rewrite this title NVIDIA Powers OpenAI with 10GW to Train Next-Gen AI Models

Through a newly signed letter of intent, the two companies outlined a goal to deploy 10 gigawatts of computing infrastructure. OpenAI will use this setup to train and run future AI models.

NVIDIA is preparing to invest up to $100 billion. That funding will be used to set up the necessary computing equipment, expand data centers, and ensure sufficient power is available to run the systems.

The first part of this expansion is expected to begin operations in the second half of 2026 and will use NVIDIA’s Vera Rubin platform.

Jensen Huang, CEO and founder of NVIDIA, noted that this collaboration marks a major step forward, which is helping to launch a new generation of AI development.

Sam Altman, OpenAI’s CEO and co-founder, emphasized the importance of computing resources in shaping the future economy. He said the infrastructure they are building with NVIDIA will support new AI capabilities and help bring those tools to a wider audience.

As part of the agreement, NVIDIA will serve as a key technology provider to OpenAI. The two companies will work closely to align their hardware and software plans.

Recently, OpenAI announced plans to make changes to how ChatGPT handles sensitive conversations. What do the changes include? Read the full story.