Anthropic's IPO Prospectus Reveals $16B Revenue, $80B Loss, and $600B+ in Computing Commitments

Anthropic has filed a draft IPO prospectus that lays bare the staggering scale — and staggering costs — of building a frontier AI company. The Claude-maker generated $16.3 billion in revenue in the first half of 2026, up from just $1.1 billion in the same period a year earlier — a roughly 15x increase in twelve months. At the same time, computing costs nearly quadrupled, hitting $9 billion in the first half of 2026 alone, more than the $7.3 billion the company spent on infrastructure in all of 2025.

Despite the revenue explosion, Anthropic posted a net loss of $80.7 billion for the first half of 2026 — one of the largest half-year losses ever reported by any company. The headline number is heavily distorted by accounting: $72 billion of that loss stems from changes in the fair value of convertible notes and other financing instruments, not operational cash burn. Still, the underlying operating losses are substantial, and the company lost $42 billion in all of 2025. The IPO is expected to push Anthropic's valuation toward $2 trillion.

The company's computing commitments are eye-watering. Anthropic is locked into roughly $450 billion in noncancelable cloud contracts over the next seven to ten years, including $111 billion with Google, $110 billion with Amazon Web Services, and $31 billion with Microsoft — all regardless of actual usage. On top of that, a separate deal with Broadcom to lease Google-designed chips adds more than $160 billion in obligations, and the company is negotiating additional deals that could push total future infrastructure commitments past $750 billion. Anthropic says it has already bumped up against the limits of what its major cloud partners are willing to provide, and is pursuing its own data centers and third-party operators — some of which have limited track records.

Revenue concentration is a notable vulnerability. About one-fifth of that $16.3 billion in first-half revenue came from just two customers, and the company explicitly warns that a small number of clients will likely drive a disproportionate share of its business for the foreseeable future. The prospectus also dedicates roughly a third of its pages to risk factors — including the possibility that its own AI models could exhibit self-preserving behavior, resist being shut down, or generate outputs resembling blackmail. Anthropic acknowledges it bears direct liability for certain harmful model behaviors and warns that a catastrophic AI incident could produce damages too large to insure against.

On the human side, CEO Dario Amodei and President Daniela Amodei — the siblings who co-founded the company — each drew base salaries of about $655,000 in 2025, with total compensation of roughly $18 million and $16.5 million respectively, placing them near the median for S&P 500 CEOs but below typical big-tech peers. Both have pledged 80% of their personal Anthropic equity to philanthropy, and the company separately announced a plan to commit 5% of its shares to a new public benefit fund.