Rising global temperatures pose an increasing risk to the resilience of well over half of the world’s top 100 data centre hubs, according to new research from risk intelligence company Verisk Maplecroft.
The risks span several key areas, including pressure on water and electricity resources from rising cooling requirements. These factors expose data centres to increased costs, potential conflict with local communities, and power outages where energy infrastructure is less resilient to climate hazards.
Data centres are built with resilience measures, such as backup generators, in mind, and water and energy efficiencies are improving. However, the research suggests that the challenges of climate change could test this increasingly essential infrastructure as temperatures rise and demand for AI, data storage, and cloud computing grows.
Heatwaves already present a major risk to data centres. During extreme temperature spikes, operators can be forced to shut down servers to prevent damage from overheating, leading to service disruptions. In 2022, summer heatwaves in the UK and US caused several data centres to go offline, disrupting businesses reliant on cloud-based servers.
Verisk Maplecroft’s Cooling Degree Days Index measures how often and by how much temperatures exceed thresholds that trigger cooling requirements for buildings. Currently, 56% of the top 100 data centre hubs are rated as ‘high’ or ‘very high’ risk for cooling degree days. In a high-emissions scenario (SSP585), this increases to 68% by 2040 and 80% by 2080.
This means that around three-quarters of the world’s data centre hubs will need to meet significant and increasing cooling demands for longer periods each year, resulting in higher water and energy usage and costs. The top 100 data centre hubs are expected to see an average 83% increase in cooling degree days from 2030 to 2080.
Over half of top data centre hubs are in highly water-stressed areas
For many data centres, the key ingredient for cooling is water, although indirect usage from electricity generation is also significant. The average mid-size data centre uses about 300,000 gallons (1.4 million litres) of water a day, and these requirements will rise with temperatures.
By 2030, under the SSP585 scenario, 52% of global data centre hubs will be classified as facing ‘high’ or ‘very high’ risk in Verisk Maplecroft’s Water Stress Index, which evaluates total water use relative to total annual available flow. By 2050, this rises to 58%. Given the growing demand for data centres and the associated increase in water use, water scarcity is likely to become both a resilience and a sustainability risk in supply chains for IT and cloud-based services.
These risks are not limited to availability. They could spill over into social and political domains. Conflict over access to water with local communities can cause unrest and become a political issue, threatening social licence to operate and the reputations of organisations using affected sites for digital services.
Unsurprisingly, the Middle East faces the highest exposure to water stress risks, with Abu Dhabi, Dubai, and Istanbul all rated ‘very high’ risk by 2030. In Africa, Lagos, Johannesburg, and Nairobi are projected to be in the same category by 2050. In North America, half of the top data centre hubs are expected to score ‘high’ or ‘very high’ risk for water stress within just five years. Los Angeles, San Diego, Denver, Phoenix, and Mexico City face some of the highest risks.
Data centre capacity in Asia is set to nearly double by the start of the next decade, further exacerbating water availability concerns. Between 2030 and 2050, water stress risk is projected to increase significantly in Manila and Bangkok and remain ‘very high’ in several Indian hubs, including Chennai and New Delhi.
Intensifying climate hazards are a key risk to energy reliability
Data centres’ energy use is already straining power grids globally. They currently account for around 1.5% of global electricity demand, but this is expected to rise to 3% by 2030. Cooling accounts for up to 40% of that demand, a proportion that will increase as temperatures rise.
Verisk Maplecroft’s Energy Infrastructure Index assesses countries on their energy generation capacity, diversity, distribution reliability, and innovation. The importance of energy reliability in investment decisions is reflected in this index, which currently categorises 70% of top data centre hubs as ‘low’ risk.
However, extreme heat reduces the efficiency of power transmission, creating further complications for energy infrastructure amid rising demand. Data centres are highly sensitive to supply variations, so consistent power is essential to reduce outage risk. While short-term backups exist, switching to these systems is not entirely reliable and can cause grid instability due to sudden demand shifts.
Ageing power infrastructure will become increasingly vulnerable as the climate changes, even in markets currently assessed as ‘low’ risk.
According to Verisk Maplecroft’s Climate Hazard Index, which measures eight acute and chronic climate hazards, 27% of the top 100 data hubs – mostly in Asia and the Middle East – are expected to be at ‘high’ or ‘very high’ risk by 2050 under a high-emissions scenario. All but one will face at least ‘medium’ risk. This means many data centre hubs will experience more hazardous and extreme conditions from storms and rising temperatures, threatening energy reliability and operational resilience.
Combatting resilience and sustainability challenges
Data centres are long-term investments that will be exposed to shifting risk landscapes throughout their lifespan. Anticipating where these risks may emerge is a strategic priority for global organizations that operate or depend on these services.
“Data centre operators are innovating solutions to increase resilience and offset sustainability concerns, but rising temperatures increase these challenges,” says Capucine May, Senior Analyst at Verisk Maplecroft. “The onus is on operators, customers, and investors alike to assess rising climate threats alongside social and political risk factors – not only for their own resilience, but because of a growing regulatory focus on third-party risk management.”
