NVIDIA to Invest $100 Billion in OpenAI's Computational Infrastructure

NVIDIA and OpenAI have announced a strategic partnership aimed at tackling one of the most critical challenges in modern AI: the severe shortage of computing power. Under the agreement, NVIDIA plans to invest up to $100 billion in OpenAI. However, this capital will be primarily used by OpenAI to purchase NVIDIA's own graphics processing units (GPUs) and hardware to build out new data centers.

The plan calls for OpenAI to develop a network of data centers with a combined capacity of at least 10 gigawatts. This staggering scale is equivalent to the power consumption of several large metropolitan areas. According to NVIDIA CEO Jensen Huang, achieving this will require an estimated 4 to 5 million GPUs — a figure that rivals the total number of chips the company expects to ship throughout the current year.

The first phase of the collaboration is slated for the second half of 2026. It is then that OpenAI will begin deploying its new AI clusters based on NVIDIA's next-generation Vera Rubin accelerators. NVIDIA's investments will be disbursed in stages, aligned with the rollout and commissioning of the new infrastructure. As part of the partnership agreement, NVIDIA will become OpenAI's priority hardware supplier.

Analysts point out that this cooperation model is mutually beneficial. NVIDIA not only secures a massive, guaranteed outlet for its products but also solidifies its market position amid growing competition from AMD and major cloud providers. For OpenAI, the partnership alleviates the acute challenge of securing access to scarce and expensive AI accelerators, which are essential for training next-generation models.

This alliance between two industry leaders underscores that the race for more powerful artificial intelligence is entering a phase where decisive advantages will come not only from algorithms but also from access to colossal energy and infrastructure resources. NVIDIA had previously entered another strategic partnership, though on a smaller scale of $100 billion, to jointly develop several generations of new processors.