Skip to content
Friday, October 23, 2026 · Today’s edition, 6:30 a.m. ETFri, Oct 23 · 6:30 a.m. editionSign inMy Desk 0EditionsSign in

OpenAI's data center chief exits as $750B buildout churns

The latest data center exit adds to the churn as OpenAI heads toward an IPO with a $750 billion compute budget.

At a glance

  1. Malone arrived from Meta, where he spent nearly five years as a distinguished engineer, after nearly 12 years at Google focused on data center infrastructure and IT hardware R&D. He worked alongside Keith Heyde, head of infrastructure and another Meta alum, until Heyde left earlier this year.
  2. The exits frame the reorganization OpenAI made in March, when it split compute and infrastructure work into three teams: one designing data centers, one handling commercial partnerships with cloud providers and chip companies, and one managing the data centers OpenAI uses.
  3. The churn is not confined to infrastructure: chief revenue officer Denise Dresser, chief operating officer Brad Lightcap, and Fidji Simo, described as Sam Altman's second-in-command, have all left the company, Data Center Dynamics reported.

The Wall Street Journal first reported that Chris Malone, OpenAI's head of data centers, has left after a year and a half, Data Center Dynamics said. He is the latest senior departure from the group carrying a roughly $750 billion compute buildout.

Malone arrived from Meta, where he spent nearly five years as a distinguished engineer, after nearly 12 years at Google focused on data center infrastructure and IT hardware R&D. He worked alongside Keith Heyde, head of infrastructure and another Meta alum, until Heyde left earlier this year. Peter Hoeschele, who ran the Stargate joint venture, departed around that same stretch with several AI infrastructure senior leaders.

The exits frame the reorganization OpenAI made in March, when it split compute and infrastructure work into three teams: one designing data centers, one handling commercial partnerships with cloud providers and chip companies, and one managing the data centers OpenAI uses. The org chart has since filled in from outside, with Intel's Sachin Katti now managing everything tied to Stargate, while Spas Lazarov, formerly Apple's director of data center engineering, was named head of data center engineering in January. Uday Ruddarraju, previously at xAI, was promoted to CTO of compute in July, and Brent Mayo, another xAI hire, heads data center build and delivery.

The churn is not confined to infrastructure: chief revenue officer Denise Dresser, chief operating officer Brad Lightcap, and Fidji Simo, described as Sam Altman's second-in-command, have all left the company, Data Center Dynamics reported. These departures arrive as OpenAI prepares to go public next year and, last month, raised projected compute spending to around $750 billion through 2030.

That puts the exits on a buildout now set by power politics and permitting queues rather than available capital—the people leaving are the ones who navigate exactly those constraints. The March reorganization was an attempt to make the buildout bigger than any single executive, but the personnel record since then shows how hard that is to enforce. The bench is now Katti, Lazarov, Ruddarraju and Mayo, and it has been rearranged repeatedly over the past year. An IPO prospectus will be written while that lineup is still being tested, and that is the kind of execution risk public investors will price.

Reported by the Data centers & fiber desk from Data Center Dynamics. Published . Corrections

Related

Infrastructure

OpenAI names Uday Ruddarraju to run compute and Stargate, replacing Katti

Sachin Katti, who took over the Stargate effort in March, is expected to remain as an adviser.

Infrastructure

Nvidia hires former OpenAI data center head Chris Malone as VP, DSX Platform

Data Center Dynamics reports the move; Nvidia this week launched DSX POD, a containerized version of its Vera Rubin-powered reference design.

Infrastructure

Crusoe files for a $4.8bn data center campus in Jayton, Texas

The two buildings are slated to go live in 2029, and the filings name no tenant, landowner or power source.

Infrastructure

Aleria plans 8,640 Blackwell GPUs for 25MW US site

Aleria plans to expand the US build to 16,000 GPUs and calls the UAE Vera Rubin deployment one of the first in the region.

Infrastructure

Foresight signs term sheet for 50.01% of Polish data center developer

Warsaw is planned at 130MW on a secured 32,000 sqm plot with an option on a further 22,000 sqm.

Private Infrastructure Dailyfrom Private Wealth Daily

Energy transition, digital infrastructure and core-plus funds. Weekdays at 7:25 a.m. ET.

Free. Weekday mornings. Unsubscribe from any edition in one click.

From other sections