OpenAI's Data Center Chief Exits Amid Broader Leadership Exodus Before IPO

Chris Malone, OpenAI's head of data centers, departed the company last week after less than 18 months on the job — the latest in a string of senior-level exits rattling the AI giant as it navigates a path toward a 2027 IPO. Malone, a veteran infrastructure engineer who previously served as a distinguished engineer at both Meta and Google, joined OpenAI in March 2025 specifically to oversee the buildout of the Stargate initiative, the sprawling data-center effort announced alongside Oracle and SoftBank.

Stargate got off to a turbulent start, prompting OpenAI to pivot away from building its own facilities toward signing deals with cloud providers. The company is now reviving an internal construction strategy — leasing entire data centers rather than simply renting compute from hyperscalers — but that effort is being led by others, not Malone. The organizational reshuffle earlier this year already marginalized his role, placing VP Sachin Katti atop the infrastructure group under President Greg Brockman, while Malone was repositioned as co-head of a narrower technical engineering and design team.

Malone's exit is part of a broader leadership drain that has accelerated in recent weeks. Chief Revenue Officer Denise Dresser, Chief Operating Officer Brad Lightcap, and Fidji Simo — who functioned as Sam Altman's de facto number two — have all left in short succession. The departures arrive at a sensitive moment: OpenAI is aggressively competing with Anthropic for enterprise customers while simultaneously ramping up its infrastructure ambitions to a staggering scale.

On the compute spending front, OpenAI has raised its projected outlay to roughly $750 billion through 2030, up from an earlier estimate of around $600 billion. It also recently signed a 10-gigawatt data-center lease in Ohio with SoftBank's SB Energy, backed in part by an Nvidia financial guarantee. To stabilize its infrastructure leadership, the company promoted Uday Ruddarraju to chief technology officer of computing capacity in July, and brought in Brent Mayo — poached from Elon Musk's xAI — to keep construction projects on schedule. Both Ruddarraju and Mayo have direct experience building Musk's Colossus supercomputer cluster in Memphis.