Direct answer: Anthropic, the company behind the Claude family of AI models, has filed its S-1 registration statement with the US Securities and Exchange Commission and plans to list on the Nasdaq stock exchange under the ticker “ANT”. The filing reveals $16.3 billion of revenue in the first half of 2026, up from $1.2 billion in the same period of 2025, alongside a net loss of $80.7 billion driven largely by non-cash accounting charges, and a governance structure that leaves its seven co-founders holding 50.1% of the votes.
What happened
Anthropic filed its public S-1 on Monday 26 October 2026, after weeks of leaked draft filings circulating in the press. The public filing is the formal step that turns months of speculation about one of the most closely watched technology listings in years into a regulated, public process. From here, the company must answer the SEC’s questions, publish a price range, run an investor roadshow, and only then set a final price and start trading.
The filing itself does not yet say how many shares Anthropic will sell, what price it hopes to achieve, or how much money it plans to raise. Those details normally arrive in later amendments to the filing, once the company and its bankers have tested investor appetite. What the filing does give, for the first time in audited, public form, is a detailed picture of how fast Anthropic is growing, where its money comes from, and how much control its founders intend to keep.
TechCrunch reported the filing on the day it landed, and the company’s own announcement confirmed that a registration statement had been filed but had not yet become effective, which means no shares can be sold until the SEC process completes. Goldman Sachs and Morgan Stanley were named first among 24 underwriters, according to Decrypt’s report on the filing, a roster that underlines how large bankers expect this deal to be.
The timing matters. Anthropic had already filed a confidential draft with the SEC in June 2026, a route that lets companies work through regulators’ questions in private before going public. Reuters reporting cited in the coverage said backers had hoped for a valuation above $2 trillion, while the Financial Times, also cited in coverage of the filing, put investor expectations in the $1.5 trillion to $1.75 trillion range. Those figures are reported expectations, not numbers in the filing itself, and they should be treated as such until a price range is published. What is certain from the filing is the scale of the business underneath: in May 2026 Anthropic raised $65 billion at a $965 billion valuation, a round that already made it one of the most valuable private companies in the world.
The numbers behind the filing
The headline figures in the S-1 are, by any normal business standard, extraordinary. Revenue for the six months to 30 June 2026 was $16.29 billion, against $1.16 billion in the same period a year earlier. Full-year revenue for 2025 was $4.59 billion, itself up from $386 million in 2024 and $41 million in 2023. In other words, Anthropic’s revenue has multiplied roughly a hundredfold in about two and a half years, a trajectory that has very few precedents in the history of enterprise software.
The loss figures need more care, because the biggest number in the filing is also the most misleading if read on its own. Anthropic reported a net loss of $80.7 billion for the first half of 2026. Most of that, about $72.4 billion, was a non-cash charge from changes in the fair value of convertible notes and other instruments, driven primarily by the rise in Anthropic’s own valuation. When a company’s value climbs sharply, the accounting value of the convertible debt it has issued climbs too, and that increase shows up as a paper loss even though no cash has left the business. Strip that charge out and the operating picture looks very different: an operating loss of $693 million for the first half of 2026, a sharp improvement on the $2.1 billion operating loss recorded in the first half of 2025.
That improvement is the number investors will focus on, because it suggests the business is moving toward covering its day-to-day costs from revenue, even while it continues to spend heavily on computing power, research staff, and infrastructure. Operating cash flow, according to one analysis of the filing, turned positive in the period, though readers should wait for the filing’s own cash-flow statement to be digested by mainstream financial coverage before treating any single derived figure as settled.
The customer numbers tell the same story of rapid enterprise adoption. As of the end of September 2026, Anthropic had more than 11,900 customers spending $100,000 or more a year, 1,900 customers spending $1 million or more, and about 250 customers spending $10 million or more. That is not a consumer app with a long tail of free users. It is a supplier to large organisations, selling through direct deals and through the cloud marketplaces of Amazon and Google, both of which are also major investors and infrastructure partners. That concentration is a strength in revenue terms and a risk in strategic terms, and it is exactly the kind of detail the IPO process is designed to force into the open.
