Frontier Briefing No. 01 · Tuesday, December 15, 2026 · 10–11:30am ET · Free
Harvard Data Science Initiative
Frontier Briefing No. 01 From Harvard Data Science Review

The True Impact of AI Infrastructure Is Decided Before It’s Built. Here’s How to Measure It.

A data center is never just a building. Where it goes, what powers it and how it is cooled decide how much carbon its power carries and who lives with the pollution. See how Harvard researchers measure those impacts facility by facility, positive and negative, and how to weigh your options with independent evidence before anything gets built. Ninety minutes, live, with the researchers doing the work.

Prof. Francesca Dominici Dr. Gianluca Guidi
With Prof. Francesca Dominici Harvard T.H. Chan School of Public Health · Director, Harvard Data Science Initiative · Dr. Gianluca Guidi Harvard T.H. Chan School of Public Health
You walk away with a one-page memo for the decision in front of you, whether you are a community member, developer, utility, investor or regulator: the options on the table, which levers (site, power source, cooling) move the numbers and by how much, the tradeoffs in power, water, air and health, who carries them, the numbers to ask for, and what to do next.
Cost Free
Capacity 200 attendees
When December 15 Tuesday · 10am ET
Frontier Briefing No. 01

The frontier, first.

A Frontier Briefing puts you in the room with the researchers while the findings are still new. You hear the method from the people building it, ahead of the reports, the consultants and the headlines.

  • Early. The study behind this briefing is published November 16 in Nature Sustainability.¹ It has not reached a textbook or a course.
  • Live. Ninety minutes with the researchers, including your questions.
  • Applied. You leave with a one-page memo for the decision in front of you, from your side of the table.
How it runs

Learn the method from the researchers. Apply it from your seat.

Thirty minutes of prep with Paski, your AI tutor. Ninety minutes live with the researchers. Thirty minutes writing your memo.

1
Before · Self-paced

Your seat, read against the evidence.

Tell Paski where you sit: the facility, asset, territory, case or community in front of you. It shows you which impacts will matter and what to ask at the briefing.

About 30 minutes
2
Frontier Briefing · December 15

The Frontier Briefing.

The true impact measured, the method behind it, the levers that move it, and 20 minutes of your questions. Live with Prof. Francesca Dominici and Dr. Gianluca Guidi.

90 minutes live
3
After · Self-paced

Write your internal memo.

Work through your own decision with Paski, from your side of the table, starting from the same facts as every other side.

About 30 minutes

Recorded. Your work with Paski stays open afterward if you miss it.

Inside the briefing

Measure, attribute, trace, decide.

The study measured power and carbon across the country. The researchers are now turning it into one set of independent facts for any proposed data center, the same evidence for every side. This is about data centers as physical infrastructure, not the footprint of a single prompt or a company's own AI use.

  • Measure. Electricity and carbon at 403 hyperscale data centers, facility by facility, from May 2024 to April 2025: 68 to 99 terawatt-hours in a year, the high end approaching Indiana, Michigan or Tennessee.¹
  • Attribute. Carbon from the mix of power plants on each facility's regional grid, not a national average. 54% of the power came from fossil fuels, and a unit of power in Wyoming carried more than twice the carbon of one in Oregon.¹
  • Trace. The next step, now in development: for a single proposed site, trace its power, map the plume from the plants that supply it, count who breathes it, and check the water, before anything is built.
  • Decide. Where a data center goes, where its power comes from and how it is cooled can all cut carbon and pollution. See which lever moves the number, by how much, and who carries the cost. We bring the evidence; the decision is yours to make.

¹ Guidi, G., Dominici, F. et al., Assessing the Carbon Emissions and Energy Consumption of US Hyperscale Data Centers, Nature Sustainability, November 16, 2026.

Who leads it

Two researchers. One method.

The authors of a new peer-reviewed study in Nature Sustainability covering 403 of the country’s biggest data centers, walking through the method live, not a recording of them.

Prof. Francesca Dominici

Prof. Francesca Dominici

Harvard T.H. Chan School of Public Health · Director, Harvard Data Science Initiative

Clarence James Gamble Professor at the Harvard T.H. Chan School of Public Health and Director of the Harvard Data Science Initiative. Senior author of the study. Her work has informed US air quality standards. TIME100 Health, Member of the National Academy of Medicine.

Dr. Gianluca Guidi

Dr. Gianluca Guidi

Harvard T.H. Chan School of Public Health

Postdoctoral Research Fellow at the Harvard T.H. Chan School of Public Health, with a PhD in artificial intelligence. Lead author of the study. He studies how AI infrastructure affects energy, the environment and public health, and previously led the team's research on US bitcoin mining.

Frontier Briefing No. 01

The researchers, live. The evidence the debate is missing.

Tuesday, December 15, 10:00–11:30am ET. Free. Registration closes Monday, December 14 at 5pm ET.

Reserve your spot
Reserve your spot · free