Anthropic wants to be worth $2 trillion. Its own IPO filing shows why that’s such a big bet

Anthropic is heading for what could be the largest stock market debut ever, and this week investors got their first real look inside it. Details from the Claude maker’s confidential prospectus, reported by Reuters, show a company growing faster than almost any business in history while losing tens of billions of dollars, signing up for half a trillion dollars of computing, and warning that its own technology could pose “catastrophic or existential risks to humanity.” Wall Street is split on whether that adds up to a $2 trillion company.

When is the Anthropic IPO?

There’s still no date, ticker or price. Anthropic confidentially submitted a draft S-1 to the US Securities and Exchange Commission on June 1, saying the timing “will depend on market conditions and other factors”, and it has declined to comment since.

The timetable has slipped once already. Reuters reported in September that the roadshow, when the company pitches its shares to big investors, would start in mid-October at the earliest, with Morgan Stanley, Goldman Sachs, JPMorgan and Citigroup running the sale. The listing was first pencilled in for days before the November 3 US midterms; Reuters’ reporting on the prospectus this week said it would more likely come after them. Most reports expect it on the Nasdaq, raising up to $100 billion at a valuation above $2 trillion, more than double the $965 billion the company was valued at in May.

Traders are betting it happens soon. On prediction market Kalshi, the odds of Anthropic announcing its IPO this year stood at about 78% on Wednesday, CNBC reported, and the chance of an announcement before November jumped from under 5% on Monday to 16% once the prospectus details came out.

What the IPO filing shows

The confidential prospectus hasn’t been published, but Reuters has reported its key figures (we covered them in detail in our story on the filing):

• Revenue: up 12-fold in 2025, to nearly $4.6 billion.
• Losses: a $42 billion net loss, of which about $34 billion is an accounting charge, plus an operating loss of more than $8 billion.
• Commitments: $518 billion for cloud, computing and infrastructure, around 80% of it non-cancellable, against about $20 billion in cash at the end of 2025.
• Customers: two customers brought in nearly a quarter of revenue, and 47% of sales went through Amazon’s and Google’s cloud marketplaces.
• Risks: nearly a third of the document, including warnings that its models can “resist shutdown” and behave in ways “resembling blackmail”.

This year’s numbers look very different

The prospectus figures are for 2025, and Anthropic has grown much faster since. The company told investors that its revenue more than doubled to $11.6 billion in the three months to June, up from $4.73 billion in the first quarter and $787 million a year earlier, the Wall Street Journal reported in August. That single quarter was more than twice its revenue for the whole of 2025, and put it ahead of OpenAI, which told its own investors it made $6.7 billion in the same period.

Anthropic also told investors it made a small adjusted operating profit in the quarter, thanks to using its computing power more efficiently. The Journal noted that the company hasn’t explained exactly how it calculates that figure, and that earlier updates to investors left out the cost of paying staff in shares. OpenAI’s operating loss, by contrast, widened to $12.3 billion. Much of Anthropic’s growth has come from businesses paying for Claude, and especially Claude Code, its coding tool.

At that pace Anthropic would bring in more than $45 billion a year, which makes the $2 trillion price tag look less outlandish: roughly 40 to 45 times current annualised sales, rather than more than 400 times last year’s.

Where the $518 billion goes

Anthropic doesn’t own most of the chips it runs on. It rents them, mostly from the same tech giants that invest in it and compete with it. The prospectus breaks down the biggest commitments, according to reporting based on the filing Reuters saw:

About 80% of the $518 billion total is non-cancellable. If Anthropic uses less than it has promised Google or Amazon, it still has to pay the difference.

That’s the heart of the debate. If revenue keeps doubling, those bills look like the cost of keeping up with demand. If growth stalls, Anthropic is locked into paying for computing it doesn’t need, and the companies on the other side of those contracts, several of them among the biggest on the stock market, would feel it too.

The bull case: “perhaps the greatest business story”

The growth is the pitch. A 12-fold jump in a single year is almost unheard of for a company already earning billions, and Wall Street noticed: European tech shares rose 2.4% and the US chip index 1.5% after the first details came out, Reuters reported. Morningstar’s Michael Field reckoned the company would trade at around 18 to 20 times sales, which he called “much more reasonable than the recent SpaceX IPO.” Dan Coatsworth of AJ Bell pointed to the “stellar revenue growth” as the reason investors were paying attention.

Even sceptics of the AI boom are impressed. Hedge fund manager Bill Ackman told Bloomberg TV on Wednesday:

Anthropic is perhaps the greatest business story I’ve ever seen.

