Thinking Machines Lab, led by Mira Murati, is in talks to raise between $5 billion and $6 billion on a pre-money valuation of roughly $40 billion. Accel is poised to lead the funding round, with Nvidia expected to contribute about half the capital, underscoring a strategic partnership spanning investment and hardware supply.
- Raise size updated from $1 billion to $5-6 billion, with Nvidia and Accel as major backers
- Funding supports deployment of at least 1 gigawatt of Nvidia AI infrastructure by 2027
- Revenue from GPU cluster services underpins commercial model despite free open-weight models
Market signal
Thinking Machines Lab’s forthcoming $5 to $6 billion raise at a $40 billion pre-money valuation signals escalating capital requirements in advanced AI infrastructure startups. Early reports of a $1 billion raise underestimated the scale needed to support deployments of cutting-edge Nvidia hardware. This recalibration aligns the capital raise with ambitious projects delivering at least one gigawatt of AI compute power, positioning the company to compete directly with other infrastructure-heavy AI players.
The involvement of Accel since the seed stage and Nvidia’s commitment to invest approximately half of the new round highlights a convergence of chipmaker investment and AI model development. Nvidia’s position as both investor and GPU infrastructure supplier reflects a growing pattern in the AI ecosystem, where hardware providers are financially backing companies that drive demand for their chips, embedding themselves into long-term tech stacks.
Operator impact
Operators evaluating AI infrastructure or platform services should note that Thinking Machines’ model monetizes access to GPU clusters through its Tinker API service rather than charging for the AI models themselves. This approach demonstrates a shift towards service-based revenue models that emphasize compute consumption and user control over data and algorithms, appealing to enterprises wary of ceding model ownership or control.
For operators involved in cloud or edge AI deployments, Thinking Machines’ strategy signals growing opportunities for partnerships with hardware vendors like Nvidia, offering combined access to innovative AI models and specialized compute infrastructure. The startup’s reported annualized revenue in the hundreds of millions, while self-reported, sets a benchmark for similar ventures seeking large-scale adoption and commercial traction while emphasizing open-weight models.
What to watch next
The pace and size of Thinking Machines’ capital raise will be key to watch, given earlier ambitions exceeded $50 billion valuations and the current round is reportedly at least 20% below that. How quickly the company can leverage this capital to deploy gigawatt-scale Nvidia Vera Rubin systems will shape competitive dynamics in AI hardware and platform services.
Additionally, Nvidia’s continued role as a chip supplier and substantial investor in AI startups signals an evolving business model for semiconductor companies. Observers should monitor how this ecosystem orchestration influences pricing, access, and innovation in AI infrastructure markets, especially as more AI labs consider similar structures balancing open model distribution with monetized hardware usage.