Anthropic Wants a $2 Trillion IPO After Losing $42 Billion: The AI Bet Gets Bigger

Revenue Surged, But AI Computing Costs Are Exploding

Anthropic's revenue grew almost 12-fold in 2025, reaching nearly $4.6 billion. The rapid growth has been driven by demand for its Claude AI models and enterprise adoption.

But building and operating increasingly powerful AI models is extremely expensive.

Anthropic spent $7.33 billion on compute and infrastructure in 2025, roughly three times its spending in 2024. That represented more than half of its $12.65 billion in total operating expenses.

The company expects those costs to rise substantially. Its prospectus lists $518 billion in future cloud, computing and infrastructure obligations, highlighting the enormous capital requirements behind its AI ambitions.

Anthropic also disclosed $20.28 billion in cash, cash equivalents and short-term investments as of December 31.

Anthropic's AI Safety Warning Adds Another Twist

The financial disclosures arrive alongside unusually strong warnings about AI risks.

In its IPO prospectus, Anthropic warned that advanced AI could create “catastrophic or existential risks to humanity.” It specifically discussed the possibility of models developing unexpected behaviours, including resisting shutdown, concealing information or manipulating people.

The company devoted 80 of its 261-page prospectus to risk disclosures, according to Reuters. It also acknowledged that evaluating AI safety is becoming harder as models become better at recognising and responding to monitoring.

That creates an unusual contrast for Anthropic: the company is simultaneously warning about the risks of increasingly powerful AI while seeking enormous amounts of capital to build and deploy it.

A Major Test for the AI Market

Anthropic's potential IPO would give public-market investors a direct way to invest in one of the leading AI model companies.

The company competes with OpenAI, Google, Meta and xAI, while Amazon and Google are also major strategic partners and infrastructure providers.

The IPO could therefore become an important test of whether investors are willing to assign trillion-dollar valuations to AI companies despite massive infrastructure spending and continued losses.

For Anthropic, the central question is no longer simply how quickly Claude can grow. It is whether that growth can eventually justify the extraordinary cost of building frontier AI.