Anthropic has filed a confidential S-1 document for its upcoming IPO, which is expected to take place in November. The company aims for a valuation of $2 trillion or more and seeks to raise up to $100 billion. Nvidia, which previously invested $10 billion, may participate as an anchor investor.
The financial data shows significant growth, with revenue rising from $400 million in 2024 to $4.6 billion last year. In the second quarter of this year alone, sales reached $11.5 billion. Despite this, the company reported an operating loss of $8 billion in 2025, with computing infrastructure costs accounting for over 91% of that figure.
Anthropic projects infrastructure spending to reach $518 billion over the next decade. Approximately 80% of these commitments are non-cancellable. Major partners include Amazon, Microsoft, and Google, while the company also holds $161.2 billion in equipment lease contracts with Broadcom.
The prospectus highlights potential risks associated with advanced AI, including existential threats and self-preserving behaviors. The company warns investors that ongoing development could increase the likelihood of models causing harm, noting that the risk factors section of the filing spans 80 pages.
