Overview

• The companies signed a letter of intent to deploy at least 10 gigawatts of Nvidia systems, with investment released progressively and final terms expected in the coming weeks.
• OpenAI will treat Nvidia as its preferred compute and networking partner, and both will co‑optimize hardware and software roadmaps for future model development.
• The first phase targets the second half of 2026 using Nvidia’s Vera Rubin platform, marking the start of the multi‑gigawatt rollout.
• Nvidia CEO Jensen Huang said 10 gigawatts equates to roughly 4 million to 5 million GPUs, underscoring the scale of the planned buildout.
• Reports say OpenAI would pay cash for systems while Nvidia takes a non‑controlling equity stake, with an initial $10 billion tied to a definitive purchase agreement, and observers noted potential antitrust scrutiny of the tie‑up alongside OpenAI’s existing work with Microsoft, Oracle, SoftBank and Stargate.