Former OpenAI CTO Mira Murati’s Thinking Machines partners with Nvidia for gigawatt-scale AI infrastructure

Thinking Machines Lab, an innovative startup spearheaded by Mira Murati, the former CTO of OpenAI, has announced a significant partnership with Nvidia. This collaboration aims to establish a robust AI infrastructure, deploying at least one gigawatt of the cutting-edge Nvidia Vera Rubin systems to enhance the training of advanced AI models. Set to commence early next year, this infrastructure rollout will facilitate Thinking Machines in developing scalable and customizable AI systems. The partnership will not only focus on building training and inference systems specifically optimized for Nvidia's technology but also aims to broaden access to state-of-the-art AI models for businesses, research organizations, and the scientific community. In addition to the partnership, Nvidia has made a notable investment in Thinking Machines Lab, although the financial specifics remain undisclosed. Jensen Huang, Nvidia's founder and CEO, remarked on the transformative potential of AI, stating, "AI is the most powerful knowledge discovery instrument in human history. Thinking Machines has brought together a world-class team to advance the frontier of AI. We are thrilled to partner with them to realize their exciting vision for the future of AI." Mira Murati expressed enthusiasm about the collaboration, highlighting that Nvidia's technology underpins the entire AI field. She emphasized that this partnership will significantly enhance their ability to create AI systems that can be tailored to individual needs, ultimately unlocking human potential. Founded in February 2025, after Murati's tenure at OpenAI, Thinking Machines has kept many of its projects confidential, revealing limited information about its long-term goals. The startup's first product, an API named Tinker, was launched last October, enabling researchers and developers to refine AI models to better suit their needs.