Bill Ackman, Pershing Square

There’s a timing argument too. Axios argued that Anthropic needs the money, could have the market to itself, and has seen some of its political risk ease since it signed the White House’s voluntary AI accord on Tuesday. A public listing would also give it shares to use as currency for deals and hiring, and pile pressure on OpenAI.

The bear case: the losses, the bills and the warnings

The same filing gives sceptics plenty to work with. Anthropic has committed to spending roughly 25 times its year-end cash on computing, most of it locked in whether or not demand keeps growing. It depends heavily on Amazon and Google, which are its investors, suppliers and rivals all at once. And it is now one of two AI labs under investigation by the US Federal Trade Commission, while still fighting the Pentagon, which an appeals court has allowed to keep Anthropic on its supply-chain blacklist.

“There are a lot of question marks around these valuations,” Newton Jones of Steward Partners told Reuters, pointing to “companies that are losing billions of dollars.” And with nearly a quarter of 2025 revenue coming from two unnamed customers, losing one big client would hurt far more than it would at a typical software company.

Ackman himself won’t buy, saying his fund avoids fast-growing businesses “consuming a huge amount of capital”, and questioned whether frontier labs can hold their ground against cheaper open-weight models. Michael Burry, the investor famous for betting against the 2008 housing bubble, went much further in a post on X:

For the benefit of humanity, the markets should tank hard and prevent the OpenAI and Anthropic IPOs.

Michael Burry, investor

Central bankers are worried about the wider picture. On Thursday Bank of England governor Andrew Bailey said that in the AI boom “everybody is currently priced to be a winner”, and that history suggests otherwise. His bank has warned that AI shares could fall much harder than they did in July.

Who would be in charge?

Not shareholders. According to the filing, a “Founder LLC” owned by Anthropic’s seven co-founders will hold a single special share carrying 50.1% of the votes on key matters. Anthropic will stay a public benefit corporation, and its Long-Term Benefit Trust, set up to keep it focused on safety, will elect four directors. The company warns this could lead to decisions “that may conflict with short-, medium-, or long-term financial interests.”

That’s unusual for a company asking the public for up to $100 billion. Founders keeping control isn’t new (Meta and Alphabet do it too), but few prospectuses tell investors so plainly that the company may put safety ahead of their returns.

The board itself is a heavyweight one. Daniela Amodei, Anthropic’s president, chairs it, and its members include Netflix co-founder Reed Hastings, Novartis chief executive Vas Narasimhan and former Microsoft finance chief Chris Liddell, according to a Reuters factbox on the filing. The trust that picks four directors includes former Federal Reserve chair Ben Bernanke and national security expert Richard Fontaine, and is chaired by Neil Buddy Shah, head of the Clinton Health Access Initiative.

Some of Anthropic’s own early staff have been eyeing remote land to escape to if AI goes wrong, which gives a sense of how seriously the people running it take the risks they’re listing.

The race with OpenAI

OpenAI also filed confidentially for an IPO in June, according to Reuters, but Anthropic now looks likely to get there first: Kalshi traders put the odds of Anthropic listing before OpenAI at around 93%. Going first matters. The first frontier lab on the market will set the price investors put on the whole sector, and will raise its money before the second one competes for the same funds.

The last mega-listing is a useful guide. When SpaceX went public in June, US brokers including Charles Schwab, Fidelity, Robinhood, SoFi and E*Trade offered IPO shares to ordinary customers, but demand was so high that many got only a fraction of what they asked for.

What happens next

If Anthropic follows the usual path, the next few weeks look like this:

1. Public filing: the prospectus appears on the SEC’s EDGAR website at least 15 days before the roadshow starts. That’s when the ticker, exchange and full financials become official.
2. Roadshow: a week or two of meetings with big investors, starting mid-October at the earliest, with a price range published at the start.
3. Pricing: the final share price is set the evening before trading, based on demand.
4. First trade: the shares start trading the next morning, and anyone with a suitable brokerage account can buy them.
5. Lock-up: staff and early investors usually can’t sell for around six months, which is often when a newly listed share price wobbles.

Can you buy Anthropic stock?

Not yet. Anthropic is still private, so its shares don’t trade on any exchange. When it lists, there are two ways in:

• At the IPO price: only if Anthropic sets shares aside for ordinary investors, which it hasn’t said it will. If it follows SpaceX, they would go through US brokers, usually to US customers.
• Once it’s trading: anyone with an account that offers US shares can buy, including investors in the UK, Canada, Australia, Germany and the rest of Europe. In India, purchases count towards the RBI’s $250,000-a-year overseas investment limit.

