A recent report by 404 Media reveals that a small research team from Epoch AI is mapping the rapid growth of datacenters across the United States. This initiative employs open-source intelligence, including publicly available information and satellite imagery, to track facilities that often evade public scrutiny. The project aims to provide insight into the scale and pace of artificial intelligence (AI) development within the country.
Using interactive mapping tools, Epoch AI has documented the construction of datacenters that require substantial amounts of electricity and water. Many communities are often unaware of these facilities until construction is underway, leading to significant concerns regarding resource consumption and environmental impact. The interactive map includes visual markers that link to satellite views and project details, enhancing public visibility of these operations.
Highlighting Key Developments
One notable site is located in New Albany, Ohio, where Meta’s “Prometheus” datacenter complex has been established. Epoch AI estimates that the project has incurred costs of approximately $18 billion and draws an impressive 691 megawatts of power. This facility reflects Meta’s strategic shift towards AI, characterized by a combination of weatherproof tents, colocation facilities, and traditional datacenter buildings. Users of the map can explore a timeline showcasing the complex’s development, with satellite images illustrating the addition of new buildings and cooling systems over time.
Much of Epoch AI’s research focuses on the cooling infrastructure necessary for modern AI systems, which generate significant heat. According to the organization, “Modern AI data centers generate so much heat that the cooling equipment extends outside the buildings.” The team meticulously counts the number of cooling fans, measures their size, and evaluates their positions to estimate energy consumption accurately. This data is integrated into a custom model that helps infer compute capacity and construction costs.
Challenges and Future Goals
Despite its detailed analysis, Epoch AI acknowledges that its mapping project is incomplete. Variations in state and local disclosure laws mean some projects can evade public attention. Additionally, smaller facilities may not be captured in the dataset. As of November 2025, Epoch AI estimates that the current dataset represents around 15 percent of global AI compute delivered by chipmakers.
The map reveals markers across the nation, including one near Memphis, Tennessee, which identifies xAI’s Colossus 2 project. Epoch AI highlights that the company has installed natural gas turbines across the Mississippi border, likely to expedite approval processes for construction. The report also references earlier tweets from Elon Musk, indicating that 110,000 NVIDIA GB200 GPUs are operational at the site.
Even with extensive mapping efforts, there remain significant blind spots. Epoch AI emphasizes that “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.” Looking ahead, the organization aims to expand its search globally, shedding light on the infrastructure that is increasingly shaping the future economy, often without public visibility and awareness.
