Not because the authors built some shiny new model. They did something ruder and more useful - they pointed at an awkward feedback loop nobody in tech likes to linger on. Europe is heating up, fires are getting meaner, hospitals are already feeling it, and the AI boom may be quietly asking fossil fuels for one more drink before closing time.[1]

There is a harsh elegance to the argument. Heat harms bodies. Fire harms lungs. Fossil fuels make both worse. Then AI arrives like the overachieving intern who somehow also triples the office power bill.

This is not an AI paper in the usual nerdy sense

Rajagopalan, Munzel, and Gasparrini are not comparing transformers, arguing about benchmarks, or inventing a model with a name that sounds like a startup and a kitchen appliance at once. Their 2026 European Heart Journal piece is a climate-and-health warning. The core claim is simple: Europe is already in a heat and fire crisis, and the exploding electricity demand from AI data centers risks locking in more fossil-fuel use right when emissions need to fall fast.[1]

That matters because heat is not just "wow, summer is rude now." Heat raises the risk of heart attacks, heart failure decompensation, dehydration, kidney injury, and heat stroke. Add wildfire smoke and you get a nasty combo platter for the cardiovascular and respiratory systems.[1][5] Europe, meanwhile, is warming faster than any other continent in the world according to the European State of the Climate reporting.[6]

The body keeps the score. The grid sends the invoice.

The paper lands its punch with real numbers. A major 2023 study in Nature Medicine estimated about 61,672 heat-related deaths in Europe during the summer of 2022.[2] A follow-up estimated 47,690 heat-related deaths in 2023, even with adaptation measures doing some damage control.[3] Another 2025 Nature Medicine analysis projected that, without stronger mitigation and adaptation, temperature-related mortality across European cities could rise substantially over this century.[4]

That is the health side of the ledger.

The AI side is what makes this editorial sting. The International Energy Agency estimated in 2025 that data centers consumed about 415 TWh of electricity in 2024 and could reach roughly 945 TWh by 2030 in its base case - with AI as a major driver.[7] That is not a cute little rounding error. That is "your autocomplete has industrial consequences" territory.

And here is the ugly bit: AI workloads do not politely line up with when wind is blowing or the sun is doing its one job. Training and inference can create huge, spiky demand. Storage and cooling are relentless. If grids cannot meet that with clean power plus long-duration storage, gas plants and coal plants get called back onstage like a reunion tour nobody asked for.[1][7]

Why the paper feels uncomfortably plausible

This is where the authors stop sounding like alarmists and start sounding annoyingly right.

Recent work backs the surrounding context. A 2024 Lancet study found a substantial global mortality burden from air pollution caused by landscape fires.[5] A 2025 Nature paper went further and linked a meaningful share of observed heatwaves to emissions from specific carbon majors, which is the kind of sentence that should make fossil-fuel PR teams spill their cold brew.[8]

Meanwhile, current reporting from the IEA says data-centre electricity use surged again in 2025, with AI-focused facilities growing even faster than the rest.[9] The broad pattern is hard to miss: more compute, more power demand, more pressure to keep dispatchable fossil generation around unless policy gets much stricter and infrastructure gets much better.

There is a kind of wabi-sabi honesty in this paper. It does not pretend every problem has a sleek software fix. Sometimes the missing feature is restraint. Sometimes the smartest machine in the room still runs on a very dumb fuel.

The quiet thesis underneath all this

The paper is really about alignment, just not the Silicon Valley kind. It asks whether energy policy, health policy, and AI policy are aligned with physical reality.

If governments let data-center expansion outrun clean generation, then "net zero" claims start to look decorative. If hospitals plan for heat as if it were a rare surprise instead of a recurring stress test, people die in very old-fashioned biological ways. And if tech companies keep treating renewable energy credits like moral dry shampoo, the atmosphere will remain unimpressed.[1]

The authors argue for a cleaner, stricter approach: link large AI and data-center growth to verifiable clean power, build real heat-health protections, and count healthcare costs when making energy decisions.[1] Sensible. Unspectacular. Necessary. Not every connection needs to be ornate for the whole design to hold.

AI can help science, medicine, and energy systems. No question. But if your miracle machine needs a hidden coal encore, that is not magic. That is outsourcing the smoke.

References

1. Rajagopalan S, Munzel T, Gasparrini A. Europe's heat and fire crisis and the new fossil-fuel surge from artificial intelligence. European Heart Journal. 2026. DOI: https://doi.org/10.1093/eurheartj/ehag078
2. Ballester J, Quijal-Zamorano M, Mendez Turrubiates RF, et al. Heat-related mortality in Europe during the summer of 2022. Nature Medicine. 2023;29:1857-1866. DOI: https://doi.org/10.1038/s41591-023-02419-z
3. Ballester J, Quijal-Zamorano M, Mendez Turrubiates RF, et al. Heat-related mortality in Europe during 2023 and the role of adaptation in protecting health. Nature Medicine. 2024;30:3101-3105. DOI: https://doi.org/10.1038/s41591-024-03186-1
4. Masselot P, Mistry MN, Rao S, et al. Estimating future heat-related and cold-related mortality under climate change, demographic and adaptation scenarios in 854 European cities. Nature Medicine. 2025. DOI: https://doi.org/10.1038/s41591-024-03452-2
5. Xu R, Ye T, Huang W, et al. Global, regional, and national mortality burden attributable to air pollution from landscape fires: a health impact assessment study. The Lancet. 2024;404:2447-2459. DOI: https://doi.org/10.1016/S0140-6736(24)02251-7
6. European Commission. 2024 warmest year on record in Europe, finds European State of the Climate report. 15 April 2025. https://commission.europa.eu/news-and-media/news/2024-warmest-year-record-europe-finds-european-state-climate-report-2025-04-15_en
7. International Energy Agency. Energy and AI. 10 April 2025. https://www.iea.org/reports/energy-and-ai
8. Ekwurzel B, Quesada M, Nauels A, et al. Systematic attribution of heatwaves to the emissions of carbon majors. Nature. 2025. DOI: https://doi.org/10.1038/s41586-025-09450-9
9. International Energy Agency. Data centre electricity use surged in 2025, even with tightening bottlenecks driving a scramble for solutions. 23 April 2026. https://www.iea.org/news/data-centre-electricity-use-surged-in-2025-even-with-tightening-bottlenecks-driving-a-scramble-for-solutions

Disclaimer: This blog post is a simplified summary of published research for educational purposes. The accompanying illustration is artistic and does not depict actual model architectures, data, or experimental results. Always refer to the original paper for technical details.