OpenAI hits $1.25 trillion valuation as Anthropic nears IPO

OpenAI has been valued at $1.25 trillion, according to The Verge, as rival Anthropic prepares for its own initial public offering.

CEO Sam Altman said recently that OpenAI is prioritizing safety instead of going “guns blazing” toward an IPO. The $1.25 trillion figure places the company among the most valuable private firms in history, exceeding the market capitalization of Meta and approaching that of Amazon.

Anthropic, OpenAI’s primary competitor, is nearing its public debut. An outside research group said Anthropic is “rapidly accelerating” AI development even as calls for an industry slowdown grow louder in Washington following the midterm elections.

The valuations arrive at a moment of political shift. Democrats, who won both chambers of Congress, campaigned on reining in AI development. The Verge reported that fears of AI regulation contributed to a stock selloff this week.

Separately, Microsoft is reportedly “winding down” a cloud AI deal with Oracle, turning off rented server capacity from neoclouds that powered its AI efforts. Bloomberg reported that the Oracle capacity used older Nvidia chips that had become less useful.

What the coverage missed

The coverage focused on the numbers. The part that got less attention is the contradiction at the center of both companies’ positions.

Altman says OpenAI is prioritizing safety. Anthropic says the same. Both companies are raising capital at record valuations that imply enormous revenue growth. A company that is prioritizing safety over growth does not command a $1.25 trillion valuation. Investors are not paying for caution. They are paying for speed.

The outside research group’s finding that Anthropic is “rapidly accelerating” development is the detail that does not fit the safety narrative. A company that slows down does not need to go public. A company that speeds up does.

Microsoft’s decision to wind down its Oracle deal is also a signal. The company is not reducing its AI investment. It is consolidating control over its own infrastructure. The neocloud era, in which AI companies rented capacity from third parties, is ending. The buildout is not stopping. It is moving in-house.

What I think

The valuations are a bet that AI will generate enough revenue to justify prices that exceed the output of most nations. That bet may be right. It may be wrong. What it is not is a bet on safety.

A $1.25 trillion valuation assumes OpenAI will produce products that generate hundreds of billions in revenue. Products that generate that much money are products that are widely deployed. Widely deployed AI is not cautious AI. The valuation and the stated priority are in conflict, and the valuation is the one that reflects reality.

The political timing matters. Democrats won on a message of affordability and accountability. They have promised oversight of AI. If they follow through, the regulatory environment that these valuations assume may not arrive. The market is pricing in a future that the electorate just voted against.

Anthropic’s IPO will be the test. If it prices at a valuation comparable to OpenAI’s, the market is saying that the AI buildout is a two-company race. If it prices lower, the market is saying that OpenAI’s lead is worth more than its competitor’s safety claims. Either way, the safety narrative is marketing. The valuations are the truth.