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

Top-tier AI labs have followed one script for two years: valuations only go up. A new negotiation reportedly breaks that pattern.

The Information reported that Thinking Machines Lab is in discussions to raise at least $1 billion at a valuation of roughly $40 billion, with existing backer Accel in talks to lead the round. Multiple outlets, including Reuters, have also reported that Nvidia is discussing joining the round.

What stands out is the direction of the number. Late last year, Thinking Machines reportedly sought a valuation above $50 billion for a $4-5 billion raise. The terms now under discussion are smaller on both counts — a rare case of a top-tier AI lab negotiating down from its own prior target.

Thinking Machines was founded in February 2025 by Mira Murati, OpenAI’s former CTO. In June 2025, it closed a $2 billion seed round — one of the largest seed rounds in Silicon Valley history — led by Andreessen Horowitz with participation from Nvidia, GV, Lightspeed, and Conviction Partners, at a $12 billion valuation. In March 2026, the company struck a strategic investment and gigawatt-scale Vera Rubin supply deal with Nvidia. That earlier round was largely a bet on Murati’s pedigree and the roster of former OpenAI researchers she recruited.

Since then, several of those co-founders have left. Andrew Tulloch departed for Meta in October 2025. Barret Zoph and Luke Metz both returned to OpenAI in January 2026. On July 27, co-founder Lilian Weng also left for OpenAI, writing internally that she couldn’t sustain the pace a startup demands for health reasons — not because she’d lost interest in the research.

In July, Thinking Machines introduced its first revenue-generating product: Inkling, an open-weight model built on Tinker, the model-customization platform it launched in October 2025. Inkling charges usage-based compute fees for adapting models to proprietary data. A source familiar with the company’s financials said the resulting annual revenue run rate has already topped $100 million — though a $40 billion valuation against that figure implies an extraordinarily high revenue multiple.

A Broader Cooling, or a One-Off?

Frontier AI labs have mostly raised at ever-higher valuations. Safe Superintelligence, led by Ilya Sutskever, reportedly took a $5 billion equity investment from Nvidia to scale up compute tenfold. Against that backdrop, Thinking Machines negotiating down from its own prior target is worth watching as a signal — even if, for now, it may be specific to this one company. The talks remain unconfirmed, and neither Accel nor Thinking Machines immediately responded to requests for comment.