Three key takeaways from IEA’s first Energy and AI report

AI-driven data centers consume huge amounts of electricity, which could change industry development.(Photo: iStock)

International Energy Agency (IEA) released its first Energy and AI report on April 10, highlighting how the data center boom driven by artificial intelligence is giving the tech sector unprecedented influence over the energy industry.

IEA Executive Director Fatih Birol stated that AI represents one of the most important developments in energy today, yet policymakers and markets lack the tools to fully understand the wide-ranging impacts. RECCESSARY has summarized three key takeaways from the report.

1. AI power demand surges past traditional heavy industries

The IEA notes that AI-focused data centers today consume as much electricity as 100,000 households, with the largest under construction projected to use 20 times that amount—underscoring the explosive growth in power demand driven by AI adoption.

The report estimates that global data center electricity consumption will double within five years, reaching 945 terawatt-hours (TWh)—roughly equivalent to Japan’s current annual power consumption.

In the United States, data centers are expected to consume more electricity than the combined usage of all energy-intensive manufacturing sectors, including aluminum, steel, cement, and chemicals. The tech industry is already a major electricity and emissions contributor; with AI workloads increasing, the pressure to secure renewable energy and accelerate decarbonization is mounting.

Comparison of data center electricity consumption. (Source: IEA)

Yet ramping up electricity supply isn’t easy—especially as many grids are already stretched thin. In developed countries, building new transmission infrastructure can take four to eight years. Lead times for key components like transformers and cables have doubled over the past 5 years. Without addressing these grid constraints, up to 20% of planned data center projects could face delays.

2. Data center clustering intensifies electricity supply challenges

The IEA reports that U.S. data centers accounted for 45% of global electricity consumption in 2024, followed by China and Europe. ASEAN economies are emerging as critical players, with data center power demand expected to double—especially in hubs like Singapore and Johor province in Malaysia.

Global Map of Major Data Centers (2024). (Source: IEA)

To meet low-latency requirements for services like video streaming and web hosting, data centers are often built near end users in densely populated areas. This, combined with uneven power supply across countries, creates sharp regional disparities. In Japan, over half of the power demand growth is driven by data centers; in the U.S., the share is nearly 50%; and in Malaysia, around 20%.

The report stresses that the outlook for AI-related electricity demand remains highly uncertain. Key factors include the speed of AI adoption, efficiency improvements, and technological breakthroughs in the energy sector. Under different scenarios, electricity demand could vary by as much as 1,000 TWh by 2035. Demand for gas-fired power could differ by a factor of four, while nuclear power projections vary even more widely.

3. AI powers energy shift toward renewables and gas

While AI’s power demands are straining infrastructure, it also offers significant potential to transform the energy sector. The IEA sees major opportunities for AI to enhance resource supply, optimize electricity transmission, extend asset lifetimes, and reduce carbon emissions. But its greatest promise may lie in accelerating innovation—helping identify advanced materials, battery chemistries, and carbon capture molecules more quickly.

By 2035, the IEA projects that half of the new data center electricity demand will be met by renewables, supported by energy storage and enhanced grid capacity. Including wind and solar power, the total electricity generation will exceed 450 terawatt-hours (TWh). Natural gas will also play a key role due to its reliability and market competitiveness, particularly in the short term. Coal and gas remain dominant in meeting near-term needs.

Breakdown of energy sources for data centers. (Source: IEA)

Earlier this year, Chinese startup DeepSeek claimed to challenge the notion that AI is inherently energy-hungry. Microsoft CEO Satya Nadella remarked that as AI becomes more computationally efficient, its applications—and therefore energy demand—will continue to grow. The IEA now underscores this trend, positioning AI as a central force in shaping the future of the global energy system.

Source: IEA, France 24, Bloomberg