How Anthropic makes its money
Anthropic’s business rests on the Claude family of models, sold principally as an application programming interface and through enterprise products. Developers and companies pay to send text, documents, images and code to Claude and receive answers, summaries, analysis and actions in return, and the price scales with how much work the model does. Around that core sit products such as Claude Code, an agentic coding tool that writes, tests and edits software under a developer’s direction, which has become one of the company’s fastest-growing lines.
The model for a business like this is closer to a utility than to traditional software. Customers do not buy a licence once and use it forever. They pay for consumption, and their bills rise as they build more of their operations around the model. That is why the customer-spend tiers in the filing matter so much: a company that starts with a small team experimenting with Claude can grow, within a year or two, into a seven-figure annual account, and the filing shows that migration happening at scale.
It is also why computing cost is the central question for the whole sector. Every answer Claude gives consumes processing power, and the company must buy or rent that power from chip suppliers and cloud providers. Revenue growing faster than computing cost is the operating leverage story that would justify a very large valuation. Revenue growing only because computing spend grows in lockstep would not. The S-1 gives investors the raw material to test that question, and the debate over gross margins, which one analyst flagged publicly before the filing, will now play out with real numbers rather than estimates.
For context, Anthropic sits in a small group of frontier AI laboratories that train the largest general-purpose models, alongside OpenAI, Google DeepMind, Meta and xAI. Each has a different funding model: OpenAI has moved toward a for-profit structure with huge infrastructure commitments, Google and Meta fund their labs from advertising and platform profits, and xAI leans on its sister companies. Anthropic’s IPO would make it the first of the pure-play frontier labs to submit itself fully to public-market discipline, with quarterly reporting, shareholder votes and a share price that moves on every earnings release.
The governance fight in the filing
The most controversial part of the S-1 is not a financial figure at all. It is who gets to run the company after it goes public. Anthropic’s seven co-founders, Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah, will form a Founder LLC that collectively controls 50.1% of shareholder voting power. Separately, the company’s Long-Term Benefit Trust, established in 2023 to ensure the company pursues its public benefit mission, will have the authority to appoint most of the board.
The stated purpose is to preserve Anthropic’s safety mission. The company has argued since its founding that building increasingly powerful AI systems carries risks that ordinary corporate governance, driven by quarterly earnings pressure, may not handle well. A founder bloc with majority voting power, combined with a trust whose mandate is the company’s long-term benefit rather than short-term shareholder returns, is designed to insulate safety decisions, such as whether to delay or restrict a model release, from market pressure.
Critics, including those quoted in TechCrunch’s coverage, argue the opposite: that it means public shareholders will have little practical control over a company building technology that, by its own admission in the filing, poses catastrophic or existential risks. Voting power will not be proportional to economic ownership, so investors who supply most of the capital after the IPO will not be able to vote out directors or block major decisions in the usual way. That tension, between a safety mission that asks to be trusted and a public market that expects accountability, is likely to dominate the roadshow and the early analyst coverage.
It also has a practical consequence for anyone who uses Claude. Decisions about model behaviour, safety restrictions, pricing and product direction will continue to be made by a leadership team that is unusually insulated from shareholder pressure, for better or worse. Users who value stability and a safety-first approach may see that as reassurance. Users who want faster feature releases or lower prices may see it as a frustration. Either way, the structure is now public and cannot be quietly changed later without a major governance fight.
What it means for UK small businesses
For a small business in the UK, an American IPO filing can feel remote. This one is not, because Claude and its rivals are rapidly becoming part of the everyday toolkit for firms that would never have described themselves as technology companies: accountants summarising client documents, estate agents drafting listings, clinics triaging admin, trades firms quoting jobs, and online sellers writing product descriptions and answering customers.
