The International Energy Agency (IEA) examined the opportunities and challenges brought by AI in terms of global energy.
Training and deployment of sophisticated AI models occurs within vast, power-hungry data centers. “Typical AI-focused data centers consume as much electricity as 100,000 households,” the IEA notes, with the largest facility under construction projected to demand 20 times more.
Data Center Investment is Expanding
Global investment in data centers has almost doubled since 2022, reaching $50 trillion in 2024, sparking concerns about growing electricity needs.
Data centers account for around 1.5% (about 415 terawatt hours, TWH) of global electricity consumption in 2024, but local impacts are far more important. Consumption has increased by about 12% each year since 2017, significantly outpacing the overall increase in electricity demand.
The US leads this consumption (45%), followed by China (25%) and Europe (15%). Almost half of US data center capacity is concentrated in just five regional clusters.
Going forward, IEA will more than double the global data center’s power consumption by 2030, reaching approximately 945 TWH. To put it in context, it is slightly higher than Japan’s current total electricity consumption.
AI has been identified as “the most important driver of this growth.” The US is projected to see the biggest increase in data centers that could account for almost half of all electricity demand growth by 2030. By the end of the decade, US data centers are expected to consume more electricity than they would be used by the aluminum, steel, cement, chemicals and other energy-intensive manufacturing industries.
The IEA “base case” extends this trajectory, predicting approximately 1,200 Twh of global data center power consumption by 2035. However, there is a 2035 projection of 700 TWH (the “headwind” case” to 1,700 TWH (the “lift-off case”), and there is significant uncertainty according to AI Uptake, Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and Efficiency Gain, and E
Fatih Birol, Executive Director of IEA, said: “AI is one of the biggest stories in the energy world today, but up until now, policymakers and markets have lacked the tools to fully understand the widespread impact.
“In the US, data centers account for almost half of the increase in electricity demand, making it more than half in Japan and a fifth in Malaysia.”
Meet the global AI energy needs
Powering this AI boom requires a diverse energy portfolio. The IEA suggests that renewable energy and natural gas will lead, but emerging technologies such as small modular nuclear reactors (SMRs) and advanced geothermal have a role to play.
Renewable energy supported by storage and grid infrastructure is projected to meet half of global data center demand growth by 2035. Natural gas is also important, especially in the US, with the basic case increasing by 175 TWH by 2035 to meet data center needs. Nuclear power has contributed similarly, particularly in China, Japan and the US, with the first SMR expected around 2030.
However, simply increasing production is not sufficient. The IEA highlights the important needs of infrastructure upgrades, particularly grid investments. The existing grid is already tense, and the long lead times of complex connection queues and key components such as transformers could delay around 20% of globally planned data center projects.
The possibilities of AI to optimize energy systems
Beyond energy demand, AI offers great potential to revolutionize the energy sector itself.
The IEA details many applications:
Energy Supply: The oil and gas industry (early adopters) use AI to optimize exploration, production, maintenance and safety, including reducing methane emissions. AI can also support important mineral exploration. Power Sector: AI can improve forecasts for fluctuating renewable energy and reduce reductions. It enhances grid balance, fault detection (30-50% reduction in outage period) and unlocks critical transmission capacity through smarter management. Use of Termination: In the industry, the widespread adoption of AI for process optimization could result in energy savings equivalent to the total energy consumption of Mexico today. Transport applications such as traffic management and route optimization must monitor the rebound effect from autonomous vehicles, but can save 120 million units of energy. The possibilities for optimizing construction are important, but are hampered by slower digitization. Innovation: AI can dramatically accelerate the discovery and testing of new energy technologies such as advanced battery chemistry, synthetic fuel catalysts, and carbon capture materials. However, the energy sector is currently not fully utilizing AI for innovation compared to fields like biomedicine.
Collaboration is the key to navigating issues
Despite the possibilities, significant barriers are preventing AI from fully integrated into the energy sector. These include data access and quality issues, inadequate digital infrastructure and skills (AI talent concentration is low in the energy sector), and regulatory hurdles, and security concerns.
Cybersecurity is a double-edged sword. AI enhances defense capabilities, but it also equips attackers with sophisticated tools. Utility cyberattacks have tripled over the past four years.
Supply chain security is another important concern, particularly with important minerals such as gallium (used in advanced chips), where supply is highly concentrated.
The IEA concludes that deeper dialogue and collaboration between the technology sector, the energy industry, and policy makers are paramount. Addressing the challenges of grid integration requires smarter data center positioning, operational flexibility research, and streamlining.
AI presents opportunities for significant emission reductions through optimization, exceeding the emissions generated by data centers, but these benefits are not guaranteed and can be offset by rebound effects.
“AI is a tool and potentially very powerful, but it depends on us, whether it’s our society, our government, our businesses, etc,” Dr. Birol said.
“The IEA will continue to provide data, analytics and forums for dialogue to help policymakers and other stakeholders navigate the path ahead as the energy sector shapes the future of AI and AI shapes the future of energy.”
(Photo: Javier Miranda)
See: UK forms AI Energy Council to coordinate growth and sustainability goals
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London. The comprehensive event will be held in collaboration with other major events, including the Intelligent Automation Conference, Blockx, Digital Transformation Week, and Cyber Security & Cloud Expo.
Check out other upcoming Enterprise Technology events and webinars with TechForge here.
