Anthropic’s IPO Filing Offers a Look Under the Hood of the AI Boom

This story was featured in the Financial Poise Weekly newsletter.

Read the full newsletter for more of the week’s top stories.

As Anthropic prepares for its $2 trillion-plus IPO, its new prospectus is revealing just how expensive its rapid growth has become.

Echo Wang at Reuters, the prospectus showed that:

• Anthropic’s revenue grew twelvefold in 2025 to nearly $4.6 billion, but its operating loss widened from $2.98 billion to $8.06 billion.
• Computing is one of the biggest expenses. Anthropic spent $7.33 billion on compute and infrastructure last year, three times its 2024 spending. It has also committed to another $518 billion in cloud, computing, and infrastructure obligations in the coming years.

A large portion of the prospectus was devoted to AI-related risks.

As the New York Times’ Dealbook newsletter explains:

• Anthropic warns that increasingly autonomous models could exhibit “self-preserving behaviors” or attempt to “conceal or manipulate information,” potentially exposing the company to security risks and legal liabilities.
• More strikingly, Anthropic reportedly warns that its technology could pose “catastrophic or existential risks to humanity.” This comes after CEO Dario Amodei publicly called for a slowdown in the AI industry due to safety concerns, which we covered here.
• The implications go beyond Anthropic. Amazon and Google hold large stakes in the company, so a delayed IPO or lower-than-expected valuation could force those investments to be marked down. This could also put pressure on valuations elsewhere in the interconnected AI industry.

Our take? Anthropic is effectively asking investors to believe both that AI could transform the global economy and that the technology is powerful enough to create risks it describes as potentially catastrophic.

The financial risks are easier to put numbers around: enormous infrastructure commitments, heavy losses, customer concentration, and a valuation that assumes extraordinary future growth. The harder question for investors is how to price risks that even the company developing the technology says are difficult to predict or control.

If you want to read an argument in favor of making Anthropic and its competitors slow down, read If Anyone Builds It, Everyone Dies. It’s really uplifting…