million in the first half of 2026. Operating performance also improved substantially, with the company generating $974 million in operating income during Q2 2026 and positive operating cash flow of $3,929 million for the first six months of the year. However, Anthropic continued reporting substantial net losses, largely affected by noncash accounting adjustments related to convertible financial instruments, while its growing cash reserves were supported primarily by external financing.
The company’s principal challenges include intense AI competition, substantial computing infrastructure commitments, model safety and security risks, and uncertainty over sustainable profitability. Its unusual governance structure also gives public shareholders limited influence over major corporate decisions. For beginner investors, the central takeaway is that Anthropic has demonstrated extraordinary commercial growth and improving operating performance, but its long-term financial success depends on sustaining those improvements while funding the substantial costs of developing and operating advanced AI systems.
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behavior or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional
Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
