63/100

Moderate evidence

ProblemPublished recordClimateProblem Space

Artificial intelligence raises operational carbon emissions from data centers

Truvace record
TRV-2026-2878raw data (JSON)
Version
v1Published
First published
Evidence type
Observational study
Sources
1 sourcePeer-reviewed
sha256
70867cdadc46…check it

Artificial intelligence is expanding computing demand, yet how this expansion affects the operational carbon emissions that arise from purchased electricity at data centers remains poorly quantified. Existing work has not systematically isolated this operational-phase outcome or examined the mechanisms, thresholds, and spatial patterns that shape it. Here we show that artificial intelligence development raises operational carbon emissions from data centers. A bottom-up accounting of 1,090 data centers across 162…

The problem

A study of 1,090 data centers in 162 Chinese cities from 2015 to 2024 found that a one-standard-deviation rise in an AI index was associated with an extra 68.82 million tonnes of operational carbon emissions.

What this doesn’t fix

The 68.82 million tonne figure is an association rather than a proven causal effect, and the data cover only Chinese cities from 2015 to 2024.

Sources

1 on file
  1. Environmental Science and Ecotechnologydoi:10.1016/j.ese.2026.100789
    Peer-reviewed2026-11-01

Cite this record

v1

Truvace Impact Record TRV-2026-2878, v1: “Artificial intelligence raises operational carbon emissions from data centers.” Truvace, 2026-11-13. /record/TRV-2026-2878 (accessed on the day you copy it). sha256 70867cdadc468f90…

Change history

1 version

Every change made to this record since it was first published, and why. Nothing has changed since it went up.

  1. Publishedv170867cdadc46…

    Certified into the record

Check this record hasn’t been altered

Check this record hasn’t been altered
How to check it yourself, without trusting this page

Download any version from /api/record/TRV-2026-2878 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.