(PDF) The data heat island effect: quantifying the impact of AI data centers in a warming world
Marinoni, Andrea et al, The data heat island effect: quantifying the impact of AI data centers in a warming world, March 2026, DOI:10.48550/arXiv.2603.20897 License CC BY-NC-ND 4.0
The article below is excerpted from the papers cited above.
The urban heat island (UHI) effect
plays a key role in the impact of anthropogenic activities on climate change and
global warming. In a recent paper Andrea Marinoni at the University of
Cambridge and their colleagues saw that the amount of energy needed to run a
data center had been steadily increasing and was likely to “explode” in the
coming years, so they wanted to quantify the impact.
From humble origins as rudimentary storage facilities to
their current status as the lifeblood of the Internet and the Cloud, data
centers have become foundational pillars supporting our digital lives.
Yet, their energy footprints and building structures
are having an impact on climate change.
The researchers utilized 20 years of satellite measurements
of land surface temperatures and cross-referenced the data against the locations of
more than 8,400 AI data centers. To eliminate the possibility of impact from
the urban heat island effect and recognizing that surface temperature could be
affected by other factors, the researchers eliminated data centers in or near
urban locations (like Loudoun County) and instead focused their investigation
on only on data centers away from populated areas.
The goal was to quantify the land surface temperature
increase caused by the establishment of an AI hyperscaler in a location, determine
the region of influence of this increase; and estimate the population affected
by the temperature increase.
 |
| from Marinoni, Andrea et al |
They found that the average land surface temperature increase across the data centers was 2.07°C in the months after an AI data center started operation. The land surface temperature increase minimum and maximum were 0.3 °C and 9.1 °C, respectively. The 95th percentile of the land surface temperature increase after the AI data centers began operations is between 1.5°C and 2.4°C.
In addition, the researchers found that the impact of land surface temperature increase impacted a very large area and reached up to 10 km distance from the AI hyperscaler facilities. The data heat island effect seems to reduce its intensity to 30% within 7 km around the data centers. It was pointed out in 2023 by Kilgore et al that "Data centers account for 2.5% to 3.7% of global GHG emissions," which exceeds the greenhouse gas emissions from the aviation industry recorded at 2.4%.
 |
| from Marinoni, Andrea et al |
The increasing demand for AI-based services, processes and
operations is leading to the proliferation of data centers worldwide that are
extremely power hungry. This study shows a rather remarkable impact of the AI
data centers on their local regions, which was found to be consistent across data
centers worldwide and extends for several kilometers around the AI
hyperscalers. The consistency, scale and extent of these effects lead the
researchers to suggest that data centers are creating local climate zones - that they call the data
heat island effect - is real, significant, and may have a non-trivial impact on
global warming and climate transformation.
The data heat island effect could have a significant impact
on the on planet since the trends of
data center energy consumption are expected to show a steep growth in the
foreseeable future. The data heat island effect could become an additional
factor in the changing climate, hence having a robust impact on communities at
local, regional, and international level.
No comments:
Post a Comment