Live · 1,592 Projects · Updated 2026-04-22

US Data-Center Build-Out

National snapshot of disclosed data-center capacity, status, and operator concentration across all U.S. metros.

Read first. Underlying data is LLM-generated swarm research, medium confidence overall. Disclosed capacity is on 41% of rows; all MW totals are lower bounds. Dollar amounts, water consumption, and 22 other prompt-specified columns are missing from the delivered CSV.

Total Disclosed

–

across rows with MW

Pipeline MW

–

announced / permitting / approved / under construction

Operational MW

–

currently running, disclosed

Pipeline / Operational

–

forward-loaded build-out

Hyperscaler Operated

–

share of disclosed MW

US Footprint

Marker size scales with disclosed MW. Color reflects status. Lat/long is approximated; cluster, do not site-pick.

Pipeline (announced / permitting / approved / under construction) Operational Paused Cancelled / Rejected Other / Unknown

Status Mix

Counts vs. capacity, side-by-side. Counts treat all projects equally; MW shows where the build-out actually sits.

Projects by status

Capacity (MW) by status

Top states by total capacity

Stacked by phase: operational (blue), pipeline (green), paused/cancelled/rejected (coral).

Top operators by total capacity

Hyperscaler operators in bold. JV capacity (Stargate, Crusoe-for-OpenAI, etc.) routed through third-party developers is not re-attributed here.

Top metros by pipeline capacity

Pipeline only; phase split is unreliable at metro grain (multiple campuses per metro).

Project Table

Click any column to sort. Filter by name or pick a state / status. Showing rows with disclosed MW only.

– rows

Project	Developer	Operator	MW	St	City	Status	Status Date	Source

Data sourced from public filings, project announcements, and LLM-assisted research. Disclosed capacity covers approximately 41% of rows; total MW figures are lower bounds.

Dollar amounts, water consumption, and several originally specified columns are not present in the underlying CSV. Lat/long values are approximate and intended for cluster visualization only.