Accel is in talks to lead $1B for Thinking Machines at a $40B valuation
It would quadruple the price Mira Murati's lab carried in July 2025 — and come in well below the $50bn-plus it sought late last year. Revenue run rate is above $100 million.
Accel is in talks to lead a round of at least $1 billion for Thinking Machines Lab at a valuation of at least $40 billion, according to reporting on 3 September. Nvidia has discussed participating, and has been reported as considering as much as $2.5 billion.
Thinking Machines was founded by Mira Murati after she left OpenAI as chief technology officer. It raised $2 billion in July 2025 at a $10 billion pre-investment valuation, which makes the mooted price a fourfold increase in a little over a year.
It is also a markdown against the company's own ambition. Thinking Machines sought a valuation above $50 billion late last year and did not get it. Forty billion is a strong outcome that is nonetheless below the ask, which tells you something about where the market for pre-product frontier labs has settled.
The revenue is small and the model is unusual
Annual revenue run rate is above $100 million. At $40 billion, that is a multiple of roughly 400 times revenue — a number that is only comprehensible as an option on the team rather than a valuation of the business.
What the business actually sells is tooling for enterprises to customise models on their own data, while the company's own open-weight models are given away. That is a deliberate inversion of the frontier lab model: the weights are free and the adaptation layer is the product.
There is a real thesis in it. If open-weight quality keeps converging on closed frontier quality — and this month's releases from Z.ai, Alibaba and MBZUAI suggest it is — then the scarce thing is not the model but the ability to make a model work on a particular company's data. Selling that layer while giving away the weights is a bet that the commodity moves down the stack.
The counter-argument is that this is precisely the layer the hyperscalers, the deployment consultancies and the model vendors are all moving into simultaneously. Google and Accenture announced a joint deployment unit this week; Microsoft, OpenAI, Amazon and Anthropic all now run forward-deployed engineering practices.
Nvidia on the cap table again
Nvidia's discussed participation continues a pattern that is now the defining feature of AI financing. Within the same fortnight it agreed to buy Hugging Face for $12.9 billion, committed about $2 billion to Nscale's pre-IPO round, backed Lambda as it signed a $35 billion contract with Anthropic, and took the lease on a Texas data centre from Hut 8.
Each investment creates a customer for Nvidia hardware, and the revenue that follows is booked as demand. It is legitimate vendor financing and it is also increasingly circular, which makes Jensen Huang's forecast of roughly 70 percent revenue growth for fiscal 2028 harder to assess from outside.
What is not confirmed
Nothing is signed. This is reported talks, and Thinking Machines has not commented on the terms, the participants or the raise. Nvidia has not confirmed its involvement or the $2.5 billion figure.
The company has also not disclosed gross margin, customer count, or what share of the $100 million-plus run rate comes from a small number of enterprise contracts. At 400 times revenue, those are the numbers that would distinguish an early price from an indefensible one, and none of them are public.
Runs the newsroom. Rename this profile in the studio to your own byline.
Related
Every weekday, the AI stories that moved money or shipped code.
No cross-posting, unsubscribe anytime. See all newsletters