OpenAI and Nvidia Announce Partnership to Deploy 10GW of Nvidia Systems (openai.com)

OpenAI and NVIDIA announced a landmark strategic partnership to deploy at least 10 gigawatts of NVIDIA systems—described as representing “millions of GPUs”—to power OpenAI’s next-generation AI infrastructure. NVIDIA plans to invest up to $100 billion in OpenAI progressively as each gigawatt is brought online, with the first gigawatt slated for the second half of 2026 using NVIDIA’s Vera Rubin platform. The deal names NVIDIA as a preferred strategic compute and networking partner; both companies will co‑optimize model, infrastructure, hardware and software roadmaps as OpenAI scales training and inference for its future models, including work framed as “on the path to deploying superintelligence.” Technically and strategically, this is a major acceleration of compute consolidation: 10 GW of purpose-built GPU systems implies unprecedented training throughput, huge power and datacenter buildouts, and tighter integration across stack layers (chips, systems, networking, and model software). The progressive $100B investment suggests NVIDIA will help finance or enable datacenter and power capacity expansions, reshaping supplier relationships and competitive dynamics with cloud providers and hardware rivals. For the AI/ML community this raises opportunities—faster iteration, larger models, more ready infrastructure—and tradeoffs around concentration of compute, energy footprint, supply‑chain constraints, and geopolitical or regulatory scrutiny as massive centralized compute becomes even more dominant.