Chris Malone lasted about 17 months as OpenAI's head of data centers. The utility contracts he was hired to help land run 25 years.
Malone has left the company, The Wall Street Journal's Anissa Gardizy reported Monday. An OpenAI spokesperson confirmed the departure to Bloomberg the same day. Neither Malone nor the company offered a reason, and Bloomberg reported he did not immediately respond to a message. Absence of a stated reason is the norm at this level, and it is evidence of nothing in particular.
The job was already being split before he left
Malone joined in March 2025, two months after OpenAI announced the Stargate Project with Oracle and SoftBank. He arrived from Meta, where he spent nearly five years, and before that more than a decade at Google, working on data center infrastructure as a distinguished engineer. That is close to the only resume that qualifies someone for this work, and not many people hold it.
Earlier this year, OpenAI reorganized around him. Malone stopped reporting directly to president Greg Brockman and began reporting to Sachin Katti, the former Intel chief technology and AI officer who joined OpenAI in late 2025 and now serves as vice president of compute strategy. Malone became co-head of a data center technical engineering team rather than sole owner of the buildout.
The seat is not being backfilled as one seat. Per the Journal, OpenAI has appointed a chief technology officer for computing capacity, has not named that person publicly, and has revived a push to lease entire finished facilities rather than build them from raw land. Three executives now hold pieces of the brief: Uday Ruddarraju leads the data center team, Brent Mayo runs build and delivery, and Spas Lazarov leads data center engineering.
In a statement to TechCrunch, OpenAI said it had "recently reorganized" its "infrastructure organization to support the scale and pace of our work," adding: "We have a strong, deeply experienced data center team in place, with clear leadership and the technical expertise to execute our plans."
The churn is real, and mostly ordinary
Business Insider has tallied 13 executive departures at OpenAI in 2026. Brad Lightcap, chief operating officer for eight years, said on Aug. 11 he was leaving to start something new. Denise Dresser, chief revenue officer for roughly eight months, was replaced two weeks ago by Dali Rajic of the Alphabet-owned security firm Wiz. Fidji Simo, effectively the number two, stepped down in July citing chronic illness.
Brockman has argued the scrutiny is disproportionate, telling the Journal that the "spotlight" on OpenAI means "every departure gets scrutinized in a way that it doesn't otherwise." He has a point. Companies remake their benches before a listing, and chief financial officer Sarah Friar told staff last week that OpenAI intends to go public in 2027 or sooner. Equity vests, and people who survive an intense build phase often leave when the next phase looks like procurement rather than construction.
What is not ordinary is the workload
The honest story is not that one executive left. It is what the role has become.
OpenAI has raised projected compute spending to roughly $750 billion through 2030, up from about $600 billion earlier this year, against Sam Altman's headline figure of $1.4 trillion in total infrastructure commitments. Its computing capacity more than tripled during 2025, to roughly 1.9 gigawatts. The Oracle agreement alone is a $300 billion cloud contract starting in 2027 requiring about 4.5 gigawatts.
Altman has been candid that the physics are the hard part. In his "Abundant Intelligence" post he set the goal as a factory "that can produce a gigawatt of new AI infrastructure every week," then conceded: "The execution of this will be extremely difficult; it will take us years to get to this milestone and it will require innovation at every level of the stack, from chips to power to building to robotics."
Georgia is the live test. Project Camellia, OpenAI's planned campus on 1,400 acres in Effingham County, is budgeted above $20 billion, and Katti has said the fully developed site would probably exceed $30 billion. It has secured 3.2 gigawatts from Georgia Power, phased between 2028 and 2032 — roughly three times the electricity Atlanta consumes. This month, Georgia Power voluntarily delayed the start of its 25-year service agreement after Public Service Commission staff raised objections, resetting the deadline to Aug. 26.
Commissioner Peter Hubbard told The Current GA he wanted a fuller review: "The magnitude of this contract, the single contract, it is far larger than what we typically approve in the commission under integrated resource plans when we're planning out 20 years into the future for load growth."
Nothing in the reporting connects Malone's exit to any of that. But it is the environment his successors inherit: interconnection queues, transformer lead times, construction labor and regulators who no longer rubber-stamp load requests.
Financing is equally crowded. SoftBank is in talks over a $10 billion to $20 billion bond sale, possibly as early as September, partly to refinance the $40 billion bridge loan it took in March. Its final $10 billion installment into OpenAI is due Oct. 1, taking its total to roughly $64.6 billion and about a 13 percent stake.
What to watch
First, the name of that unannounced chief technology officer for computing capacity — continuity in this function lives in individual relationships with utilities, grid operators and contractors. Second, whether announced timelines slip at Abilene, Georgia or Ohio; those public dates are the only measurable test of whether the reshuffle mattered. Third, the prospectus. A company listing in 2027 must put its compute commitments in writing for regulators, and key-person risk is a standard disclosure.