Nvidia is investing up to $100 billion in OpenAI as part of an AI data center deal (www.businessinsider.com)

Nvidia announced a strategic partnership with OpenAI in which the chipmaker will invest up to $100 billion to support the build-out and deployment of at least 10 gigawatts of AI data centers running Nvidia systems. The deal cements a long-running collaboration that Nvidia says stretches from its first DGX supercomputers through to the era of ChatGPT, and explicitly ties enormous new hardware deployments to OpenAI’s future model training and inference needs. Technically, 10 GW of dedicated AI data-center power implies a massive expansion of GPU-based clusters—thousands of racks of Nvidia accelerators and associated networking, cooling, and software stacks—aimed at training and serving next-generation large models and multimodal systems at scale. For the AI/ML community this accelerates the compute arms race: it guarantees OpenAI preferential access to vast Nvidia capacity, likely speeds model iteration, and strengthens tight hardware-software co-design. It also reshapes the market dynamic around cloud providers, GPU supply chains, and competitive access to exascale-level training resources, with implications for cost, latency, and who can realistically build frontier models.