Pranav Hotkar
Houston, Texas – June 22, 2026 - Chevron has signed a 20-year agreement with Microsoft to provide dedicated natural gas-fired power for a major AI data center campus in West Texas, marking one of the largest long-term energy partnerships announced to support hyperscale artificial intelligence infrastructure.
The agreement centers on Project Kilby, a co-located power generation facility that will supply electricity directly to Microsoft's planned data center campus near Pecos, Texas. Rather than relying solely on the regional grid, the project will pair on-site power generation with the AI campus, a model that is gaining traction as hyperscalers seek reliable energy sources for increasingly power-intensive workloads.
According to Chevron, Project Kilby will ultimately deliver up to 2.67 GW of generating capacity through a phased buildout, with initial power expected in 2028. The Microsoft campus is projected to add approximately 2 GW of new data center capacity over the next five to seven years, representing one of the largest AI infrastructure developments currently planned in the United States.
The project is being developed in partnership with Engine No. 1 and will primarily utilize natural gas produced from the Permian Basin. Most of the electricity will be generated using turbines supplied by GE Vernova, with additional generation equipment provided by Solar Turbines, a Caterpillar subsidiary. Chevron expects to make a final investment decision before the end of 2026, subject to regulatory approvals and other customary conditions.
Unlike traditional grid-connected developments, the co-located design is intended to provide Microsoft with a dedicated, dispatchable power source while reducing pressure on the Texas electricity network. The facility is expected to connect to the grid at a later stage, allowing excess generation to support the broader ERCOT market when available.
The announcement reflects a broader shift across the data center industry, where energy availability has become a primary constraint on AI infrastructure expansion. As GPU clusters require significantly higher power densities than conventional cloud deployments, technology companies are increasingly securing long-term energy agreements or developing dedicated generation assets alongside new campuses. Recent research suggests accommodating future AI-driven data center demand will require substantial investments in new generation, transmission, and storage infrastructure.
For Chevron, the agreement represents a significant expansion beyond its traditional upstream business into large-scale power generation for digital infrastructure. The company said the project leverages abundant Permian Basin natural gas to provide reliable, long-term electricity for AI computing while creating a new market for domestic energy resources. Microsoft, meanwhile, continues to diversify its energy portfolio as it scales cloud and AI services, complementing previous investments in renewable and nuclear energy with dedicated natural gas generation for mission-critical infrastructure.
