Anthropic IPO & leveraged ETFs

What Does Anthropic Do? Products, Revenue and Business Model

Anthropic builds the Claude AI models and sells access to businesses and developers. Usage fees were 79% of revenue in the first nine months of 2026.
What Does Anthropic Do? Products, Revenue and Business Model
Anthropic, PBC is an AI research company that builds the Claude family of AI models and sells access to them to businesses, developers and individual subscribers. Most revenue comes from usage-based fees metered in tokens, and the rest comes from subscriptions. Anthropic reported $33.2 billion of revenue for the nine months ended September 30, 2026 in its S-1/A.

Last updated: November 18, 2026

Key Takeaways

  • Anthropic was founded in 2021 and describes itself as a research lab whose mission is the responsible development and maintenance of advanced AI for the long-term benefit of humanity.
  • Its products are the Claude models, the Claude Platform for developers, and applications such as Claude Code, Claude Cowork and Claude.ai.
  • Usage-based revenue was $26.3 billion, or 79% of total revenue, for the nine months ended September 30, 2026, and subscriptions were $6.9 billion, or 21%.
  • Compute and infrastructure is the largest cost: $18.8 billion, or 57% of revenue, over the same nine months.
  • Anthropic is a Delaware public benefit corporation whose Class A stock was approved for listing on Nasdaq under the ticker ANT.

What Does Anthropic Do?

Anthropic develops large AI models under the Claude name and earns revenue when customers use them. Businesses build Claude into their own software and workflows, developers use it to write code, and individuals subscribe to Claude apps. Anthropic also runs a research program on AI safety, which it says shapes how its models are designed and deployed.

The company was founded in 2021 by seven researchers and had approximately 5,300 employees as of September 30, 2026, according to its S-1/A filed November 17, 2026. A majority of them work in research, engineering and product development. Its headquarters is in the San Francisco Bay Area.

Anthropic's own description is narrower than "AI company." The filing says it is "first and foremost a research lab," and its company page describes the goal as building AI systems that are reliable, interpretable and steerable.

What Products Does Anthropic Sell?

Anthropic sells access to Claude in three layers: the models themselves, a developer platform, and finished applications. Customers can pay per unit of usage, subscribe to a plan, or do both. The S-1/A describes four model classes, each tuned for a different balance of capability, cost and speed, with the same models powering every product.

ProductWhat it isTypical buyer
Claude modelsFour model classes: Mythos (including Fable), Opus, Sonnet and HaikuEnterprises and developers
Claude PlatformDevelopment environment for building, deploying and scaling applications and agents on ClaudeSoftware teams
Claude CodeCoding agent that plans, writes, tests and debugs softwareDevelopers and engineering organizations
Claude CoworkKnowledge work agent that carries out multi-step tasks on a user's desktopNon-technical professionals
Claude.ai plansSubscription tiers named Pro, Max and UltraIndividual knowledge workers

The filing also lists newer products, including Claude Security and Claude Design, and industry versions such as Claude for Financial Services and Claude for Healthcare. In September 2026, Anthropic began merging Claude Cowork and Claude.ai into one application.

Who Are Anthropic's Customers?

Anthropic's customers are mostly businesses. As of September 30, 2026, the S-1/A counted more than 11,900 customers with $100,000 or more of trailing twelve-month revenue, more than 1,900 with $1 million or more, and approximately 250 with $10 million or more. Individual subscribers, which the filing calls prosumers, are a second group.

The filing says over 50% of the Fortune 500 and over 30% of the Global 2000 were customers above the $100,000 level on that date. Growth inside the base has been fast: the $100,000 group stood at more than 950 customers one year earlier.

Distribution runs through partners as well as direct sales. Claude is available on Amazon Web Services, Google Cloud Platform and Microsoft Azure, and through systems integrators such as Accenture, Deloitte, KPMG and PwC. Anthropic says it keeps a direct contract with customers who buy through a cloud marketplace and pays the cloud provider a fee.

How Does Anthropic Make Money?

Anthropic makes money in two ways. The main one is usage-based pricing, which the company calls pay-for-intelligence: customers pay for the tokens their requests consume, at a price set by the model they choose. The second is subscriptions, where a customer pays a fixed fee for a set level of usage and features.

