NVIDIA and OpenAI sign $100 billion deal to supply new AI servers

Under the agreement, OpenAI will gain access to NVIDIA’s advanced computing resources, including supercomputers based on H100 GPUs. This will allow the company to build AI data centers with a total capacity of at least 10 gigawatts, where millions of GPUs will be used.

In parallel, NVIDIA plans to invest up to $100 billion in OpenAI, with funding to be provided gradually — with the commissioning of new data centers.

For OpenAI, known for its GPT language models, the agreement opens access to resources without which it is impossible to scale training and support the operation of global AI systems. For NVIDIA, this step consolidates its status as the main supplier of hardware for artificial intelligence and allows it to already integrate its developments into key industry models.

The collaboration extends beyond infrastructure to software: the companies intend to work together to optimize and accelerate the processes of creating AI products. This is especially relevant against the backdrop of competition from players such as Google DeepMind, Anthropic and Meta.

The final details of the deal are expected to be agreed in the coming weeks, but observers are already saying that the union of the two technology giants could determine the trajectory of the artificial intelligence industry for years to come.