Chevron (NYSE: CVX) and Microsoft (NASDAQ: MSFT) have announced a 20-year power purchase agreement to construct a dedicated natural gas-fired power plant in West Texas. The project, named Project Kilby, spans more than 2,000 acres in Reeves County near Pecos, deep in the Permian Basin, and will supply electricity exclusively to a Microsoft data center campus—making it one of the most significant energy deals between a U.S. oil major and a hyperscaler.

Bloomberg first reported the deal, which builds on an exclusivity agreement the two companies reached with investment firm Engine No. 1 in late March. Project Kilby is structured as a behind-the-meter installation, with the power plant co-located at the data center rather than feeding electricity into the public ERCOT grid. This design reflects the priority tech companies now place on securing reliable, dedicated power supplies rather than depending on broader regional energy infrastructure.

Access to reliable electricity has become the most pressing constraint in the AI infrastructure race, outpacing concerns around chips, construction timelines, and regulatory permits. The scale of this agreement signals that the AI economy has moved well beyond software ambition, now requiring the securing of land and the locking in of long-term energy supply commitments. Chevron and Microsoft represent two distinctly different industries converging around the same strategic conviction: that demand for AI compute will be large and durable enough to justify infrastructure built to last decades.

The 20-year commitment marks substantial capital allocation from both companies, underscoring confidence that the buildout of AI infrastructure will remain a defining economic force well into the future. For energy companies like Chevron, the deal highlights new revenue opportunities from direct partnerships with technology firms hungry for stable, large-scale power supply. For Microsoft, securing a dedicated generation asset insulates its data center operations from grid volatility and positions the company to scale AI workloads without facing the electricity bottlenecks that have constrained competitors.