Mira Murati’s AI lab is back in the fundraising market, and the number attached to it is enormous. Thinking Machines, the startup founded early last year by the former OpenAI CTO, is in talks to raise $1 billion at a valuation of at least $40 billion, according to TechCrunch AI, citing reporting from The Information. Existing backer Accel is in discussions to lead the round.
What stands out here is the speed and the scale. This company has almost no public product history, yet investors are lining up to value it in the same league as established AI players.
📊 The core numbers
Here’s what TechCrunch AI reports:
- New round: roughly $1 billion in fresh capital.
- Target valuation: at least $40 billion.
- Lead investor: Accel, already on the cap table, in talks to lead.
- Revenue run rate: over $100 million annually, per a source with knowledge of the financials.
That last figure matters. A $40 billion valuation against $100 million in run-rate revenue works out to a multiple around 400x. Even in a frothy AI market, that’s extraordinary. Investors aren’t paying for today’s revenue. They’re paying for who’s building it and what they expect it to become.
📉 A markdown from the earlier ambition
The $40 billion figure is actually a step down from what the company reportedly wanted. Late last year, Thinking Machines was said to be seeking a $50 billion valuation. Landing at $40 billion, if the round closes, signals that even the hottest names are feeling some gravity in a market that spent 2025 pricing AI labs at dizzying levels.
Still, look at the trajectory. The prior raise, a $2 billion round that ranks among the largest seed financings ever, valued the company at $12 billion. Andreessen Horowitz led it, joined by Nvidia, GV, Lightspeed, and Conviction Partners. If this new round closes at $40 billion, that’s more than 3x in under a year.
🧠 Why investors keep writing checks
The original bet was pedigree. Murati ran engineering and research at OpenAI. She brought a roster of former OpenAI researchers with her, and that talent concentration is what backers paid for.
The product story is starting to fill in too. In July, the company launched Inkling, an open-weight model. It earns money through usage-based compute fees for adapting models on proprietary data via the company’s Tinker platform. That’s a real revenue mechanism, not just a research demo, and it explains how a lab this young already posts a nine-figure run rate.
⚠️ The talent question
Not everything is pointing up. Thinking Machines has seen several high-profile exits. Some co-founders, including Lilian Weng and Luke Metz, went back to OpenAI, according to TechCrunch AI. When a company’s entire thesis rests on the people, departures like these are the risk investors have to weigh against the valuation. Accel and Thinking Machines didn’t immediately respond to requests for comment.
🔭 What to watch next
This is significant because it’s another data point on how the AI funding market is behaving in late 2025. A few things to track:
- Does the round actually close at $40 billion, or does the final number drift as due diligence plays out?
- Whether Inkling and Tinker adoption grows fast enough to justify the multiple.
- Whether more senior researchers stay or follow the earlier departures out the door.
- How rival labs price their next raises against this benchmark.
For practitioners, the takeaway is practical. Capital is still flowing to open-weight and model-customization plays, and platforms that let teams adapt models on their own data are commanding premium valuations. That’s a signal about where the money thinks the next wave of value sits.
The deal isn’t done yet. Full details are at the original source.