Former OpenAI head of data centers joins Nvidia to help lead DSX Platform team
Chris Malone announced his new role two months after reports said he had left OpenAI.
Two months after leaving OpenAI, Chris Malone, who used to lead data centers at the AI company, has joined Nvidia as vice president of DSX Platform.
In a post on LinkedIn, Malone announced his new role. He said that after spending a few weeks with the team, he was “even more convinced of the opportunity ahead.”
Malone spent 18 months at OpenAI. Before joining, he worked at Meta for nearly five years as a distinguished engineer. Before Meta, he spent nearly 12 years at Google, focusing on the company's data center systems and research and development for IT hardware.
At OpenAI, he worked with Keith Heyde, the head of infrastructure, who had also previously worked at Meta. Malone left OpenAI earlier this year.
Peter Hoeschele, who led the Stargate joint venture, also left around that time. Several senior leaders who worked on AI infrastructure left with him.
The team was brought together under Sachin Katti. But Katti, a former Intel executive, left earlier this month.
Uday Ruddarraju, who previously led teams at xAI, is now in charge of OpenAI's infrastructure and computing work. He has brought several former xAI employees to OpenAI. One of them is Brent Mayo, who has been named head of data center construction and delivery.
“AI is fundamentally changing how data centers are designed, built, and operated. The scale, power density, cooling requirements, and speed of deployment demand new approaches across the industry,” Malone wrote in his post.
“I see tremendous potential for DSX to help meet that challenge. By bringing computing, power, cooling, and operations together in a common platform, we can help partners build AI factories faster, use energy more efficiently, and get more useful computing from every megawatt.”
Nvidia first introduced its DSX AI Factory Platform in March 2026 at GTC San Jose. The platform is a guide for building data centers that support Nvidia's latest CPU-GPU racks, as well as its Spectrum-X Ethernet networking and storage systems.
This week, Nvidia said the DSX Platform could help customers build AI data centers that use up to 100 percent less water for cooling. The company also said these data centers could use up to 30 percent less power while producing the same amount of computing work.
To support the platform, Nvidia launched DSX POD at GTC Berlin this week. DSX POD is a modular version of Nvidia's reference design for AI data centers powered by Vera Rubin.
The units are built inside containers and put together and tested before they are delivered. Nvidia said they can be connected to a prepared site’s power, cooling, networking, and other support systems. This could reduce the time needed to set up a data center from six months to 14 days.