Newsletter
Anthropic is making a big bet that it isn’t entirely sure about itself. Its IPO prospectus, reviewed by Reuters, points to a listing that could value the AI company at more than $2 trillion.
What’s happening: that valuation would surpass the record $1.77 trillion valuation Elon Musk’s SpaceX reached in its IPO. But there’s a catch.
Some numbers:
- 2025 revenue: $4.59 billion, up 1,088% year on year
- Operating loss: $8.06 billion
- Compute and infrastructure spending: $7.33 billion, up 190%
- Cash and short-term investments: $20.28 billion
- Future cloud, compute and infrastructure obligations: roughly $518 billion
Those future obligations are more than 100 times Anthropic’s 2025 revenue. Nearly a quarter of that revenue came from two customers, and the company warns that many of its largest clients are not tied to long-term contracts. Two of its biggest backers, Amazon and Google, also provide the cloud infrastructure it pays for.
An IPO would mark a rapid rise for a company founded just five years ago. It could also set a benchmark for how Wall Street values leading AI companies, including rival OpenAI.
The financial figures are striking, but the prospectus also spends considerable space on the risks of AI.
Roughly 80 of its 261 main pages are devoted to risk factors, nearly twice the 48 pages used to describe the business.
Anthropic says AI could transform the global economy more profoundly than industrialisation, electricity and the internet. At the same time, it warns that advanced AI could pose “catastrophic or existential risks to humanity.” Its models could develop unexpected capabilities, conceal or manipulate information, resist shutdown, or exhibit behaviour resembling blackmail.
The warnings come as Anthropic continues to expand its AI business.
It introduced Claude Opus 5.5 last week, and its IPO is reportedly likely to take place in November.
