OpenAI is losing its compute strategist. Sachin Katti announced on October 14, 2026, that he is leaving his role as the company’s vice president of compute strategy and GPT-Infra. His last day will be November 15, 2026, after which he will continue supporting OpenAI as an advisor.
“After an incredible year at OpenAI, I’ve decided to move on to build something new,” Katti wrote. The departure arrives as OpenAI tries to turn enormous infrastructure plans into operating reality—a task involving power contracts, semiconductor suppliers, data centers, financing, and the usual cheerful assumption that the grid will cooperate.
Uday Ruddarraju will replace Katti as OpenAI’s CTO of compute and lead the company’s overall compute organization. Ruddarraju joined OpenAI in July 2026 after serving as head of infrastructure engineering at Elon Musk’s xAI.
A year of infrastructure at industrial scale
Katti wrote on LinkedIn that he helped orchestrate “one of the largest AI compute infrastructure buildouts in history.” His teams secured double-digit gigawatts of new compute capacity across projects in Ohio, Portsmouth, Georgia, and Trainium, giving OpenAI a broad footprint for training and serving models.
The work involved partners including NVIDIA, Broadcom, AMD, Oracle, Microsoft, Google, AWS, Cerebras, and CoreWeave. Katti described the work as a tour through “every dimension of the compute world,” including silicon, systems, datacenters, energy, supply chains, financing, and the economics of AI.
“Building AI infrastructure at this scale has given me an incredible view into every dimension of the compute world, from silicon and systems to datacenters, energy, supply chains, financing, and the economics of AI,” Katti wrote. He also said, “I’m enormously proud of what our teams accomplished together and grateful to the partners who helped make it possible.”
One piece of that work, the GPT-Infra effort, moved from an idea to a real business with customers and partners in just three months. The speed matters because OpenAI’s compute organization is tasked with developing infrastructure to train and serve models, not simply drawing diagrams about it.
OpenAI is targeting roughly $600 billion in total compute spend by 2030 and is involved with multi-billion-dollar data center projects in Ohio, Texas, and Georgia. The infrastructure map includes the Stargate Oracle AI data center in Abilene, Texas, with an aerial view dated August 26, 2026.
Leadership turnover meets a giant spending plan
OpenAI announced Project Camellia in Effingham County, Georgia, on July 22, 2026. The company contracted Georgia Power Company for 3.2 gigawatts of power for the project, with delivery scheduled between 2028 and 2032.
That power commitment helps show the scale of OpenAI’s infrastructure ambitions, but it also makes the leadership transition more consequential. Katti’s remit reached across the technical and financial machinery needed to turn compute demand into physical capacity, while Ruddarraju now inherits the organization responsible for carrying that work forward.
Katti’s career spans academia as a professor at Stanford, startup building, a tenure as CTO at Intel, and his work at OpenAI. His next venture remains unnamed, but he framed the move as a return to the role he prefers: “I’m a builder at heart, and I’m looking forward to building something new.”
He also offered a clear handoff to his successor: “I’m proud of what we’ve built and have total confidence in Uday and the team to carry it forward.” In another version of the message, Katti said, “I have total confidence in Uday and the team to carry it forward.”
The transition comes during a wider period of leadership change at OpenAI. Fidji Simo, the former product and business chief, Brad Lightcap, the former OpenAI executive, and Denise Dresser, the former revenue chief, left the company in 2026, while Greg Brockman remains OpenAI’s president.
Katti’s exit does not remove him from OpenAI entirely; his advisor role keeps him connected to the company. Still, moving from leading compute strategy to building something new marks a notable shift just as OpenAI’s infrastructure plans approach a scale where every gigawatt, supplier, and data center becomes part of the business strategy.
Based on
