Overview
- NVIDIA and Thinking Machines announced a multiyear partnership to provide at least one gigawatt of next‑generation Vera Rubin systems.
- The agreement includes a significant NVIDIA investment in Thinking Machines, with financial terms not disclosed.
- Deployment is targeted for early next year, with the computing power primarily dedicated to training the startup’s AI models.
- Industry executives estimate a one‑gigawatt build can cost about $50 billion, highlighting the scale of the planned capacity.
- The arrangement reflects NVIDIA’s growing role as both financier and chip supplier, with reported large stakes in OpenAI and Anthropic and an earlier investment in Thinking Machines’ $2 billion seed round.