Data Center Power Visualizations

AI Data Centers: Power Demand Visualizations

This page uses a sample of our Aterio U.S. data centers dataset (real project data) to explore how much electricity AI-related data centers are designed to use and how that load is distributed across states and providers.

Visualization 1: How big are most data centers in terms of power demand?

Visualization 2: Which providers dominate power usage in the top five states?

Underlying Data (Sample)

Sample of the project dataset used to generate the visualizations below. Each row is one data center.

Data Center Name	Provider	State	Total Power (MW)
HAII1 AlohaNAP Data Center	1547 Critical Systems Realty	Hawaii	1.0
Hawaii Data Center	DRFortress	Hawaii	4.883
HIL1 Data Center	OVHCloud	Hawaii	7.382
Portland Data Center 3 - (PTL3)	Digital Fortress	Oregon	1.0
Portland Data Center 2 - (PTL2)	Digital Fortress	Oregon	1.0
Portland Data Center	H5 Data Centers	Oregon	2.994
Downtown Portland Data Center 1 - (PTL)	Digital Fortress	Oregon	2.0
Portland OR \| PTOR1	1547 Critical Systems Realty	Oregon	6.0
Bellingham Data Center	Lunavi	Washington	0.8
Tacoma Data Center	Colocation Northwest	Washington	2.0
San Francisco	Data Canopy	California	0.8
SF1: San Francisco	Fortress Data Centers	California	2.994
Downtown Seattle Data Center - (SEA)	Digital Fortress	Washington	1.0
NYI Seattle SEA1	NYI	Washington	4.0
Seattle Data Center (SEA1)	Evocative	Washington	4.0
Seattle Data Center (SEA2)	Evocative	Washington	0.0
Data Center SEA 2	H5 Data Centers	Washington	1.0
Seattle 5th Avenue Data Center (SEA1)	Sabey Data Centers	Washington	30.0
Seattle Intergate West Campus (SEA2)	Sabey Data Centers	Washington	4.0
Seattle Intergate East Campus (SEA3)	Sabey Data Centers	Washington	8.0
San Francisco Data Center SF (200 Paul)	H5 Data Centers	California	2.994
Seattle Data Center	Colocation Northwest	Washington	15.0
ColoHouse Seattle	ColoHouse	Washington	0.0
DataBank SEA1 - Tukwila	DataBank	Washington	14.0
Sabey SEA1 - Seattle	Sabey Data Centers	Washington	15.0
Sabey SEA2 - Tukwila	Sabey Data Centers	Washington	17.0
Seattle	QTS Data Centers	Washington	40.0
Seattle (SEA1)	QTS Data Centers	Washington	100.0
Cologix WA SEA1	Cologix	Washington	0.25
Cologix WA SEA2	Cologix	Washington	1.0

Note: the full dataset contains 5,176 rows; this table shows the first 30 records.

Visualization 1: How Big Are Data Centers (by Power Demand)?

What this chart shows: We grouped every data center in the dataset into buckets based on its planned power demand (in megawatts). The chart shows how many facilities fall into each power range.

Key takeaways:

Most data centers are under 100 MW of power demand (the first four bars).
There is still a noticeable group of very large facilities in the 100–200 MW and 200–500 MW ranges.
This “long tail” of huge sites likely represents hyperscale AI and cloud campuses that matter a lot for grid planning and community impact.

In other words: the typical data center is mid-sized, but a smaller number of giant facilities are responsible for a big chunk of total electricity demand.

Visualization 2: Which Providers Dominate in the Top Power States?

What this chart shows: We focused on the five states with the highest total data center power demand (Virginia, Texas, Georgia, Arizona, Pennsylvania). Each bar shows the total power demand in that state, broken down by major provider.

Key takeaways:

Virginia and Texas are huge outliers in terms of total data center power demand.
In Virginia, a mix of “Company Not Disclosed” sites and Amazon AWS account for a large share of load.
Texas also has a large amount of undisclosed capacity plus big blocks from cloud and AI providers.
States like Georgia, Arizona, and Pennsylvania have smaller totals but are still meaningful clusters in the emerging AI infrastructure map.

This suggests that AI and cloud infrastructure is highly concentrated in a handful of states and dominated by a small set of providers. That concentration has direct implications for local grids, water, and communities.