Microsoft and Chevron Ink Natural Gas Deal for Texas AI Data Center

Microsoft and Chevron have signed a 20-year agreement to power “Project Kilby,” a massive data center complex in West Texas, using natural gas from the Permian Basin. According to project specifications, the facility will reach 2.67 gigawatts of capacity by 2028, operating independently of the public power grid to meet the energy demands of artificial intelligence.

Why is Microsoft moving toward autonomous natural gas power?

Microsoft is bypassing the traditional electrical grid to avoid the instability and rising costs associated with the U.S. energy infrastructure. Project Kilby will use GE Vernova turbines to generate its own electricity on-site, utilizing cheap gas sourced directly from the Permian Basin.

The project, developed alongside investment fund Engine No. 1, expects a final investment decision by the end of this year. At full capacity, the 2.67 gigawatt output could power more than 530,000 homes, illustrating the sheer scale of energy required for modern AI clusters.

How are AI inference companies like Groq scaling their infrastructure?

While giants like Microsoft build gigawatt-scale complexes, specialized AI firms are focusing on rapid, targeted expansion. Groq plans to quadruple its capacity to 200 megawatts by the end of 2027, expanding its existing network of 13 data centers.

This growth is fueled by a $750 million funding round closed in September 2025, which valued the company at $6.9 billion. Groq’s investor list includes major industry players such as Samsung, Cisco, and BlackRock, signaling a broader market bet on AI inference infrastructure.

Will AI expansion derail corporate net-zero commitments?

The reliance on natural gas suggests a conflict between AI growth and environmental goals. Microsoft has previously committed to covering 100% of its electricity consumption with renewable energy by 2030, but this target is now in question.

According to a report from Bloomberg in May, the company is considering delaying or abandoning that commitment due to the energy intensity of AI. Microsoft is currently racing against Alphabet and Amazon to dominate the AI sector, with plans to double its data center footprint within the next two years.

What happens to the U.S. power grid as data center capacity doubles?

The surge in AI infrastructure is putting unprecedented pressure on national energy supplies. Data from BloombergNEF indicates that data center capacity in the U.S. is projected to double by 2030.

This trend likely leads to higher electricity costs for residential consumers as utilities struggle to keep up with industrial demand. The shift toward autonomous plants like Project Kilby is a direct response to this scarcity, allowing Big Tech to secure energy without competing with the general public for grid access.

Comparison: Scale of AI Energy Infrastructure

Frequently Asked Questions