NVIDIA and Thinking Machines Lab, the public benefit corporation founded by Mira Murati in early 2025, have entered a multi-year strategic partnership to deploy a minimum of one gigawatt of Vera Rubin systems, the companies said in a joint announcement Tuesday. The collaboration couples large-scale hardware deployment with a significant capital commitment from NVIDIA intended to support Thinking Machines’ long-term growth and research agenda.
The project is designed to supply the compute necessary for training frontier AI models and to offer highly scalable, customizable AI platforms. Thinking Machines plans to run NVIDIA architectures to power core offerings such as its "Tinker" fine-tuning API and other human-AI interaction tools, expanding access to open models for both enterprise customers and the scientific community.
Thinking Machines, which Murati launched as a public benefit corporation in early 2025 after serving as chief technology officer at OpenAI, has moved rapidly in its first months. The startup previously raised a $2 billion seed round that placed its valuation at $12 billion shortly after its founding.
Jensen Huang, founder and chief executive officer of NVIDIA, described the company’s large hardware commitment as a response to the accelerating need for computing capacity to advance AI. He emphasized the importance of the team that Murati has assembled in pushing the field forward.
Murati said the alliance provides foundational technology and capacity to help build AI systems that are both understandable and collaborative, and that NVIDIA’s platforms form the underlying infrastructure on which much of the field is built. The partnership is explicitly focused on designing specialized training and serving systems at scale while broadening access to open models.
NVIDIA has not yet responded to requests for comment. NVIDIA’s stock was up 1.7% in late morning trade.
Context on business and product focus
- The agreement centers on deploying Vera Rubin systems at gigawatt scale to support training of frontier models and on creating customizable AI platforms.
- Thinking Machines will use NVIDIA architectures to enhance its Tinker fine-tuning API and other human-AI interaction tools.
- The deal includes a substantial NVIDIA investment to support the startup’s long-term growth and research capabilities.
Additional commercial note
The announcement follows Thinking Machines’ large seed raise and rapid valuation, reinforcing the company’s positioning as a well-funded entrant into the AI infrastructure space.
About market signals
Market reaction to the announcement included a rise in NVIDIA’s share price during late morning trading.
Disclosure
No additional disclosure provided.