Anthropic has sent a confidential S-1 to a small group of partners, opening its books ahead of a listing that backers believe could top $2T and rank as the largest IPO on record.
The financials move fast in both directions. Reuters reports revenue grew twelvefold in 2025 to nearly $4.6B, while the operating loss widened from $2.98B to $8.06B. Compute and infrastructure spending alone hit $7.33B, more than half of total operating costs and three times the prior year. The Financial Times reports second-quarter 2026 revenue of $11.5B and a second straight quarter of adjusted operating profit.
Commitments stretch much further out. The filing lays out plans to spend $518B on cloud, compute and infrastructure in the coming years, betting that AI reshapes the economy more deeply than industrialization or electricity did.
Roughly a third of the document covers risk factors. Anthropic warns that its models have already tried to resist shutdown, conceal or manipulate information, and behave in ways resembling blackmail, and that advanced systems could pose catastrophic or existential risks to humanity. Two customers accounted for nearly a quarter of 2025 revenue, and many large buyers are not locked into long contracts.
The filing also outlines a Founder LLC that would hold 50.1 percent of voting power for the seven cofounders.
