Al Jazeera reported on 11 June that the study used NASA satellite data from 2004 to 2024 and cross-referenced those data with more than 11,000 data-centre locations. The report said the authors focused on 6,733 centres outside densely populated areas and found temperature effects up to 10 kilometres away.
The arXiv paper, "The data heat island effect: quantifying the impact of AI data centers in a warming world", says the average land-surface-temperature increase after operations began was about 2C. It also estimates that more than 340 million people could be affected and reports a range from 0.3C to 9.1C across sites.
Those are land-surface temperatures, not near-ground air temperatures. Land surface temperature is derived from satellite observations of how hot the ground surface appears, while people experience air temperature, humidity, shade, ventilation and indoor cooling. The distinction matters because a warmer roof, road or bare-soil surface does not automatically prove the same rise in human heat exposure.
Table: What the preprint reports and what remains limited
| Finding or method | Reported value | Publication status | Main limitation |
|---|---|---|---|
| Average post-opening land-surface-temperature rise | 2C | arXiv preprint | Correlation design; not peer reviewed |
| Reported range across sites | 0.3C to 9.1C | arXiv preprint, reported by Al Jazeera | Site-level outliers need scrutiny |
| Maximum reported distance of effect | Up to 10km | arXiv preprint, reported by Al Jazeera | Mechanism and confounders remain disputed |
| Population potentially affected | More than 340 million people | arXiv preprint | Exposure estimate depends on the model and siting assumptions |