Be careful before then. On Tuesday the SEC charged a fund adviser with defrauding investors in pre-IPO funds holding stakes in SpaceX and other companies, warning that “fraudsters can exploit the allure of exclusive, high-return pre-IPO access”. Anyone offering you Anthropic shares today is almost certainly not legitimate. This isn’t financial advice.

Frequently asked questions

When will Anthropic go public?

There’s no confirmed date. Reuters has reported that Anthropic will start marketing its shares in mid-October at the earliest, with a listing expected around or after the November 3 US midterm elections. Betting market Kalshi puts the chance of an announcement this year at about 78%.

Is Anthropic publicly traded?

No. Anthropic is still a private company, so you can’t buy its shares on any stock exchange yet. It confidentially filed for an IPO with the SEC on June 1, 2026.

What will the Anthropic IPO price per share be?

It hasn’t been set. A price range will appear in the public prospectus at the start of the roadshow, and the final price is set the night before trading. Reports put the company’s valuation at more than $2 trillion.

What is Anthropic’s ticker symbol?

Anthropic doesn’t have one yet. It will be revealed in the public prospectus. Most reports expect a Nasdaq listing. Any app or site already showing an “Anthropic stock price” isn’t showing a real one.

How do I buy Anthropic stock at the IPO?

Only through a broker that gets an allocation, and only if Anthropic sets shares aside for retail investors, which it hasn’t said it will. When SpaceX listed in June, US customers of Robinhood, Fidelity, Schwab, SoFi and E*Trade could request IPO shares. Once Anthropic starts trading, you can buy it on any of them.

How to invest in Anthropic before the IPO?

Ordinary investors realistically can’t. Pre-IPO stakes trade privately, mostly through funds open only to accredited investors, and Anthropic restricts share transfers. Unsolicited offers of Anthropic shares are very likely scams.

Who owns Anthropic?

Its founders, staff and investors. Amazon and Google are both big investors, but neither controls it. After the IPO, a Founder LLC owned by the seven co-founders will hold 50.1% of the votes on key matters, according to the prospectus. The exact stakes will be listed in the public filing.

Does Amazon own Anthropic?

No. Amazon owns a minority stake, as does Google, and both supply much of Anthropic’s computing power. Owning Amazon or Alphabet shares gives you only a small, indirect slice of Anthropic.

Is there an Anthropic ETF?

No ETF tracks Anthropic on its own. Some funds hold small private stakes, and index and tech ETFs may add the shares once it lists and becomes eligible.

How can I buy Anthropic shares from India, the UK or Australia?

Wait for it to list, then buy through a platform that offers US shares. IPO-price shares usually go only to customers of US brokers. In India, overseas share purchases count towards the RBI’s $250,000-a-year Liberalised Remittance Scheme limit.

Is Anthropic profitable?

Not on paper yet. Its prospectus shows a net loss of about $42 billion in 2025, most of it an accounting charge, on revenue of nearly $4.6 billion. It told investors it made a small adjusted operating profit in the second quarter of 2026, according to the Wall Street Journal.

What is Anthropic’s revenue in 2026?

Anthropic told investors it made $4.73 billion in the first quarter of 2026 and $11.6 billion in the second, according to the Wall Street Journal. That puts it on track for more than $45 billion a year.

Will Anthropic go public before OpenAI?

It looks likely. OpenAI has also filed confidentially, but Kalshi traders give Anthropic around 93% odds of listing first.

What is a public benefit corporation?

A company legally required to weigh a public mission alongside profits. Anthropic’s mission is the safe development of AI, and it will stay a public benefit corporation after listing, which its filing warns could lead to decisions that conflict with shareholders’ financial interests.

MadRobot’s take

Anthropic’s filing is the most honest IPO pitch a big tech company has made in years, and that’s exactly what makes it hard to value. Its growth is real, but a $2 trillion price assumes it becomes one of the most profitable companies on earth while paying off more than half a trillion dollars of computing bills and competing with the giants that supply them. Investors also have to accept that, by design, they won’t control it.

The sharpest point in the whole debate comes from the company itself: it says it must build more powerful AI to make AI safe, while warning that more powerful AI is more dangerous. Buying Anthropic shares means betting that it can keep both of those promises at once. Anyone thinking of buying on day one should read the risk factors first. Anthropic clearly wants them read.

Sources: Anthropic, “Anthropic confidentially submits draft S-1 to the SEC”; Reuters (prospectus, board factbox, market reaction, IPO timing); Bloomberg TV interview with Bill Ackman; Michael Burry on X; Axios; Wall Street Journal (2026 revenue, via Yahoo Finance); CNBC (Kalshi odds); SEC press release 2026-98.