For the nine months ended September 30, 2026, usage-based revenue was $26.3 billion, or 79% of the total, and subscription revenue was $6.9 billion, or 21%. Sales through cloud marketplaces made up about 44% of revenue in that period.

A hypothetical example shows how the usage model scales. Suppose a company runs 10,000 coding tasks a month and each task costs $2 in tokens. The monthly bill is $20,000. If the company moves the same work to a model priced at half as much, the bill falls to $10,000, unless lower cost leads it to run more tasks. That tradeoff between price cuts and volume is a main driver of the figures on the Anthropic revenue page.

Anthropic's Financial Snapshot From the S-1/A

Anthropic's revenue has grown quickly and the company has reported a net loss in every year since its founding. The figures below come from the summary financial data in the S-1/A filed November 17, 2026.

PeriodRevenueOperating lossNet loss
2023$41 million$551 million$1.5 billion
2024$386 million$3.0 billion$8.3 billion
2025$4.6 billion$8.1 billion$42.0 billion
Nine months ended September 30, 2026$33.2 billion$1.3 billion$104.0 billion

The gap between operating loss and net loss is mostly a non-cash accounting item: changes in the fair value of convertible notes and similar instruments, which totaled $94.3 billion in the 2026 period. The page on whether Anthropic is profitable walks through each line.

Compute: Anthropic's Largest Cost

Compute is Anthropic's largest expense. Compute and infrastructure costs were $18.8 billion for the nine months ended September 30, 2026, equal to 57% of revenue, down from 172% of revenue in the same period of 2025. The S-1/A says these costs are expected to keep rising in absolute dollars.

The commitments are long dated. Anthropic disclosed purchase commitments and leases with aggregate committed spend of approximately $900.9 billion, substantially all for compute and infrastructure, with about $451.6 billion expected before December 31, 2030. It runs workloads across AWS Trainium chips, Google TPUs, NVIDIA GPUs and AMD GPUs, and treats that capacity as one shared pool for research, training and serving customers.

Who Does Anthropic Compete With?

Anthropic names Google and OpenAI as its primary competitors among companies building frontier models for developers and enterprises. The S-1/A also lists a broader group: large technology companies with their own models, specialized AI labs, and open-weight or open-source models that compete on cost for less demanding workloads.

The filing adds a less obvious competitor: inertia. Enterprises with long-established systems and workflows may be slow to adopt any AI model. A direct comparison is in Anthropic vs OpenAI, and the wider field is mapped in Anthropic competitors.

How Anthropic Is Governed

Anthropic is a Delaware public benefit corporation. Its directors must balance the financial interests of stockholders, the interests of people materially affected by the company's conduct, and a stated public benefit purpose. The guide to Anthropic as a PBC explains what that means for shareholders.

Two bodies hold control. The Long-Term Benefit Trust, a group of financially disinterested trustees, elects a majority of the board through a special Class T share. A company owned by the seven co-founders holds a Class F share with 50.1% of the general stockholder vote. Public investors buy Class A shares with one vote each. Background on the group is in the profile of Anthropic's founders.

Risks Anthropic Flags in Its Own Prospectus

The risk factors in the S-1/A are specific, and several bear directly on the business model described above.

  • Losses: the company has a history of net losses and says it may not achieve or sustain profitability.
  • Growth rate: Anthropic expects its revenue growth rate to decline over time.
  • Channel concentration: a limited number of cloud providers act as distribution and billing channels for a large share of revenue.
  • Compute dependence: most capacity is contracted 12 to 18 months ahead under multi-year commitments, and short-term capacity costs significantly more.
  • Policy violations: an estimated 5% to 7% of revenue in the 2026 nine-month period came from customers later found to be violating usage policies or commercial terms.
  • Regulation and litigation: the filing cites a $1.5 billion copyright class action settlement recorded in 2025 and an FTC civil investigative demand issued in October 2026.

Is Anthropic a Public Company?

Yes. Anthropic, PBC registered its Class A common stock with the SEC and was approved to list on the Nasdaq Global Select Market under the ticker ANT. Reuters reported on November 18, 2026 that the IPO priced at $83 per share, and trading was scheduled to begin on November 19, 2026.

