Behind Anthropic’s stunning $42 billion loss lies the massive stakes of Google and Amazon
A leaked copy of Anthropic’s S1 prospectus shows the Claude maker reported $8 billion in operating losses last year. Total losses exceeded $42 billion due to the way Google and Amazon’s stakes in the AI company were accounted for.
Getty Images
Leaked Anthropic prospectus revealed that the AI giant made $4.6 billion and lost $8 billion in operating terms last year. But its total losses last year were $42 billion.
That would have been one of the worst annual losses ever for a U.S. company, nearly equivalent to the $29.1 billion loss Citigroup reported during the 2008 financial crisis. But Anthropic didn’t spend that money on IT, talent or buyouts. Rather, they are bookkeeping costs measured according to generally accepted accounting principles (GAAP).
This charge is largely due to increased stakes from two major Anthropic investors. Google first invested in the company in mid-2023, but later that year it will invest $2 billion in debt known as convertible notes. These can then be exchanged for shares in the AI startup. Amazon also invested $8 billion at the end of 2024 for a valuation of $46 billion, similarly via convertible debt.
“This loss is the consequence of good news.”
Now, as the company targets a $2 trillion valuation for its IPO, the value of shares promised to Google and Amazon has also exploded. Anthropic must consider this an accounting loss.
It is unclear whether the $34 billion in losses also affects other investors.
“The loss is the consequence of good news,” wrote Joshua Ronen, an accounting professor at the NYU Stern School of Business, in an email to Forbes. “The burden is enormous precisely because the valuation has soared.”
Google said in its latest earnings report that it has $124.3 billion in private startups. The search giant did not disclose the name of these investments, but noted that they largely came from a single company. The size of its startup investments is so vast that it has created an accounting problem even for a company the size of Google. In July, it recorded $99 billion in “other income” from its investments in SpaceX and “another private startup.” THE New York Times reported in March 2025 on court filings showing Google owned 14% of Anthropic.
Amazon also noted in its July earnings report that its net profit jumped to $62.6 billion from $18.2 billion last year, largely due to its investment in Anthropic. This increase in revenue was largely attributable to a $53.4 billion markup on its stake in the startup. Amazon said it invested $18 billion in Anthropic and has the right to invest an additional $20 billion.
Amazon and Google continued to invest in subsequent rounds for Anthropic, which raised more than $126 billion, according to Pitchbook. Part of Amazon and Google ratings have already been converted into shares.
Anthropic still spent a large sum of money to run its business: developing its Claude models, paying salaries and purchasing computing power. But $8 billion is nothing compared to the $518 billion he plans to spend on AI infrastructure over the next decade. Google’s cloud division and Amazon’s AWS arm will likely be among the biggest beneficiaries of this spending windfall.
Anthropic, Google and Amazon did not immediately respond to requests for comment.
Anthropic’s market debut is highly anticipated, after the San Francisco-based startup confidentially filed its prospectus in June – a landmark event that would also shape the IPO environment for rival OpenAI. It also comes as the general public debates the existential safety of AI and regulations that could serve as safeguards, as AI models from the world’s largest AI labs have been hacked by third parties.
Anthropic CEO Dario Amodei was one of a handful of high-tech executives who signed a deal with President Donald Trump at the White House yesterday to personally control the development of what the president called “super intelligence.” Earlier this month, CEO Dario Amodei called for a slowdown in the development of advanced AI, urging “caution” as labs advance their models and amid a series of hacking incidents involving AI agents.
This is not the only tragedy that has devastated the company. Earlier this year, Amodei engaged in a high-profile confrontation with the War Department over the Pentagon’s ability to use Anthropic’s AI for domestic mass surveillance and autonomous weapons. After tense exchanges, the government labeled Anthropic a “supply chain risk” – a devastating blow for any company that does business with the military. Anthropic then sued the DoW and a judge issued an injunction regarding the designation.
More from Forbes
Gn headline
