How Much of the United States Can Still Host New Hyperscale Data Centers? A Constraint-Based Feasibility Analysis

TL;DR

This study asks where in the contiguous United States additional hyperscale data centers could be physically and infrastructurally hosted under current constraints. It develops a constraint-first geospatial framework on a hexagonal grid and builds two unsupervised feasibility surfaces— similarity-based and kernel-density–based—trained on revealed hyperscale siting patterns and guided by power-grid adjacency, climate risk, and land-use factors. The analysis identifies a limited feasible land envelope and estimates tens of gigawatts of physically feasible capacity (approximately $24$–$78$ GW under similarity and $26$–$106$ GW under KDE), substantially lower than naive land-availability assumptions. The results offer a conservative, national-scale ceiling for infrastructure planning under present-day conditions and constraints.

Abstract

The rapid expansion of hyperscale data centers, primarily driven by cloud computing and generative AI is placing growing pressure on electricity systems, land, and climate-sensitive infrastructure. While existing maps document where data centers are currently located, a major unanswered question remains: where can hyperscale data centers still be built under present-day physical, infrastructural, and environmental constraints? Here we address this question, focusing on the United States, using a national-scale, constraint-first geospatial framework that infers feasibility from revealed hyperscale siting patterns rather than from demand forecasts or optimization assumptions. By combining power-grid adjacency, environmental limits, land-use constraints, and climatic constraints within a uniform hexagonal spatial system, we estimate the feasible hyperscale hosting capacity. Our presented approaches converge on a limited feasible land envelope, implying a substantial contraction relative to naive land-availability assumptions. Based on observed build-out patterns, we estimate that total physically feasible U.S. hyperscale capacity lies in the tens of gigawatts rather than the hundreds. The results of this piece are intended to support national-scale reasoning about infrastructure feasibility under modern constraints.

Figures (5)

• Figure 1: USA-wide visualization of current data center locations and power transmission lines.
• Figure 2: Illustrative subset of the national data center dataset within the Northern Virginia hyperscale corridor. Points show existing data center locations scaled by upper-bound sustained power demand, while lines indicate high-voltage transmission infrastructure.
• Figure 3: Mean July temperature aggregated to H3 resolution 4 hexagons.
• Figure 4: Environmental and infrastructural context layers: elevation, built-up intensity (log scale), surface water share, and distance to power generation facilities.
• Figure 5: Comparison of feasibility envelopes derived from (top) similarity-based and (bottom) KDE-based approaches. Darker regions indicate hexagons selected under multiple feasibility thresholds.