NVIDIA and OpenAI Collaborate on Historic AI Infrastructure Initiative

In a significant development for the artificial intelligence sector, NVIDIA and OpenAI have announced a partnership to deploy a massive AI infrastructure powered by millions of NVIDIA GPUs. This collaboration aims to scale OpenAI’s computing capabilities with at least 10 gigawatts of data centers, according to NVIDIA.

Expanding AI Capabilities

The announcement was made during a discussion with CNBC, where NVIDIA’s founder and CEO Jensen Huang described it as “the biggest AI infrastructure project in history.” The initiative is set to transition AI from research labs to real-world applications, leveraging NVIDIA’s advanced systems, including the NVIDIA Vera Rubin platform.

OpenAI CEO Sam Altman emphasized the unmatched scale and speed that NVIDIA brings to this initiative, stating, “There’s no partner but NVIDIA that can do this at this kind of scale, at this kind of speed.” The partnership includes a commitment from NVIDIA to invest up to $100 billion in OpenAI as each gigawatt is deployed.

Meeting Growing AI Demand

OpenAI has experienced rapid growth since the launch of its ChatGPT, which reached 100 million users in record time. The company now boasts over 700 million weekly active users and offers advanced capabilities such as AI reasoning and support for multimodal data. This explosive growth necessitates a robust infrastructure to support both training and inference demands of its sophisticated AI models.

Altman noted that the increasing computational power would lower the cost per unit of intelligence while raising the frontier of AI capabilities. This expansion allows for more impactful use cases without forcing choices between critical applications like cancer research and education.

Long-Term Collaboration

The partnership builds on a long-standing relationship between NVIDIA and OpenAI, dating back to when NVIDIA delivered its first DGX system to OpenAI in 2016. Greg Brockman, president of OpenAI, highlighted the monumental leap in computational power, stating, “This is a billion times more computational power than that initial server.”

Huang expressed that this is merely the beginning of a global AI infrastructure buildout, predicting further expansions beyond the initial 10 gigawatts. “We’re literally going to connect intelligence to every application, to every use case, to every device — and we’re just at the beginning,” he said.

Image source: Shutterstock