New research · Nature Sustainability, 2026
The cloud is a place.
From a new peer-reviewed study in Nature Sustainability, Nov. 16, 2026.
New research · Nature Sustainability, 2026
From a new peer-reviewed study in Nature Sustainability, Nov. 16, 2026.
01 · Power
America’s biggest data centers now use as much electricity as entire states. Where that power comes from makes all the difference.1
We traced the electricity of America’s 403 hyperscale data centers back to the power plants that supply them.
02 · Air
The way a data center is powered can affect the air people breathe, the water communities rely on, and the climate we all share.2
And the wind can carry what the stacks release across county and state lines.
03 · People
Yet the data on these impacts is rarely publicly available.3
Each dot is a census tract: a neighborhood of about 4,000 people.
04 · The explorer
Pick a site and the explorer traces its power, maps the plume, counts who breathes it and checks the water, before anything is built.
05 · Every side
Developers, big tech companies, communities, local officials and regulators. They need one set of facts, measured independently.
The study · 1 of 5
hyperscale data centers, tracked from May 2024 to April 2025, and the number is expected to increase.
The study · 2 of 5
21–31 millionmetric tons of CO₂, two to three times an earlier estimate for 2018.
The study · 3 of 5
68–99terawatt-hours of electricity used by hyperscale data centers across the country in a year. The high end approaches Indiana, Michigan or Tennessee.
The study · 4 of 5
54%of the power came from fossil fuels, mostly natural gas and coal, and about a quarter of the demand was in grid regions dirtier than the average fuel mix.
The study · 5 of 5
2×A unit of power in Wyoming carried more than twice the carbon of one in Oregon.
The Project
The study measured power and carbon across the country. Now we are turning that science into something people can act on: one set of independent facts for any proposed data center, before anything is built.
The platform
An objective, science-based platform for every side of the decision.
We are building an objective, science-based platform that integrates rigorous, peer-reviewed evidence about the impacts of data centers on electricity, greenhouse gases, air quality, water, noise, and economic and community benefits.
We turn rigorous health, climate, and energy data into evidence-based solutions that lower the risks of siting and securing computing power. The platform is also designed to enable the different stakeholders to make decisions on how and where to build a data center, so that communities can thrive, data center developers can complete their projects, and regulators can rely on objective data about compliance.
The explorer
Pick a site and the explorer traces its power, maps the plume, counts who breathes it and checks the water, before anything is built.
The explorer turns this research into evidence for a single site: air quality and public health, water use, noise, emissions, energy costs and local economic value, with every assumption visible and adjustable.
The same evidence, for every side. The explorer takes no side on whether a project should go ahead. It gives developers, big tech companies, communities, local officials and regulators the same facts.
The Researchers
Our lab traces air pollution back to its sources. We did it for bitcoin mining. Now, AI data centers.
We use AI every day, and we believe hyperscale data centers can be built responsibly. Where a data center goes, what powers it and how it’s cooled can all cut carbon and pollution. Those choices need to be based on independent evidence and reflect the needs of the surrounding communities.
The study was led at the Harvard Data Science Initiative. The explorer’s models are built on the same research.

Lead author
Postdoctoral Research Fellow at Harvard T.H. Chan School of Public Health, with a PhD in Artificial Intelligence. He studies how AI infrastructure impacts energy, the environment, and public health for better informed policies.

Senior author
Clarence James Gamble Professor at the Harvard T.H. Chan School of Public Health and Director of the Harvard Data Science Initiative. Her work has informed U.S. air quality standards.
TIME100 Health · Member of the National Academy of Medicine
Our research is published in
Nature Sustainability, Science, The New England Journal of Medicine, Nature, JAMA, Nature Communications, Science Advances
This study measured power and carbon. Air, water, noise and health come next. Get the findings first.
The science behind our platform →The explorer turns this research into evidence for a single site: air quality and public health, water use, noise, emissions, energy costs and local economic value, with every assumption visible and adjustable. If you are planning, reviewing or living near a project, we would like to hear from you.
Who are you?
the explorer · a preview · demo data
Only Compare and the map are open in this preview.
built on research published in Nature Sustainability Science The New England Journal of Medicine Nature JAMA Nature Communications Science Advances