Anthropic incurred nearly $9 billion in computing capacity and infrastructure costs in the first half of 2026, exceeding the $7.3 billion spent during all of 2025, a new draft IPO prospectus shows.

The San Francisco company, led by Chief Executive Dario Amodei, generated $16.3 billion in revenue for the six months ended June 30, up from $1.1 billion in the year-earlier period. Computing capacity and infrastructure costs rose from $2.3 billion in that same period of 2025.

The new figures come from a draft reviewed by The Wall Street Journal for its October 20 report, rather than a public filing or company release. Anthropic declined to comment.

What Changed

AI-generated summary, reviewed by an editor. More on our AI guidelines.

First-half losses

Anthropic recorded a first-half net loss of $80.7 billion, compared with $13 billion in the same period a year earlier. Of the latest loss, $72 billion came from accounting charges tied to changes in the value of convertible notes and other financing instruments issued to raise funding.

Those charges reflect changes in the estimated value of financing obligations, rather than equivalent cash payments during the period. Subtracting the charges leaves $8.7 billion of net loss, but that remainder does not establish an operating-loss figure or measure cash spent running the business during those six months.

Later-quarter performance

In a reported private investor briefing last week, Anthropic said third-quarter revenue reached $16.8 billion, up about 45% from the previous quarter. Revenue for those three months exceeded the $16.3 billion generated in the entire first half.

Anthropic also reported an adjusted profit for that July-to-September period. That later-quarter measure is separate from the first-half net loss. The available reporting does not establish how the company defines adjusted profitability or reconcile that measure with its results under generally accepted accounting principles.

Infrastructure contracts

Anthropic has around $450 billion in cloud agreements over the next seven to ten years, with the bulk noncancelable. These include $111 billion to Google, $110 billion to Amazon Web Services and $31 billion to Microsoft, payable regardless of actual usage.

The supplier list also includes up to $86 billion to rent Nvidia chips from Nscale and Lambda through 2033 and 2034, respectively, plus other cloud deals. The company's public Amazon agreement secures up to five gigawatts of capacity to train and deploy Claude, with more than $100 billion committed to AWS technologies over ten years.

Separately, more than $160 billion in equipment lease obligations to obtain Google chips through a deal with Broadcom brings future infrastructure commitments above $600 billion. Broadcom Chief Executive Hock Tan said at a conference earlier this month that rising interest rates could make access to capital the biggest constraint on AI companies.

Know someone who'd find this useful? ✉️ Email it to a friend in one click, or they can subscribe free here.

Additional computing deals under negotiation could add as much as $150 billion, taking the potential total toward $750 billion.

Supply and liability risks

Around one-fifth of first-half revenue came from two unnamed customers.

The draft says Anthropic has sought more computing capacity than its cloud partners were willing to provide and describes those partners as increasingly viewing it as a competitor.

Anthropic is developing dedicated data centers through long-term leases and using third-party operators. The draft describes some operators as having limited operating histories and warns their facilities may be less reliable than those run by large cloud providers.

Anthropic also disclosed potential liability for harmful model behavior, including data theft and security incidents. It warned that damages from a catastrophic AI incident could be too large for it to obtain insurance covering the potential losses.

Frequently Asked Questions

How quickly did Anthropic's first-half revenue grow?

Revenue reached $16.3 billion in the six months ended June 30, 2026, compared with $1.1 billion during the same period in 2025.

Does the $80.7 billion net loss measure cash burn?

The first-half net loss includes $72 billion in accounting charges tied to changes in the value of convertible notes and other financing instruments. Subtracting those charges does not establish an operating loss or cash-burn figure.

Has Anthropic signed $750 billion in infrastructure agreements?

The reported infrastructure commitments exceed $600 billion. The potential $750 billion total includes up to $150 billion in additional computing deals still under negotiation.

How do the later profitability claims compare?

Accounts of a private investor briefing described $16.8 billion in third-quarter revenue and an adjusted profit. Those claims concern July through September, a later period than the first-half net loss.

Where do the new first-half figures come from?

They come from a draft prospectus reviewed by The Wall Street Journal for its October 20 report, rather than a public filing. Anthropic declined to comment.

AI-generated summary, reviewed by an editor. More on our AI guidelines.

OpenAI Puts Revenue Run Rate Near $50 Billion, Sets $70 Billion Year-End Target
OpenAI told investors its annualized revenue was about $50 billion at the end of September and expects to reach $70 billion or more by year-end. The nearly $70 billion figure that circulated in late S
Anthropic IPO: A Somewhat Peculiar Prospectus and Concerns About an AI Bubble
Ross Hendricks, a professional research analyst whose Ross Report focuses on financial markets, particularly energy, and who has worked directly with the chief investment officer of a $500 million glo
Broadcom Agrees to Lend Anthropic Up to $42 Billion to Lease Chips It Helps Design
Broadcom has agreed to lend Anthropic up to $42 billion through convertible notes, Anthropic's IPO prospectus obtained by Reuters shows. The notes could convert into Anthropic shares, with proceeds re
AI News

San Francisco

Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: editor@implicator.ai