The marketed deal covered 825 million Class A shares at $70 to $78, with Goldman Sachs and Morgan Stanley as lead underwriters. Public reporting means Anthropic's quarterly results, risk disclosures and ownership data are filed on SEC EDGAR, which is the source for most of the figures in this article.

Frequently Asked Questions

1. Is Anthropic the same company as Claude?

Claude is the product and Anthropic is the company. Anthropic, PBC develops the Claude family of AI models and sells them through the Claude Platform and applications such as Claude Code, Claude Cowork and Claude.ai. The stock listed on Nasdaq under the ticker ANT represents ownership in Anthropic, PBC.

2. Who founded Anthropic?

Anthropic was founded in 2021 by seven co-founders named in its S-1/A: Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah. Dario Amodei is chief executive officer and Daniela Amodei is president. Several of the co-founders previously worked at OpenAI.

3. Is Anthropic owned by Amazon or Google?

No single company owns Anthropic. Its S-1/A says Amazon and Google would hold about 14.7% and 13.7% of outstanding capital stock after the IPO, in non-voting Class N shares. Voting control rests with a company owned by the co-founders and with the Long-Term Benefit Trust, which elects a majority of the board.

4. Is Anthropic profitable?

Anthropic is not profitable under GAAP. It reported a net loss of $104.0 billion for the nine months ended September 30, 2026, most of it from non-cash fair value changes on convertible notes. Its operating loss for the period was $1.3 billion on revenue of $33.2 billion, according to the S-1/A.

5. What is Claude Code?

Claude Code is Anthropic's coding agent. According to the company's S-1/A, it works alongside software developers on end-to-end development: planning, writing, editing, testing, debugging and committing code across production codebases. Anthropic released it as a research preview in February 2025 and describes it as the clearest example of AI working as a collaborator.

6. How many employees does Anthropic have?

Anthropic had approximately 5,300 employees worldwide as of September 30, 2026, according to its S-1/A filed November 17, 2026. A majority work in research, engineering and product development. The filing reports trailing twelve-month revenue per employee of over $6.6 million on that date, up from $1.4 million a year earlier.

Conclusion

What does Anthropic do? It builds the Claude models and sells access to them, mainly to businesses that pay by usage. The business pairs very fast revenue growth with very large compute commitments and an unusual governance design. Read the prospectus sections on business, risk factors and capital stock before forming a view on ANT.

References

  1. SEC EDGAR: Anthropic, PBC, Amendment No. 3 to Form S-1 (filed November 17, 2026)
  2. Anthropic: Company page
  3. SEC EDGAR: Anthropic, PBC filing history, CIK 0002130357
  4. SEC EDGAR: Anthropic, PBC, Form S-1MEF (filed November 18, 2026)

All Guides in This Series

Disclaimer: This article is for educational and informational purposes only and is not investment, legal or tax advice, or a recommendation to buy or sell any security. Leveraged and inverse ETFs are complex products intended for short-term use by experienced traders, and investors can lose their entire investment. Read a fund's prospectus before investing. WOLF Financial is a digital marketing agency, not a registered investment adviser or broker-dealer.

KEEP READING

MORE INSIGHTS.

READ MORE
More insights
2x Short Anthropic ETF List: Every Inverse ANT Fund
Anthropic IPO & leveraged ETFs
2x Short Anthropic ETF List: Every Inverse ANT Fund
Seven inverse Anthropic ETFs by issuer: ANDD, ANS, ANSS, ANTD, ANTQ, ANTY and ANTZ, with exchange, expense ratio, launch status and risks.
Read more
Read more
Anthropic Stock vs ETF: ANT Shares or a Single-Stock Fund?
Anthropic IPO & leveraged ETFs
Anthropic Stock vs ETF: ANT Shares or a Single-Stock Fund?
ANT shares give direct 1x ownership of Anthropic. Single-stock ETFs such as ANTT and ANTZ give 2x daily exposure for a fee. Costs and risks compared.
Read more
Read more
2x Long Anthropic ETF List: Every Ticker by Issuer
Anthropic IPO & leveraged ETFs
2x Long Anthropic ETF List: Every Ticker by Issuer
Eight 2x long Anthropic ETFs by issuer: AIL, ANTC, ANTO, ANTT, ANTU, ANUU, ATHX and CLAU, with exchange, expense ratio and launch status.
Read more
Read more