It may be losing tens of billions of dollars, but Anthropic still believes it can land the highest valuation in history.
Driving the news: The AI company’s IPO prospectus revealed that, despite rapid growth over the past year, it reported a net loss of US$42 billion in 2025, per Reuters. Nonetheless, the company is reportedly targeting a $2 trillion valuation for its public market debut, which is expected in November.
According to the prospectus, Anthropic’s revenue jumped by 12-fold last year to $4.6 billion, though it also spent $7.3 billion on compute and infrastructure — more than half of its total operating costs.
Zoom in: CEO Dario Amodei spent about a third of the prospectus warning investors that the technology they would be funding “may pose existential risks to humanity.” Anthropic noted that it will continue to operate as a Public Benefit Corporation, which, in theory, allows leaders to balance investor interest with safe AI development.
The company is also creating a new “Founder LLC” group with the sole responsibility of insulating Anthropic’s leadership from the pressure of public investors and ensuring AI benefits humanity.
Why it matters: The prospectus shines a light on how safety fears could put a damper on the long-awaited IPOs of the AI industry’s two biggest players. When you couple the doomsday alarm bells with a $2 trillion valuation and soaring operating costs, betting on Anthropic or OpenAI has become a far riskier proposition.
OpenAI did little to dispel those fears, announcing that it was shelving the release of its newest model over safety concerns this week.
Our take: So far, the AI race has largely been centred around building the most powerful, cost-efficient models as quickly as possible. Now, top labs are going to be fighting a PR battle to convince all of us that they’re the ones building the “safe” AI.—LA
