A recent report by 404 Media uncovers how a small research team is meticulously mapping the rapid expansion of datacenter infrastructure in the United States. The initiative, led by Epoch AI, a non-profit research institute, utilizes publicly available information and satellite imagery to document facilities that often go unnoticed in public discussions. This project aims to shed light on the scale and pace of artificial intelligence development in the country.
Mapping the Expansion of Datacenters
Datacenter construction has emerged as a contentious issue nationwide. These facilities demand substantial amounts of electricity and water, yet many communities remain unaware of their presence until construction is underway. Through its mapping efforts, Epoch AI provides visual markers over known sites, linking each to satellite views and detailed project information. For example, a green circle on the map indicates Meta’s “Prometheus” datacenter complex in New Albany, Ohio, which Epoch AI estimates has cost around $18 billion and draws 691 megawatts of power.
Epoch AI describes the complex as a combination of weatherproof tents, colocation facilities, and traditional datacenter buildings, highlighting Meta’s strategic shift towards AI. Users can explore a timeline that showcases the complex’s growth, revealing new buildings and cooling systems added over time through satellite imagery.
Understanding Cooling Infrastructure and Energy Use
A significant aspect of Epoch AI’s analysis focuses on the cooling infrastructure necessary for modern AI systems, which generate considerable heat. According to Epoch AI, many datacenters install cooling units outside their buildings or on rooftops. The team meticulously counts fans, measures their sizes, and assesses their configurations to estimate energy consumption.
Jean-Stanislas Denain, a senior researcher at Epoch AI, explains, “We focus on cooling because it’s a very useful clue for figuring out the power consumption.” The team’s model, while comprehensive, carries some uncertainty; variations in fan speed and configuration mean the real cooling capacity could be twice as high or half as low as estimates suggest.
Despite these efforts, the map remains incomplete, as state and local disclosure laws vary, and some projects deliberately avoid publicity. Epoch AI estimates that their current dataset represents approximately 15 percent of global AI compute delivered by chipmakers as of November 2025. As users zoom out on the map, they can spot markers across the country, including one near Memphis, Tennessee, which points to xAI’s Colossus 2 project. Epoch AI notes that the company installed natural gas turbines across the Mississippi border, likely to expedite approval processes.
The report indicates that approximately 110,000 NVIDIA GB200 GPUs are operational at this site, based on previous tweets from Elon Musk.
While detailed mapping efforts have provided significant insights, challenges remain. Epoch AI acknowledges, “Even if we have a perfect analysis of a data center, we may still be in the dark about who uses it, and how much they use.” To address this, the organization plans to expand its research globally, aiming to illuminate the infrastructure that is shaping the future economy—often without public visibility.
