FundraiseOct 28, 2026, 12:48 UTC

Anthropic files for IPO that could value it at $2T, raise over $100B

What's the deal? Anthropic filed its S-1 with the US Securities and Exchange Commission on October 26, 2026, setting up what could be the largest initial public offering ever. The AI company plans to list its Class A common stock on the Nasdaq Global Select Market under the symbol "ANT," with Goldman Sachs and Morgan Stanley leading the offering. Share count and pricing are not yet set, but the listing could value Anthropic at up to $2 trillion and raise more than $100 billion.

The numbers: Revenue hit $16.3 billion in the first half of 2026, up 14-fold year on year, after jumping from $41 million in 2023 to $4.6 billion in 2025. The company reported a net loss of $80.7 billion for the half — but roughly $72.4 billion was a non-cash charge from marking convertible notes to fair value as its valuation soared. Its operating loss narrowed to about $693 million, and it generated roughly $3.9 billion in positive operating cash flow.

Standout stat: Net revenue retention stood at around 900% as of September 30 — existing customers spending ninefold more year on year. The median public cloud software company sits at 108%. Anthropic now has more than 11,900 enterprise customers paying over $100,000 annually, up from about 950 a year earlier.

Who wins? Amazon is likely the single largest beneficiary. It holds convertible notes with $6.2 billion in principal outstanding, which its own filings valued at $97.9 billion as of June 30. Google is also set for a windfall. Venture backers including Menlo Ventures, Sequoia, Sound Ventures, and Spark Capital stand to return billions, though none owns 5% — a quirk of Anthropic's scale, where a 1% stake is worth nearly $10 billion at its $965 billion private valuation.

What could go wrong? Anthropic has committed up to $807 billion to data centre, chip, and computing contracts over 10 years, many of which cannot be cancelled. It warned it may need additional financing sooner than anticipated, even after the IPO. Its top five customers account for 25% of revenue, and the company itself cautioned that growth "may not be sustainable over time."

Why now? Anthropic is going public into a jittery market. AI data centre operator Firmus pulled its IPO this month, and SK Hynix scrapped plans to list its Solidigm unit, both citing market conditions.

How analysts read it: NYU finance professor Aswath Damodaran, working from the S-1, values Anthropic's equity at about $1.475 trillion, at the low end of the $1.5-2 trillion pricing range. His base case assumes revenue approaching $1 trillion by 2036 (including a possible push into drug development), a 27.5% operating margin after a regulatory drag, $1 of reinvestment for every $3 of new revenue, a 9.74% cost of capital and a 10% chance of failure, with the present value of compute commitments counted as debt. His scenarios run from negative equity (-$373 billion) in a worst case to $4.04 trillion in a best case. He flags that 1,900 enterprises now spend more than $1 million a year on Claude, up from fewer than 100 a year earlier, but that growth slowed in the third quarter, and he sharply criticises the governance: five share classes, founders holding 50.1% of votes, a benefit trust picking the board, and the public benefit corporation structure. Insiders other than founders can sell up to 10% of their shares on day one, with the rest unlocking over 180 days and founders after a year. He says he will sit out if the IPO prices well above his value.

Chart: Aswath Damodaran, Musings on Markets

Snowflake CEO Sridhar Ramaswamy, in a widely shared thread, reads the same S-1 as a capacity bet rather than a bank-run risk: $265.6 billion of future fixed or minimum purchases and $321.4 billion of leases not yet started as of June 2026, set against compute expense falling to 55% of sales from 201% a year earlier. "The core question is whether demand keeps outgrowing committed cost," he writes; "a capacity bust is more likely than a bank run."

Update — 2 November 2026: Anthropic filed an amendment to the S-1 adding its third-quarter 2026 financials and, for the first time, a total figure for its compute commitments. Q3 revenue was $16.96 billion, up 1,127% year on year, but quarter-on-quarter growth slowed sharply to about 47% (from 144% in Q2). The company swung to a $567 million operating loss, reversing a $974 million operating profit in Q2, driven mainly by $5 billion of compute bought under costlier short-term deals rather than cheaper multi-year contracts, plus about $2 billion of non-cash donation-matching charges. Net loss for the quarter was $23.3 billion, almost all (about $21.9 billion) a non-cash fair-value charge on convertible notes, derivatives and SAFEs tied to the rise in Anthropic's valuation, not a cash cost. The amendment also disclosed aggregate purchase commitments and leases of approximately $895.4 billion as of September 30 — nearly all for compute and infrastructure — with about $447.5 billion due before the end of 2030 and roughly $69.8 billion due within the following 12 months. About $213.8 billion of that total still carries termination rights or unmet conditions such as financing or regulatory approval. Liquidity stood at $146.6 billion in cash, short-term investments and an available credit line. Anthropic still has not set an IPO price or date.

The signal: The filing offers the clearest picture yet of a frontier AI company's economics — explosive revenue, a widening but largely non-cash loss, and a compute bill now quantified at nearly $900 billion. It also sets a benchmark for rival OpenAI, whose own IPO is expected in 2027. The open question remains whether any growth projection can justify a valuation in the trillions.

Image credit: Generated with Gemini

Anthropic files for IPO that could value it at $2T, raise over $100B

More top stories

Databento lands $100M Series C at $1.1B, four months after Series B

Tyme Group hits unicorn status with $250M Series D led by Nubank

Revolut, now at $115B, eyes 2028 IPO and slower hiring as AI spreads