First, the filing is a signal about staying power. A company reporting $16.3 billion in half-year revenue, with thousands of large enterprise customers, is not going to disappear overnight, and its products are unlikely to be abandoned without a migration path. For a small firm choosing which AI tools to build processes around, that durability matters more than any single feature, because the real cost of AI adoption is not the subscription. It is the time spent training staff, rewriting templates, and embedding a tool into daily work, all of which is wasted if the supplier folds or pivots.
Second, public-company status will make Anthropic’s pricing and product decisions more visible, but not necessarily more stable. Public markets reward growth, and growth in this sector has so far come from selling more computing power to larger customers. Small businesses should expect enterprise features to arrive first, consumer and small-business pricing to follow, and occasional price or limit changes as the company balances growth against the margins investors demand. The practical response is to avoid building a business that depends on a single model at a single price point. Use tools that let you switch providers, keep your own data and prompts portable, and test alternatives, including Google’s and OpenAI’s offerings, before committing a core workflow to one supplier.
Third, the governance debate has a direct UK angle. The UK’s AI regulation debate, including the long-running discussion over an AI safety bill and the work of the AI Security Institute, has repeatedly returned to the same question Anthropic’s structure raises: who decides when a powerful model is safe enough to release, and who can overrule them? A public Anthropic will be watched by UK regulators, enterprise buyers and the press in a way a private one was not, and its disclosures, including its risk section warning of catastrophic or existential risks, will become reference material in that debate. Small firms will not be asked to solve that question, but they will live with its answers, in the form of model restrictions, content policies and compliance expectations that flow down from the frontier labs to every business user.
Fourth, there is a competitive effect worth watching. If Anthropic’s IPO succeeds at a very large valuation, it will make it easier for other AI companies to raise money and to go public, which tends to increase investment in tools, integrations and support for smaller customers over time. If it struggles, or if investors punish the governance structure with a lower price, it will be read as a verdict on the whole sector’s economics, and spending may tighten. Either way, the IPO is a useful barometer: it is the first time public investors, rather than a small circle of venture funds and tech giants, get to price the economics of frontier AI.
Finally, there is a simple procurement lesson. The filing shows that Anthropic’s largest customers are concentrated in the enterprise tier, and its revenue depends heavily on cloud marketplaces. Small businesses buying through those marketplaces, or through resellers, should check exactly what they are paying for: model version, usage limits, data retention terms and support levels. The cheapest headline price is rarely the cheapest total cost once usage, retraining and support are counted.
Honest limits / what to watch
This post is based on the filing as reported on the day it was published, and there are important things it does not yet establish. The share price, the number of shares on offer, and the amount Anthropic hopes to raise have not been disclosed, so any valuation figure quoted in coverage, whether $1.5 trillion, $1.75 trillion or above $2 trillion, is a reported expectation, not a fact from the filing. Readers should treat those numbers as speculation until a price range is published.
The net loss figure also deserves caution. The $80.7 billion headline is dominated by a non-cash accounting charge, and it would be wrong to read it as Anthropic having spent $80.7 billion in cash. Equally, it would be wrong to dismiss it entirely: the charge reflects the rising value of convertible instruments, and the conversion terms of those instruments, including large holdings by Amazon and Google, will affect how much of the company public shareholders actually own. The details of those conversions will be pored over by analysts in the coming weeks, and early summaries may be revised.
The governance structure is also untested at this scale. No public company of this size has combined a founder voting bloc above 50% with a trust that appoints most of the board. Whether index providers include the shares, whether large pension funds accept the structure, and whether the trust’s mandate survives its first serious disagreement with management are all open questions. The answers will shape not just Anthropic but the governance templates other AI companies copy.
There is also the wider market context. Investor appetite for AI infrastructure spending has become more selective during 2026, with questions about whether the enormous capital expenditure on chips and data centres will generate matching returns. Anthropic’s filing lands in that debate, not apart from it. A successful IPO would not settle the debate, and a difficult one would not end the AI boom, but both would move it.
