OpenAI trims its round to $1.25 trillion and Anthropic goes public below expectations: a first sign of a correction?

By Bloomberg · November 11, 2026.

OpenAI has had to scale back its latest funding round. According to Bloomberg, which cites people familiar with the deal who asked not to be named, the company set out in late September to raise at least $30 billion at a $1.4 trillion valuation. Some investors found the terms too ambitious. OpenAI has lowered its target valuation to $1.25 trillion and cut the amount it is seeking by a third. The company declined to comment.

For scale, its last known valuation was $852 billion. Even after the cut, the target is still almost 50% above that figure.

The second development involves Anthropic. Bloomberg reports that it has decided to go public through an IPO (initial public offering) at a $1.4 trillion valuation on a fully diluted basis, meaning it counts every share and every possible option. That figure is lower than expected. According to Bloomberg, both moves have rattled venture capital, which had already watched AI stocks fall for several weeks and now fears a correction in private valuations.

Those are the facts. What follows is our take.

Our view: this is not the AI bubble bursting. It is the point where price stops being a promise and becomes a set of numbers. As long as these companies raised money only in private rounds, their valuations rested on narratives and on limited access. With an IPO, and with investors starting to say “too expensive”, the market wants to see revenue, margins and efficiency. That is the backdrop to the caution.

Who wins and who loses. Second-tier startups lose: if the two leaders have to lower expectations, the most vulnerable valuations belong to companies that rely on the same narrative without the same revenue. Patient investors and companies with solid unit economics win. Over the medium term, discipline wins too: a bubble that deflates gradually does less damage than one that bursts.

It's worth not reading too much into this. A $1.25 trillion valuation is still extraordinary, and a price adjustment says nothing about what the models can actually do. Market corrections and technological progress generally run on different timelines: internet infrastructure survived the 2000 crash and went on to grow many times over.

The short term could be uncomfortable, with less funding for startups, layoffs and delayed projects. That would add to a labor transition that is already delicate. But the underlying thesis doesn't change. A more demanding capital market may speed up exactly what matters in the long run: making AI more efficient and cheaper, which is a precondition for its benefits, from medicine to abundance, to reach everyone. The thing to watch is whether the correction stays confined to valuations or reaches investment in compute.