Drone Flyover of Project Stargate Datacenter by Citrini Research (www.citriniresearch.com)

Citrini Research released drone footage of the Stargate datacenter campus in Abilene, Texas, visually confirming the mammoth physical scale behind today’s AI buildout. The site will host eight buildings by mid‑2026 (≈4 million sq ft) with a combined capacity of ~1.2 GW — enough electricity to power a city like Seattle — and planners are discussing another 600 MW expansion. Each building is being engineered to support up to 100,000 GPUs on a single integrated network fabric, with two buildings optimized for direct‑to‑chip liquid cooling and air cooling. The footage and reporting show heavy industrial installs at scale: generators, transformers, chillers, 2N redundant electrical systems, fuel tanks, batteries, copper busways and massive cable trays — infrastructure that produces thermal plumes visible on weather radar and already hosts early training/inference runs. For the AI/ML community this is a concrete reminder that model progress is now constrained by physical infrastructure as much as algorithms or chips. The buildout implies sustained, large demand across semiconductors, networking, power equipment, construction and cooling solutions, and it elevates natural gas and dispatchable power as critical enablers. With tens to hundreds of similar projects planned nationwide (and hyperscaler facilities like Meta, Microsoft, Google/Anthropic and xAI also expanding), expect major supply‑chain pressure, capital intensity, and regional utility impacts — all of which will shape where and how large‑scale model training can grow.