A new study on the location of data centers around the world has revealed a surprising fact in the technology world. According to ASHRAE guidance, which sets industry standards, the ideal air intake temperature for data centers should be between 18°C and 27°C. However, a global mapping study shows that approximately 7,000 of the 8,808 operational facilities worldwide are operating outside this efficient temperature range.
Data center overheating problem: 7,000 facilities operating inefficiently
Climate mismatch has become a global norm in the industry. Approximately 600 facilities are located in extremely hot regions where average annual temperatures exceed 27°C, while many others operate in cooler regions below 18°C. For example, Singapore, where temperatures hover around 33°C and humidity frequently exceeds 80%, is home to over 1.4 gigawatts of data center capacity despite the intense heat conditions.
High ambient temperatures significantly increase cooling demands at facilities, reducing electricity transmission efficiency and placing additional pressure on local power grids. According to international energy forecasts, data centers consumed approximately 415 terawatt-hours of electricity in 2024, accounting for about 1.5% of global demand. With the proliferation of high-density systems, this figure is expected to more than double by 2030.
Site selection for servers is generally based on factors such as power availability, electricity pricing, water access, and land costs. These commercial considerations often outweigh temperature suitability when operators evaluate new projects. Globally, air cooling remains the dominant approach, accounting for more than half of all systems. While liquid-based cooling is gaining traction, particularly for high-power systems, upgrading existing facilities requires substantial capital investment.
Risk assessments indicate that extreme heat could affect roughly two-thirds of major data center hubs worldwide by 2040. Countries like Singapore, Nigeria, and the United Arab Emirates already have entire facilities located in regions exceeding critical temperature thresholds. On the other hand, the rapid expansion of AI workloads is leading to unprecedented growth in data centers, sharply increasing power demands and operating costs. This increase in energy use demonstrates that economic concerns about the AI bubble are translating into a tangible infrastructure strain.
These data reveal that current expansion patterns prioritize short-term demand over long-term environmental efficiency. So, what are your thoughts on the balance between technological sustainability and increasing energy needs?
