The AI boom meets climate risk: why data-center cooling and grid resilience matter for builders

AI data centers face a growing physical threat from heatwaves and severe weather, which drive up cooling costs, strain power grids, and shift where new capacity gets built. For AI builders scaling inference and training workloads, understanding these climate risks is becoming essential for cost planning, site selection, and infrastructure resilience.

What happened

European heatwaves this week have highlighted how extreme weather stresses AI data-center infrastructure. A study by climate risk analytics firm First Street found that 79% of global data center capacity faces elevated risks from acute climate hazards such as flooding, extreme winds, and wildfires. Zurich Insurance reports that severe weather has become the leading cause of loss in its U.S. data center builders' risk portfolio over the past three years, now driving a third of the company's losses. Meanwhile, 64% of data center capacity under construction is outside traditional hubs like Northern Virginia, moving into frontier markets such as West Texas, Tennessee, Wisconsin, and Ohio, as well as the Iberian Peninsula in Europe. These areas can face heightened risks from tornadoes, hail, and high winds.

Why AI builders should care

Cooling accounts for roughly 40% of data center energy use even at normal temperatures, and this rises during heatwaves when air conditioning also drives up grid demand. Mishal Thadani, CEO of Rhizome, notes that data centers need the most energy exactly when the grid has the least available to give. This creates a compounding risk for any AI platform relying on consistent compute availability. Microsoft designs its data centers to operate reliably across a wide range of environmental conditions with redundant systems and real-time monitoring. Nvidia announced that its new AI servers can run cooling liquid at 45 degrees Celsius, up from lower temperatures, and that raising chiller temperatures by one degree can cut cooling energy costs by about 4%. These technology shifts signal that the industry is moving toward climate-adaptive infrastructure, but builders should expect these adaptations to affect deployment timelines and costs.

Practical implications

For AI builders, the practical takeaway is that site selection and cooling strategy are no longer just operational details. They directly affect uptime, energy costs, and capital availability. Insurers and investors are increasingly tying climate risk to data-center losses. Joe Macejak of Marsh Risk warns that failing to manage these risks could lead to higher costs and operational shortfalls that pose a threat to the capital stacks fueling AI-driven data center expansion. Operators may need to integrate climate-change factors into design specifications, as one European client recently did. A diverse set of cooling systems and smarter heat-transfer strategies will likely become standard.

Caveats

The evidence in this article relies on CNBC reporting and industry commentary. Specific figures such as the 79% exposure and 64% frontier construction come from cited studies and may evolve with new data. Claims about future deployments and cost savings depend on market conditions and technology adoption pace. Builders should treat these as directional signals rather than precise forecasts.

FAQs

How do heatwaves and severe weather affect AI data centers?

Heatwaves increase cooling demand, which raises energy use and strains power grids. Severe weather such as flooding, extreme winds, and wildfires can disrupt operations and drive up insurance and repair costs. A First Street study found that 79% of global data center capacity faces elevated risks from acute climate hazards.

Which regions are becoming new hubs for data-center construction outside traditional markets?

This year, 64% of data center capacity under construction is outside traditional hubs like Northern Virginia. Frontier markets include West Texas, Tennessee, Wisconsin, and Ohio in the U.S., as well as the Iberian Peninsula in Europe. These areas can face heightened risks from tornadoes, hail, and high winds.

What cooling innovations are being used to reduce data-center energy use and costs?

Nvidia announced that its new AI servers can run cooling liquid at 45 degrees Celsius, up from lower temperatures, and that raising chiller temperatures by one degree can cut cooling energy costs by about 4%. Microsoft designs its data centers to operate reliably across a wide range of environmental conditions with redundant systems and real-time monitoring.

What role do insurers play in reporting climate-related data-center losses?

Zurich Insurance reports that severe weather has become the leading cause of loss in its U.S. data center builders' risk portfolio over the past three years, now driving a third of the company's losses. Insurers are tracking climate risk exposure as part of portfolio management, and failing to manage these risks could lead to higher costs and operational shortfalls that threaten the capital stacks fueling AI-driven data center expansion.