Overview

• The companies confirmed a strategic pact to build and operate at least 10GW of GPU capacity using NVIDIA hardware to train and deploy OpenAI’s next models.
• The first deployment phase is slated for the second half of 2026 on NVIDIA’s Rubin platform, with detailed terms to be finalized in the coming weeks.
• NVIDIA said it will invest up to $100 billion progressively as each gigawatt is deployed, which the company described as its biggest commitment of this kind.
• NVIDIA CEO Jensen Huang said 10GW equates to roughly 4–5 million GPUs.
• IT之家 cites reporting that OpenAI has spent $16 billion on rented compute this year and is pursuing additional multi‑year capacity, including a reported $100 billion backup‑server plan and a separate five‑year $300 billion Oracle procurement.