Accel in Talks to Lead $1B Round for Thinking Machines at $40B Valuation

Thinking Machines is back at the fundraising table, and the numbers deserve scrutiny.

The AI lab founded by former OpenAI CTO Mira Murati is in discussions to raise $1 billion at a valuation of at least $40 billion, The Information reported Thursday. Existing backer Accel is reportedly in talks to lead the round. Neither Accel nor Thinking Machines responded to requests for comment.

The headline figure looks aggressive on its own. Compared against the company's financials, it looks extraordinary.

Thinking Machines' annual revenue run rate stands at over $100 million, according to a source with knowledge of the company's financials. Do the math: a $40 billion valuation against $100 million in revenue implies a multiple north of 400 times. That is not a typo. It is the kind of multiple usually reserved for pre-revenue moonshots, not companies with an actual commercial product generating actual cash flow.

Context matters here. This new round, if completed, would actually represent a step down from where Thinking Machines was aiming. The company reportedly sought to secure a $50 billion valuation late last year. A $40 billion outcome, even at the low end of "at least," suggests the market has recalibrated expectations somewhat, even if only modestly.

The Revenue Reality Check

Thinking Machines does have a product generating that $100 million run rate. In July, the company introduced Inkling, an open-weight model. The revenue engine behind it is Tinker, a platform that charges usage-based compute fees for adapting models on proprietary data. This is a real business model, not vaporware. Enterprises pay to customize models against their own data sets, and Thinking Machines takes a cut of the compute.

The question investors need to answer is whether that revenue trajectory can plausibly justify a valuation 400 times larger than current run rate. Multiples like this are common in early-stage AI, where investors are betting on trajectory rather than trailing metrics. But 400x is an outlier even by the frothy standards of 2026 AI fundraising. It implies either extraordinary confidence in near-term revenue acceleration, or a valuation increasingly detached from unit economics and driven instead by scarcity of access to top-tier AI talent and infrastructure.

There is also the talent question, and it cuts against the optimistic case. Thinking Machines has seen several high-profile departures since its founding. Co-founders including Lilian Weng and Luke Metz have returned to OpenAI. Barret Zoph, another co-founder, left for Google after a stint away from Thinking Machines. For a company whose original $2 billion seed round, one of the largest in history, was priced almost entirely on the pedigree of Murati and the ex-OpenAI researchers she recruited, this attrition matters. The thesis for the $12 billion valuation on that round was talent density. If the talent is drifting back to incumbents, the thesis weakens, even as the price tag on paper climbs higher.

That original seed round is worth revisiting for scale. Andreessen Horowitz led the $2 billion raise, joined by Nvidia, GV, Lightspeed, and Conviction Partners. It valued the company at $12 billion at the time. If this new $40 billion figure holds, Thinking Machines will have more than tripled its valuation in under two years, all while its financial disclosures suggest revenue generation still measured in the low hundreds of millions, not billions.

Investors backing this round are making a bet that looks less like traditional software investing and more like venture capital's version of options pricing. They are paying for optionality on frontier AI capability, not for cash flow multiples that would pass muster in a public markets context. That is not inherently irrational. Compute infrastructure, model IP, and research talent in this sector can compound quickly if a lab hits an inflection point in product-market fit. But it is a bet with a wide dispersion of outcomes, and the departure of founding researchers to rivals is exactly the kind of signal that should widen an investor's risk premium, not narrow it.

Accel's willingness to lead suggests the firm sees enough signal in Tinker's usage-based revenue model, and enough confidence in Murati's remaining team, to justify the premium. Accel is already an existing backer, which gives it informational advantages over a new entrant weighing whether to underwrite a $40 billion price tag from scratch. Existing investors often have visibility into contract pipelines, enterprise commitments, and product roadmaps that outside observers do not.

The Bottom Line

A $40 billion valuation against $100 million in run-rate revenue is not disqualifying on its own in today's AI funding environment, but it demands a much higher bar of scrutiny than the headline number suggests. The valuation cut from a reported $50 billion target is a modest signal of market discipline creeping back into AI fundraising, even at the frontier.

Watch three things going forward: whether Tinker's usage-based revenue model shows acceleration in subsequent quarters, whether further senior researcher departures continue to erode the talent thesis that justified the original valuation, and whether Accel's due diligence, once public, reveals contracted revenue or pipeline visibility that the $100 million run-rate figure alone does not capture. Absent that additional visibility, this round reads as a bet on brand and infrastructure access rather than a valuation grounded in current financial performance. That is not unusual in this sector right now, but it is a meaningfully different risk profile than the multiple alone would suggest to a casual reader.