OpenAI and NVIDIA have unveiled a massive collaborative effort aimed at constructing a new network of AI data centers. The plan involves creating computing facilities for OpenAI that will demand a colossal 10 gigawatts of power, all supplied by NVIDIA’s hardware. This infrastructure is intended to serve as the engine for OpenAI’s future artificial intelligence systems.
The partnership includes a significant financial commitment from NVIDIA, which has outlined its intention to invest up to $100 billion into OpenAI. This capital will be provided incrementally, corresponding with the completion of each segment of the computing build-out. The first installment of this new infrastructure, representing one gigawatt of capacity, is scheduled to be operational during the latter part of 2026.
The energy requirement of 10 gigawatts highlights the project’s immense scale. This amount of power is similar to what is needed to run a small industrialized nation. The data centers will be populated by millions of NVIDIA’s graphics processing units (GPUs), which are the critical components for training complex AI models.
Leaders from both companies emphasized the importance of this collaboration. NVIDIA CEO Jensen Huang noted that the two firms have driven each other’s progress for ten years. OpenAI CEO Sam Altman stated that advanced computing infrastructure is the essential foundation for the economy of the future, and that this project will help scale AI benefits to everyone.
This new agreement designates NVIDIA as OpenAI’s primary partner for computing resources. The companies will also coordinate their development plans to ensure their software and hardware evolve in sync. This new partnership exists alongside OpenAI’s current collaborations with other technology partners like Microsoft.