Meta secures Crusoe deals for 1.6 gigawatts of AI computing

Meta Platforms has locked up new computing capacity from Crusoe for two data centers in Childress, Texas, and Warrenton, Missouri, in a deal that underscores how aggressively the company is chasing the physical infrastructure behind artificial intelligence. The combined sites are expected to deliver roughly 1.6 gigawatts of capacity, a scale that signals how far the AI arms race has moved beyond software and into power, land and server availability.

The numbers are striking. A single gigawatt can power about 750,000 U.S. homes, so Meta’s reported commitment points to an appetite for electricity and cooling that rivals major industrial systems. The timing also matters: the company is adding compute as demand for AI services continues to outpace supply, and as the biggest technology firms compete for chips, data centers, transmission capacity and the land needed to build them.

Meta has already signaled how expensive that race will be. The company has said it plans to invest $600 billion in U.S. infrastructure and jobs over the next three years, and it has launched America’s Workforce Academy, a four-week training program aimed at preparing people for skilled jobs tied to AI infrastructure. Meta has also said it is investing $115 million in a data-center technician training program, a reminder that the bottlenecks in AI are increasingly as much about labor and construction as they are about models and algorithms.

Crusoe has emerged as a major beneficiary of that shift. On June 9, the company said its contracted AI infrastructure capacity had reached 4.9 gigawatts across data centers and Crusoe Cloud, with a development pipeline exceeding 40 gigawatts. It has also built out its presence in Abilene, Texas, where it announced a 900-megawatt expansion on March 27 that lifted projected total capacity across the full campus to 2.1 gigawatts. Crusoe said the first phase of that flagship data center was live on September 30, 2025.

The reported Meta agreements fit a broader strategy in which control over compute is becoming control over the AI economy’s chokepoints. Crusoe describes itself as an AI infrastructure company focused on energy-first data centers, and its materials stress the need to plan for community, water and energy from day one. That makes the Meta deal more than a procurement win. It is another sign that the next phase of AI competition will be decided by who can secure the electricity, cooling and construction scale to keep expanding.