Sachin Katti, OpenAI’s vice president of compute strategy and GPT-Infra, said on October 14, 2026 that he is leaving his role at the company. Katti wrote that his last day will be November 15, 2026, and that he will continue supporting OpenAI as an advisor after that.
“After an incredible year at OpenAI, I’ve decided to move on to build something new,” Katti wrote in a post on LinkedIn. He said AI is reshaping the entire computing stack and that he believes the industry is still in its early innings, adding that he is excited to bring together what he has learned across his career. He did not name a next venture and said he would have more to share soon.
Katti’s Account of the Compute Buildout
In his post, Katti said the past year had been extraordinary and described his work as helping orchestrate “one of the largest AI compute infrastructure buildouts in history.” He said his teams secured double-digit gigawatts of new compute capacity across projects he listed as Ohio, Portsmouth, Georgia and Trainium, along with several others, working with partners he named as NVIDIA, Broadcom, AMD, Oracle, Microsoft, Google, AWS, Cerebras and CoreWeave. He also said the GPT-Infra effort went from an idea to a real business with customers and partners in just three months.
Katti wrote that building AI infrastructure at that scale gave him a view into every dimension of the compute world, spanning silicon, systems, datacenters, energy, supply chains, financing and the economics of AI. He said he was enormously proud of what the teams accomplished together and grateful to the partners who helped make it possible.
Role and Career Path
Katti had stepped into the vice president role about two months earlier, writing in an earlier post that his remit covered the commercial strategy and execution for compute at OpenAI: widening access to capacity, bringing capacity online faster, improving the economics of what OpenAI deploys and maximizing the return on those investments. He also said he would lead GPT-Infra product efforts, applying OpenAI’s AI to problems across the AI infrastructure ecosystem from semiconductors and systems to clouds, and described what he called a recursive opportunity in which the AI OpenAI is building helps design, optimize and scale the infrastructure future systems will depend on. He thanked Greg and the OpenAI leadership team for the trust.
In his departure post, Katti traced his career from academia as a professor at Stanford to building startups as an entrepreneur, then running a large business and serving as chief technology officer at Intel before joining OpenAI. He closed by thanking Sam, Greg and the teams at OpenAI for the opportunity, trust and partnership.
The Georgia and Ohio Projects
Two of the buildout projects Katti referenced are documented in OpenAI’s own announcements. On July 22, 2026, OpenAI announced Project Camellia, a long-term datacenter project it is designing and developing in Effingham County, Georgia, contracting with Georgia Power Company for 3.2 gigawatts of power to be delivered in phases between 2028 and 2032. OpenAI committed that electricity rates will not go up for residents because of the project, that it will use a closed-loop water system, and that an annual publicly available audit by an independent firm will be conducted and released. The company also said it will provide $80 million in community benefits over the life of the project and up to $71 million in Codex credits for eligible Georgia students. Codex is OpenAI’s agentic coding tool, available within ChatGPT.
On August 17, 2026, OpenAI announced an agreement to secure approximately 8 gigawatts-IT at the PORTS-Pike Technology Campus in Pike County, Ohio, the site of the former Portsmouth Gaseous Diffusion Plant, working with SB Energy, NVIDIA and the U.S. Department of Energy. OpenAI said the project is expected to create 35,000 construction jobs during its six-year buildout through 2032 and 2,500 long-term operating jobs. Under the announced structure, SB Energy will build, own and operate the data center under a 20-year lease to OpenAI; the first 800 megawatts are expected to become available in 2028; and NVIDIA will invest $1.5 billion in SB Energy while providing credit support for the initial 4.25 IT-GW buildout. The site will exclusively host NVIDIA AI compute infrastructure. OpenAI also said it will invest $40 million in a community grant fund, building on SB Energy’s previously announced $40 million commitment, and will make up to $84 million in Codex credits available to approximately 844,000 eligible Ohio college, community college and technical school students during the 2026–2027 academic year.
Custom Silicon Work
On August 25, 2026, OpenAI reported first results for Jalapeño, its first custom inference chip. The company said the chip delivered 1.5 to 1.9 times more AI work per watt at peak throughput and 1.7 to 3.6 times lower end-to-end latency than comparison systems across three public models: GPT-OSS 120B, DeepSeek R1 and Kimi K2.5 1T. OpenAI said it tested Jalapeño on InferenceX, a public benchmark from SemiAnalysis, and that the chip is rated at 700 watts while its measured sustained power remained at or below 550 watts on the workloads tested.
A month before his departure announcement, Katti wrote that OpenAI’s compute strategy pairs deep partnerships with the industry’s leading accelerator, cloud, systems and energy companies with custom infrastructure informed by OpenAI’s own models and production workloads. OpenAI said it plans to begin deploying Jalapeño within its compute infrastructure by the end of 2026 as the first generation of a multigenerational roadmap, with Gen 2 deep in development and Gen 3 taking shape.
