Rising heat and humidity are eroding data centre free cooling, study finds

Climate change is quietly raising the cost of data centre cooling. A study published in “Scientific Reports” and led by University of Hawaii at Manoa atmospheric scientists, reported on June 26, 2026, found that rising heat and humidity are reducing how often operators can use “direct air free cooling”, one of the cheapest and least energy-hungry ways to keep servers cool. The squeeze arrives just as the artificial intelligence boom drives record data centre construction.

Direct air free cooling pulls naturally cold outside air into a building to cool equipment, avoiding power-hungry mechanical refrigeration. The study found the windows when outdoor conditions allow it are narrowing in many regions, a trend that pushes operators toward more energy or more water to keep machines within safe limits.

What the data centre cooling study found

The team, led by Christina Karamperidou, an atmospheric sciences professor in the UH Manoa School of Ocean and Earth Science and Technology, combined high-resolution hourly weather observations, climate-model simulations, and a global database of data centre locations. That let them count how often conditions exceeded the recommended thresholds for direct air free cooling, both in the past and under future climate scenarios, and to spot large-scale patterns that single-site studies miss.

Source: University of Hawaii at Manoa and “Scientific Reports”, reported June 26, 2026.

Over the past 45 years, the number of hours with high temperature and humidity that limit direct air cooling has increased, the researchers found, with the sharpest rises across the tropics and the southeastern United States. The share of data centres exposed to limiting conditions for at least one quarter of the year is rising, and projections through mid-century point to those constraints expanding as warming and humidity continue.

One finding has direct operational weight: the largest changes are not in average conditions but in the worst days. “From an operational perspective, those worst-day conditions often drive contingency planning, system overrides, redundancy requirements, and reliability decisions,” Karamperidou stated. Planning, she noted, may need to account not just for average conditions but for how the most stressful days are shifting.

Why free cooling matters as AI drives demand

The timing is the problem. Demand for data centres is climbing as AI workloads scale, exactly when the cheapest cooling method is becoming less reliable. By the IEA’s “Energy and AI” analysis, data centres used about 415 terawatt-hours of electricity in 2024, roughly 1.5 per cent of global supply, and that figure is projected to roughly double to about 945 terawatt-hours by 2030, with AI the main driver.

When free cooling is unavailable, operators fall back on mechanical refrigeration, which adds electricity load, or on evaporative cooling, which consumes water. The study sits at the meeting point of climate, computing, energy and water, a tension Winss Solutions has examined in AI and sustainability, a match made in heaven?.

The energy and water trade-off

The mechanism is straightforward. As hot, humid hours rise, the free-cooling window shrinks, and each lost hour shifts a facility toward either more power or more water. The diagram below traces that path.

Karamperidou framed the study as a map of where the strain will land. “We hope this work helps identify where emerging technologies and new cooling approaches may help address trade-offs among reliability, energy consumption, and water use,” she stated. The work was developed with researchers on cyberinfrastructure and next-generation cooling through the NSF ERC EARTH, where UH is a key partner. The broader resource pressure, including the water side, echoes Winss Solutions coverage of the important role of big data in sustainability.

Background: cooling, climate and the AI build-out

Cooling is one of the largest non-computing energy draws in a data centre, which is why operators favour free cooling wherever the climate allows it. As AI accelerates server demand, the industry has leaned on cooler and drier sites, waste-heat reuse, and liquid cooling to hold down power use. The UH Manoa study adds a climate dimension to those plans: the geography of where free cooling works is shifting, and the worst days, the ones that set reliability margins, are changing fastest. For an industry racing to build capacity, the finding reframes cooling as a long-term siting and resilience question rather than a fixed engineering assumption, a theme within the wider shift Winss Solutions follows in the future of green energy: trends and predictions.

Sources: Scientific Reports; Phys.org; International Energy Agency

Featured image: photo by panumas nikhomkhai on Pexels (free Pexels license).

For more on this topic, see our explainer on what a prompt actually costs in energy and water.

Become a Sponsor

Our website is the heart of the mission of WINSS – it’s where we share updates, publish research, highlight community impact, and connect with supporters around the world. To keep this essential platform running, updated, and accessible, we rely on the generosity of you, who believe in our work.

We offer the option to sponsor monthly, or just once choosing the amount of your choice. If you run a company, please contact us via info@winssolutions.org.

• donate monthly
• donate once only

Select a Donation Option (USD)

$5.00 (p/m) $10.00 (p/m) $20.00 (p/m) Other (p/m)

Enter Donation Amount (USD)

Figen Sekin

I specialize in sustainability education, curriculum co-creation, and early-stage project strategy. At WINSS, I craft articles on sustainability, transformative AI, and related topics. When I’m not writing, you’ll find me chasing the perfect sushi roll, exploring cities around the globe, or unwinding with my dog Puffy — the world’s most loyal sidekick.