WSJ: Anthropic Lost $80 Billion in First Half of Year, Has $600 Billion in Infrastructure Commitments

AI startup Anthropic lost more than $80 billion in the first half of 2026, according to a draft IPO prospectus viewed by the Wall Street Journal, even as the AI company’s revenue climbed and it prepared for a possible IPO that could value it at $2 trillion.

The Wall Street Journal reports that the net loss, $80.7 billion for the six months ended June 30, 2026, dwarfs the $13 billion Anthropic lost in the same period a year earlier and the $42 billion it lost in all of 2025. Most of the jump, $72 billion, came from accounting charges tied to changes in the fair value of convertible notes and other financing instruments, according to the prospectus. Anthropic declined to comment on the filing.

Anthropic’s revenue reached about $16.3 billion in the first half of 2026, up from $1.1 billion a year earlier. Full-year 2025 revenue, for comparison, was $4.6 billion. That growth has come at a steep cost: Anthropic spent nearly $9 billion on computing capacity in the first half of 2026 alone, more than the $7.3 billion it spent in all of 2025, and far above the $2.3 billion it spent in the first half of last year.

About one-fifth of the company’s first-half revenue came from just two customers, and Anthropic warned in the prospectus that it expects to keep relying on a “limited number of customers for the foreseeable future,” with a “significant portion” of revenue tied to them.

Looking ahead, the company’s infrastructure appetite is enormous. Anthropic signaled it could spend up to $750 billion in coming years on computing capacity, including deals still under negotiation, and said it is already committed to roughly $450 billion in cloud agreements over the next seven to ten years, mostly noncancelable. That includes $111 billion earmarked for Google, $110 billion for Amazon Web Services and $31 billion for Microsoft, spending the company says will happen “regardless of actual usage.” Separately, Anthropic has agreed to spend up to $86 billion renting Nvidia chips from Nscale and Lambda through 2033 and 2034, and carries an obligation of more than $160 billion to lease Google chips through a deal with Broadcom. Altogether, future infrastructure costs top $600 billion, and the company is negotiating additional deals that could push commitments up by as much as $150 billion more.

Anthropic also disclosed friction with the same cloud giants it depends on. The company said AWS, Google and Microsoft increasingly view it as a competitor and have limited the computing capacity “those partners were willing to provide.” In response, Anthropic is building its own data centers through long-term leases and third-party operators, some of which have only a “limited operating history.” Most of its current data center capacity sits in the United States, though the company is pursuing expansion abroad and has signed some letters of intent. It also warned that political opposition to data center construction could “become more organized and coordinated across jurisdictions,” and that its own prominence “may draw particular attention to projects associated with us.”

Roughly a third of the draft prospectus covers risk factors, including the potential for Anthropic’s AI models to cause real-world harm. The company wrote that its models can display “self-preserving” qualities and may generate “blackmail”-like responses in an effort to avoid being shut down. Anthropic said it could bear liability for harmful model behavior, such as stolen data or security incidents, and cautioned that insurance may be unobtainable for the “harm that could result from a catastrophic AI-related incident.”

On executive pay, CEO Dario Amodei and President Daniela Amodei, who are siblings, each earned a base salary of about $655,000 in 2025. Including stock and option awards, Dario’s total 2025 compensation was about $18 million and Daniela’s was about $16.5 million, according to data from MyLogIQ cited in the filing. That puts the pair roughly in line with median S&P 500 CEO pay, but behind many big-tech peers. Both are eligible for additional equity tied to future milestones. Each co-founder has pledged to dedicate 80 percent of their personal Anthropic equity to philanthropy, the prospectus states. On the Sunday before the article’s publication, Anthropic announced plans to commit 5 percent of its shares to a new “public benefit fund” that can make charitable donations or direct money to “specific populations.”

Anthropic and CEO Dario Amodei have been at the forefront of the AI doomer movement, with Amodei calling for an AI slowdown, which OpenAI’s Sam Altman and SpaceX CEO Elon Musk both supported:

Amodei’s essay lays out a three-step plan meant to temper the AI industry’s development pace without stopping it outright. “To be clear, pacing does not mean halting model training or technical progress, but ensuring companies take adequate time to align and safeguard their models, and for third party evaluators to confirm this,” he wrote. Anthropic claims it has already taken the first step itself. The company has “unilaterally” committed to giving third-party evaluators employee-level access so they can verify its safety practices and report incidents, a move Amodei said comes without “sacrificing commercial advantage or the United States’ lead in AI.” The second step calls on leading AI companies in democratic countries to coordinate around common safety standards. The third, and most ambitious, asks for coordination between democratic and authoritarian governments.

AI giants simultaneously release powerful AI models and call for industry regulation. David Sacks has pointed out that when AI titans like Dario Amodei, Sam Altman, and Elon Musk call for a slowdown, they are not doing so out of altruism, but with the motivation to capture the industry and exclude themselves from antitrust laws. Breitbart News social media director Wynton Hall has written his New York Times bestseller Code Red: The Left, the Right, China, and the Race to Control AI to serve as the definitive guide on how the MAGA movement can create positions on AI that benefit humanity without handing control of our nation to the leftists of Silicon Valley or allowing the Chinese to take over the world.

Read more at the Wall Street Journal here.

Lucas Nolan is a reporter for Breitbart News covering issues of AI, free speech, and online censorship.