Overview

• A joint letter of intent commits OpenAI to deploy at least 10 gigawatts of Nvidia systems for training and inference, with Nvidia investing up to $100 billion progressively per gigawatt and receiving an equity stake.
• Nvidia is designated OpenAI’s preferred compute and networking partner under a non‑exclusive arrangement that preserves its ability to work with rivals such as Anthropic or xAI.
• Analysts project an initial ~$10 billion disbursement to OpenAI within months and expect first gigawatt‑scale deployments in the second half of 2026 tied to Rubin/VR200 NVL systems.
• Market response has been broadly positive, with Barclays lifting its price target to $240 and Evercore to $225, as others maintain bullish ratings alongside cloud contract backlogs exceeding $1 trillion and multi‑trillion AI capex forecasts.
• Coverage highlights risks that include Nvidia’s revenue concentration—two direct customers made up nearly 40% last quarter—and concerns about circular financing, set against OpenAI’s extensive Oracle and CoreWeave commitments and reports of a new $10 billion Broadcom customer that analysts linked to OpenAI.