Nvidia plans to invest up to $100 billion in OpenAI as part of data center buildout (www.cnbc.com)

Nvidia said it will invest up to $100 billion in OpenAI as the lab embarks on a massive data-center buildout centered on Nvidia GPUs. OpenAI plans to deploy systems that will draw about 10 gigawatts of power — a scale that, according to Nvidia’s CEO, amounts to “monumental” infrastructure spending and underscores how tightly coupled the two companies are: OpenAI depends on Nvidia’s accelerators to train and serve its models, while demand from OpenAI helped ignite the GPU boom that followed ChatGPT. Technically, the announcement highlights the raw compute economics driving next‑generation AI: Nvidia has said building one gigawatt of AI data‑center capacity costs roughly $50–60 billion, with about $35 billion attributable to Nvidia chips and systems, implying multi‑hundred‑billion‑dollar hardware needs at 10 GW scale. The deal also places Nvidia alongside Microsoft, SoftBank and others in OpenAI’s investor roster and complements existing infrastructure partnerships (Azure, Oracle, Stargate). For the AI/ML community this signals continued concentration of training/inference capacity on a few GPU vendors and cloud partners, intensifying competition for chips, shaping model scale and deployment strategies, and making compute infrastructure a defining constraint — and commercial opportunity — for future AI progress.