Anthropic released its IPO filing, warning investors about 'existential' risks from its own AI technology. The S-1 prospectus highlights the company's rapid growth, with revenue rising twelvefold to nearly $4.6 billion in 2025, according to Reuters.
The filing also shows mounting costs, with operating losses expanding to $8.06 billion. Anthropic spent $7.33 billion on compute and infrastructure alone, more than triple the previous year's figure, and nearly half of total operating costs.
The prospectus warns that AI could pose risks such as manipulation, blackmail, and unpredictable behavior. Nearly a third of the lengthy document covers risk factors, including the potential for advanced models to act in ways that could threaten humanity, according to the Financial Times.
"Existential risks to humanity" are among the key concerns outlined in the filing, said the Financial Times, which reviewed the prospectus. The company has opened its books to potential investors, signaling its readiness for a public market debut.
The filing also reveals that Anthropic depends heavily on a small number of buyers. Two customers accounted for nearly a quarter of revenue in 2025, and many large clients aren't locked into long-term contracts, the company warns.
Anthropic plans to spend $518 billion on cloud, compute, and infrastructure over the coming years. The company expects AI to reshape the global economy more profoundly than past technological revolutions, according to the prospectus.
Source: thedecoder
