A signable evidence standard
Support accessible, dated and verifiable public records without endorsing every conclusion.
Read the principles →Research hub
Data-center research:
air, energy and permits.
Explore source-linked studies of data-center air pollution, public health and electricity use alongside permit records. Each review separates published findings from conclusions the source cannot support.
8independent source reviews
10annotated researchers
194issued permit rows analyzed
Review or contribute →EVIDENCE REVIEWS
Read the claim.
Inspect the source.
Each review separates an institution’s published findings from our interpretation, identifies the unit of analysis and states what the source cannot prove.
01 · Institutional interview
AI data centers, air pollution and health
Primary source: Harvard T.H. Chan School of Public Health · not a project partner
Read evidence review →02 · Research summary and preprintThe public-health costs of AI infrastructure
Primary source: Caltech and University of California, Riverside · not a project partner
Read evidence review →03 · Report to CongressU.S. data-center energy use through 2028
Primary source: Lawrence Berkeley National Laboratory / U.S. Department of Energy · not a project partner
Read evidence review →04 · State legislative analysisWhat Virginia’s data-center review found
Primary source: Virginia Joint Legislative Audit and Review Commission · not a project partner
Read evidence review →05 · Regulatory and epidemiological evidencePM2.5, mortality and responsible claims
Primary source: U.S. EPA / Harvard T.H. Chan School of Public Health · not a project partner
Read evidence review →06 · State utility-planning lawVirginia resource planning, what the 2026 law now requires
Primary source: Commonwealth of Virginia · not a project partner
Read evidence review →07 · Regional transmission organization forecastDominion-zone load growth, what PJM's 2026 forecast actually projects
Primary source: PJM Interconnection · not a project partner
Read evidence review →08 · Federal reliability orderComputational loads, reliability standards and the 2026 deadline
Primary source: Federal Energy Regulatory Commission · not a project partner
Read evidence review →RESEARCH PATHS
Eight ways into
the evidence
Each path separates what a primary record establishes from analysis, expert context and policy judgment.
10 source profiles
Expert research map
Annotated work on air pollution, health, water, energy, computing and Virginia policy.
Open research map →Dataset analysisWhat 194 permits show
Geographic concentration, recent issue years and program-type findings with stated limits.
Read analysis →40 source-table rowsVirginia air monitoring
DEQ's July 17 CO, NO2 and PM2.5 statistics, deployment history and instrument limits.
Open monitoring data →Open verification queueFacility identity pilot
A conservative path from state permit records to EPA facility identifiers without name-only matching.
Review the queue →Methods paper · Version 1.0Public-records method
A citable account of acquisition, transcription, identity resolution and quality assurance.
Open publication →Policy Submission 001 · submittedVirginia Energy Plan
The exact stakeholder-survey response, with a clear boundary between submission and government support.
Read submitted text →Evidence Brief 001Virginia generator rules
What the July 2026 guidance changes and why issue date alone cannot determine applicability.
Read evidence brief →Editorial standardHow claims are bounded
Discovery, normalization, source verification, correction and status labeling.
Inspect methodology →Review one defined scope
Correct a field, review a method or contribute a primary source without a blanket endorsement.
Ways to participate →Know who is—and is not—involved
Current status, funding boundaries and the difference between a citation and an advisor.
Project disclosures →