Former OpenAI Executive’s AI Startup Secures Major Partnership with Nvidia
A prominent artificial intelligence research company founded by former OpenAI chief technology officer Mira Murati has established a significant collaboration agreement with chip manufacturing leader Nvidia.
The partnership between Murati’s Thinking Machines Lab and Nvidia was announced this week, featuring a comprehensive multi-year arrangement that will see the AI research firm utilizing at least one gigawatt of Nvidia’s latest Vera Rubin computing systems beginning in 2027. Financial details of the agreement remain confidential.
Beyond the computing infrastructure deal, Nvidia has also committed to making a strategic financial investment in Thinking Machines Lab. Since launching in February 2025, the startup has successfully secured over $2 billion in funding from prominent venture capital firms including Andreessen Horowitz and Accel, as well as investments from Nvidia itself and even rival semiconductor company AMD’s investment division.
Despite being in its early stages, Thinking Machines Lab has achieved a remarkable valuation exceeding $12 billion. The company focuses on developing artificial intelligence systems capable of producing consistent, reproducible outcomes. In October, the firm launched its inaugural commercial offering, an application programming interface named Tinker.
The collaboration extends beyond hardware provision to include joint development of training and deployment systems specifically designed for Nvidia’s chip architecture. According to Murati, Nvidia’s technology serves as the fundamental infrastructure supporting the entire artificial intelligence sector, and this partnership will enhance their ability to create AI systems that users can customize and control.
However, Thinking Machines Lab has experienced notable personnel changes since its inception. Co-founder Andrew Tulloch departed for a position at Meta in October, while three other founding members – Barret Zoph, Luke Metz, and Sam Schoenholz – returned to OpenAI earlier this year.
This agreement reflects the intense demand for computing resources within the AI industry. Nvidia’s leadership has projected that companies may invest between $3 trillion and $4 trillion in AI infrastructure by 2030. The scale of such deals has grown dramatically, with OpenAI reportedly securing a $300 billion computing agreement with Oracle in 2025.
The partnership underscores the critical importance of access to advanced computing power for AI companies seeking to develop and deploy sophisticated machine learning models at scale.