Anthropic's IPO Filing Reveals $518B Compute Bill
By @sharedot · · 6 pages
- Stock Market
- Anthropic
- IPO
- AI
Anthropic's confidential IPO prospectus shows revenue jumping twelvefold to $4.6 billion against an $8 billion operating loss and a $518 billion compute commitment.
The numbers behind a potential $2 trillion listing
Anthropic's long-awaited IPO prospectus, first reported by Reuters and covered in detail by InvestmentNews, shows revenue rising roughly twelvefold last year to nearly $4.6 billion — but the maker of the Claude models lost more than $8 billion from operations in 2025.
Total operating expenses hit $12.65 billion, with compute and infrastructure alone costing $7.33 billion, roughly triple the 2024 level.
The company ended the year with $20.28 billion in cash, cash equivalents and short-term investments as it prepares for a listing that could value it above $2 trillion.
A $518 billion bet it mostly cannot cancel
The filing's most striking number is the commitment: at least $518 billion in computing infrastructure spending over the next decade, with about 80% payable regardless of how much compute Anthropic actually uses.
According to InvestmentNews, obligations include at least $111.1 billion owed to Google between April 2026 and July 2033, $110 billion to Amazon through April 2036, $31.4 billion to Microsoft, and roughly $161.2 billion in Broadcom-related equipment lease obligations that neither side can cancel except on default.
Anthropic said if its actual spend falls short, it must pay Google and Amazon the difference.
Concentration risk and tangled alliances
InvestmentNews reports the prospectus flags a sharp revenue concentration risk: two customers generated nearly a quarter of 2025 revenue, and many of Anthropic's largest clients have no long-term contracts and could reduce or stop spending.
The filing is also candid about conflicts — Amazon, Google and Microsoft each act as investor, customer, cloud provider, distributor and competitor at once, creating incentives Anthropic said "may not be fully aligned" with its own.
In response, the company is shifting from a cloud-only model toward dedicated data centers and directly leased chips.
Existential warnings and a post-midterm debut
Tech Xplore, citing a Financial Times report on the S-1, says Anthropic warns investors of potential "existential risks to humanity," flagging that powerful AI models could operate beyond their predicted parameters despite safety controls.
The offering, expected in the coming weeks, could value Anthropic at around $2 trillion — making it the biggest public stock sale of all time and more than double the roughly $965 billion it was estimated to be worth in May, per InvestmentNews, which adds Reuters reporting that the deal likely lands after the November midterms.
Whichever leading AI lab lists first is expected to set valuation benchmarks for the sector.