NVIDIA, Thinking Machines form alliance for gigawatt-scale AI
NVIDIA and Thinking Machines Lab have launched a strategic partnership to deploy Vera Rubin infrastructure, aiming to advance frontier AI models and enhance human-AI collaboration for researchers.
Tech giant NVIDIA and Thinking Machines Lab have announced a long-term strategic partnership to deploy a massive computing infrastructure focused on the next generation of artificial intelligence (AI).
The collaboration centres on the delivery of at least one gigawatt of NVIDIA Vera Rubin systems, which represent the latest rack-scale hardware designed for high-performance data centres.
This gigawatt-scale project, which is roughly equivalent to the power consumption of a major data centre campus, is intended to support the development of frontier AI models, which are highly sophisticated systems that push the boundaries of current technological capabilities in areas such as reasoning and multimodal interaction.
Thinking Machines Lab was established in 2025 by Mira Murati who previously served as the chief technology officer at OpenAI. Based in San Francisco, the startup consists of a team drawn from leading organisations like Meta and OpenAI with a mission to build advanced tools that make AI more accessible.
The company has already introduced an initial product called Tinker which is an application programming interface, or API, that allows researchers to fine-tune and customise existing AI models. To support this vision, NVIDIA has provided significant financial investment alongside the hardware agreement.
“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 Thinking Machines to realize their exciting vision for the future of AI,” said Jensen Huang, Founder and CEO of NVIDIA.
This endorsement highlights NVIDIA’s broader strategy of building what Huang describes as AI factories. These factories are large-scale installations where massive clusters of graphics processing units (GPUs) work together to process the enormous amounts of data required for modern machine learning.
The deployment of the Vera Rubin platform is currently targeted for early next year. This timeline aligns with NVIDIA’s objective to provide hyperscale AI deployments to cloud providers and strategic partners throughout 2026 and 2027.
The partnership also includes a joint effort to design training and serving systems specifically for NVIDIA architectures. This ensures that the frontier AI and open models developed by Thinking Machines Lab are optimised for the most powerful hardware available.
Mira Murati believes that this collaboration will significantly expand what is possible for individuals and researchers.
“NVIDIA’s technology is the foundation on which the entire field is built. This partnership accelerates our capacity to build AI that people can shape and make their own, as it shapes human potential in turn,” she noted.
Her perspective reflects a growing trend in the industry where researchers aim to create collaborative AI that is understandable and customisable rather than just powerful.
This deal sits within a broader context of an intense global race for AI computing power.
NVIDIA has recently formed partnerships with other major players such as Meta and Lenovo to expand the infrastructure needed for the largest AI labs. Earlier this month, NVIDIA committed $4 billion to advance optical technology for AI infrastructure through two multiyear strategic partnerships.
As the industry moves into the era of megaclusters, the ability to access multi-gigawatt capacity is becoming a defining factor for companies trying to build the next generation of AI. By securing this level of compute, Thinking Machines Lab aims to position itself as a major competitor in the field of foundational model development.