What to watch next is straightforward. First, the amended filings that add a price range and share count. Second, the roadshow, where management will have to defend both the economics and the governance structure to institutional investors. Third, the first quarterly results as a public company, which will show whether the operating-loss improvement in the first half of 2026 continues. Fourth, any response from regulators or index providers to the voting structure. And fifth, product announcements aimed at smaller businesses, which would show whether Anthropic intends to compete for the long tail of customers or remain primarily an enterprise supplier.
Key facts
| Fact | Detail |
|---|---|
| Filing | Form S-1 with the US SEC, made public 26 October 2026 |
| Planned listing | Nasdaq Global Select Market, ticker ANT |
| First-half 2026 revenue | $16.29 billion, up from $1.16 billion a year earlier |
| Full-year 2025 revenue | $4.59 billion, up from $386 million in 2024 |
| First-half 2026 operating loss | $693 million, improved from $2.1 billion a year earlier |
| First-half 2026 net loss | $80.7 billion, largely a non-cash convertible-debt charge of about $72.4 billion |
| Large customers (Sept 2026) | 11,900+ spending $100,000+ a year; 1,900+ spending $1m+; about 250 spending $10m+ |
| Founder voting control | Founder LLC holds 50.1% of shareholder voting power |
| Board control | Long-Term Benefit Trust appoints most of the board |
| Price and share count | Not yet disclosed |
| Lead underwriters named first | Goldman Sachs and Morgan Stanley, among 24 underwriters |
| Prior private valuation | $965 billion after a $65 billion raise in May 2026 |
Timeline
- 2021: Anthropic founded by former OpenAI researchers, with a public-benefit mission and a focus on AI safety.
- 2023: Long-Term Benefit Trust established to safeguard the company’s mission.
- 2024: Revenue reaches $386 million for the full year.
- 2025: Revenue reaches $4.59 billion for the full year.
- May 2026: Anthropic raises $65 billion at a $965 billion valuation.
- June 2026: Confidential draft S-1 submitted to the SEC.
- 26 October 2026: Public S-1 filed; Nasdaq listing under ANT planned; price range and share count not yet disclosed.
FAQs
Has Anthropic gone public yet?
No. Anthropic has filed its S-1 registration statement with the SEC, which starts the IPO process. It has not yet set a share price, a share count, or a listing date, so no shares are trading yet.
What ticker will Anthropic trade under?
Anthropic plans to list its Class A common stock on the Nasdaq Global Select Market under the ticker ANT, according to its filing.
Is Anthropic profitable?
Not on a net basis. It reported an operating loss of $693 million for the first half of 2026, improved from $2.1 billion a year earlier, and a net loss of $80.7 billion driven largely by non-cash accounting charges tied to convertible debt.
Who will control Anthropic after the IPO?
Its seven co-founders will collectively hold 50.1% of shareholder voting power through a Founder LLC, and its Long-Term Benefit Trust will have the authority to appoint most of the board.
What does the Anthropic IPO mean for small businesses?
It signals that Claude and enterprise AI tools are now a mainstream, revenue-backed market rather than an experiment, but it also means pricing, product direction and safety rules will increasingly be set in public, under investor scrutiny.
Where can I read the original reporting?
TechCrunch reported the filing on 26 October 2026, Decrypt covered the listing details and underwriters, and Anthropic itself announced that a registration statement had been filed. Links to the original sources are below.
Sources: TechCrunch: Anthropic officially files for its public debut, Decrypt: Anthropic makes IPO filing public, plans Nasdaq listing under ANT, CoinDesk: AI giant Anthropic files for blockbuster IPO
Related reading on WeblifyAI: our earlier coverage of Anthropic’s warning about AI models that may soon self-replicate and Claude Duet, Anthropic’s agents app.
