The AI funding race has a new headline number, and it belongs to the rebels. Thinking Machines, the frontier lab assembled from former OpenAI leadership, is reportedly in discussions with Accel to lead a $1 billion round that would value the company at $40 billion — a figure that would make it one of the most valuable private companies on earth before shipping a single flagship model.
The lab was founded on a thesis: that the world's most powerful AI models should be open, and that the closed, vertically integrated labs have overextended both their architecture and their control. Its roster of researchers, recruited at extraordinary compensation, signaled immediately that this was not a hobby project. But open-weight frontier training is expensive, and the compute commitments implied by that ambition are measured in the billions.
Accel's reported willingness to lead at a $40 billion valuation reflects how far sentiment has moved. Enterprise buyers, developers and governments are hungry for a credible alternative to the dominant labs, and investors are paying for optionality on that demand. A round of this size also mirrors the strategic logic of the incumbents: whoever can bankroll the next training run gets a seat at the table where the industry's rules are written.
Doubters note the obvious risk. A $40 billion price tag demands a business that does not yet exist publicly, and the graveyard of AI labs is deep. Yet the market keeps proving it will fund conviction over proof. If the talks close, Thinking Machines will have done something rare: priced its philosophy at a premium, and found a buyer at that number.
