Anthropic Lost $80.7 Billion in Six Months, Draft IPO Filing Shows
The Claude maker booked $16.3 billion in revenue in the first half of 2026 — fifteen times a year earlier — while the draft prospectus reveals nearly $9 billion in computing costs and future infrastructure commitments that could reach $750 billion.
Anthropic lost $80.7 billion in the first half of 2026, according to a draft IPO prospectus reviewed by The Wall Street Journal and Reuters.
The company generated $16.3 billion in revenue in the six months ended June 30. That is about fifteen times the $1.1 billion it made in the same period a year earlier. The loss compares with $13 billion a year earlier.
About $72 billion of the loss was a non-cash accounting charge. When a company’s valuation rises, the estimated value of convertible notes and other financing instruments that can turn into shares also rises, and the company books the increase as a loss. No cash left the business over that charge.
The same pattern shaped last year’s numbers. Anthropic lost about $42 billion in 2025, of which roughly $34 billion was a similar financing charge. Stripped of it, the operating loss exceeded $8 billion on revenue of about $4.6 billion.
The draft filing also lays out the other side of the AI race: computing costs. Anthropic spent nearly $9 billion on computing capacity and infrastructure in the first half, up from $2.3 billion a year earlier.
The future commitments run much larger. The Journal reported about $450 billion in mostly non-cancelable cloud agreements over the next seven to 10 years — $111 billion with Google Cloud, $110 billion with Amazon Web Services, and $31 billion with Microsoft, payable regardless of usage.
Other disclosed deals include up to $86 billion for Nvidia chips rented through AI cloud providers, and more than $160 billion for Google chips through an arrangement with Broadcom. With additional agreements still under negotiation, potential computing spending could reach $750 billion, the Journal reported. That broader figure includes prospective deals and is not all fully contracted.
The filing flags two other risks. Two customers accounted for roughly one-fifth of first-half revenue, and the company expects a small number of customers to keep generating a large share of sales. It also notes potential AI-related liabilities ahead of the listing.
Anthropic confidentially filed its draft S-1 with the Securities and Exchange Commission in June. A public listing could come as soon as November. The company declined to comment on the filing’s figures, and the numbers in the draft may still change.
Earlier reports on the confidential filing, first surfaced in September, indicated the company is targeting a valuation above $2 trillion.
