Existing investor Accel is considering leading the funding round, while Nvidia has discussed participating, according to the report, which cited unnamed sources.
The proposed valuation would represent a fourfold increase from the $10 billion pre-investment valuation Thinking Machines reached in July 2025 when it raised $2 billion, the report said. However, it is below the over $50 billion valuation the startup sought late last year.
Thinking Machines is generating at least a few hundred million dollars in annualized revenue, according to the report. The company makes money by selling businesses tools to customize AI models using their own data, while its open-weight models are available for free.
The new capital would be used to train models, rent computing infrastructure and hire employees, the report said.
The fundraising talks come as investors continue to place multibillion-dollar bets on independent AI labs despite the steep cost of developing frontier models.
When Thinking Machines raised $2 billion in July 2025, Andreessen Horowitz led the round, and Nvidia was among the investors. The deal was one of the largest seed rounds in Silicon Valley.
By November, the company was discussing another round at a valuation as high as $60 billion, although those talks had not been finalized.
Nvidia’s potential participation in the newest round would further deepen its relationship with Thinking Machines. In March, Nvidia announced an investment and multiyear partnership with the startup that included plans for Thinking Machines to deploy at least 1 gigawatt of Nvidia chips for AI training and inference.
Thinking Machines was founded in early 2025 after Murati left OpenAI. The company has positioned itself around customizable AI systems and tools that allow organizations to adapt models to their own data and needs.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
