Anthropic’s confidential IPO prospectus shows the AI company’s revenue grew roughly twelvefold to nearly $4.6 billion in 2025 while it lost more than $8 billion from operations. The filing also discloses a commitment to spend at least $518 billion on computing infrastructure over the next decade, with about 80% of that sum non-cancelable.
The prospectus, reviewed by Reuters and confirmed by multiple people who have seen the document, offers investors the first detailed look at the finances of a company seeking a valuation of more than $2 trillion in a public listing expected after the U.S. midterm elections in November.
Anthropic reported a net loss of about $42 billion in 2025. Roughly $34 billion of that was an accounting charge tied to a higher estimated value for financing arrangements that could convert into company stock. Excluding those writedowns, the operating loss topped $8 billion, up from $2.98 billion a year earlier.
Total operating expenses reached $12.65 billion. Compute and infrastructure accounted for $7.33 billion of that, roughly triple the 2024 level. The company ended 2025 with $20.28 billion in cash, cash equivalents, and short-term investments.
The filing flags concentration risk. Two customers generated nearly a quarter of Anthropic’s 2025 revenue. Many of its largest clients have no long-term contracts and could reduce or stop spending.
Anthropic said it expects to spend at least $518 billion over a decade building AI infrastructure with six partners, according to the prospectus. About 80% of that sum is non-cancelable or requires payment regardless of usage.
The company plans to spend at least $111.1 billion with Alphabet’s Google, $110 billion with Amazon, and $31.4 billion with Microsoft under long-term infrastructure service obligations over the next seven to 10 years.
‘If our actual spend falls short, we must pay Google the difference,’ the company said in the filing, adding that similar terms apply to its Amazon agreement.
The Microsoft commitment is non-cancelable except in the event of an uncured material breach by Microsoft. Anthropic also carries about $161.2 billion in Broadcom-related equipment lease obligations that are largely non-cancelable.
Agreements with Elon Musk’s xAI could lead to as much as $84.5 billion in spending on Nvidia-based computing capacity through 2029. Most of that can be canceled with 90 days’ notice. Advanced Micro Devices has agreed to buy up to $5 billion of Anthropic stock and supply computing capacity expected to exceed $20 billion.
Anthropic said the commitments are necessary because access to computing power is becoming the key constraint on AI development. The company said future demand for advanced AI systems is likely to exceed available supply and will be ‘limited principally by the availability of compute.’
Anthropic devoted more than a third of its IPO filing, around 80 of 261 pages, to laying out potential risks of the technology it is developing. The company warned that AI models pose a ‘catastrophic or existential risk to humanity.’
The filing said advanced AI models could exhibit ‘self-preserving behaviors,’ including being able to ‘resist shutdown,’ ‘conceal or manipulate information,’ and carry out behaviors ‘resembling blackmail.’
Anthropic was founded in 2021 by former OpenAI employees, including chief executive Dario Amodei, who left OpenAI after disagreements over the direction and governance of the ChatGPT maker. The company has long positioned itself as a more safety-conscious alternative.
Amodei has previously written essays warning of AI threats. In a recent essay, he urged the AI industry to slow the pace of model development with a three-step plan to reduce how quickly models improve without ‘sacrificing commercial advantage or the United States’ lead in AI.’
Anthropic confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission on June 1, 2026. The filing came days after the company closed a $65 billion funding round at a post-money valuation of $965 billion.
Anthropic CEO Dario Amodei dined one-on-one with President Donald Trump at the White House on Sunday evening, according to a tech industry source who confirmed the meeting to AFP. The private dinner was the first one-on-one meeting between the two men.
The meeting came after months of open conflict between Anthropic and the Trump administration. A federal appeals court upheld the Pentagon’s decision to classify Anthropic as a ‘supply chain risk’ two days before the dinner, prohibiting the U.S. military from using its models.
The startup had been sanctioned after refusing in February to allow its tools to be used for fully autonomous weapons or domestic mass surveillance. Trump had also criticized Amodei, saying he was trying to be ‘a perfect little angel’ after the Anthropic chief called for a pause in AI development.
Speaking to reporters aboard Marine One, Trump said: ‘Tonight, I’m having dinner with the head of Anthropic,’ referring to Amodei as ‘very highly respected.’
The dinner came ahead of a planned meeting Tuesday between Trump, U.S. House of Representatives Speaker Mike Johnson, and leading AI executives. A White House spokesperson said in a statement: ‘President Trump has been clear: America will lead the world in Super Intelligence while protecting American consumers. We will drive innovation, strengthen American competitiveness, and ensure this transformative technology works for the American people.’
Anthropic’s revenue continued to grow sharply in 2026. Revenue in the second quarter reached $11.5 billion, up from $4.73 billion in the first quarter. The company was approaching its second consecutive quarter of adjusted operating profit.
The listing could value Anthropic at more than $2 trillion, more than double its estimated $965 billion valuation in May. The IPO is expected to occur after the U.S. midterm elections in November.