Global electricity consumption from data centers is projected to more than double, according to an International Energy Agency (IEA) report, with artificial intelligence (AI) as the major driver.
The Energy and AI report projects that data center consumption globally will grow to 945TWh per year by 2030, from 415TWh in 2024.
The US will account for the largest increase, followed closely by China. Together, the countries represent nearly 80 percent of projected global growth.
In the US, consumption is expected to increase by up to 240TWh, up 130 percent compared to 2024 levels, and China is expected to grow by 175TWh, up 170 percent compared to 2024. In Europe, a 45TWh growth is projected, up 70 percent, and Japan is expected to grow by around 15TWh, a more than 80 percent jump.
Meeting demand will require a diverse range of energy assets. The report projects that renewable energy generation will grow by more than 450TWh to meet data center demand by 2035. Currently, renewables – primarily wind, solar PV, and hydro - account for 27 percent of the electricity consumed by data centers. By 2030, it is expected to make up a 50 percent share, primarily due to the growth of wind and solar across the world.
The role of dispatchable fuel sources, especially natural gas, will also play a critical role. Natural gas currently accounts for 26 percent of the electricity consumed by data centers, with nuclear energy holding a 15 percent share.
The IEA projects that natural gas power will grow by 175TWh to meet data center demand. Much of this growth will be in the US, where the pro-gas policies of the current administration have already led to several major announcements of new gas generation facilities to serve data centers behind the meter.
Reflecting on the report, head of business development for data centers at Danfoss, a power solutions firm, Jakob Jul Jensen, said: “The IEA’s new special report on Energy & AI highlights that a diverse range of energy sources will be required to meet the growing electricity needs of data centers. However, energy efficiency must also be prioritised and integrated into every layer of data center design and operation.”
The IEA sees data center demand making up around 10 percent of global electricity demand growth by 2030, which is less than industrial motors, air conditioning in homes and offices, or electric vehicles. The increase varies depending on how advanced the economy of a nation is, with data center growth in developing countries accounting for around five percent and up to 20 percent in advanced economies.
The IEA warns that unless significant investments are made into transmission infrastructure, up to 20 percent of planned data center projects could be at risk of delays. In addition, the report warns that generation equipment is likely to face supply chain issues, with manufacturers struggling to meet the demand necessary to match the growth of the sector.
To reduce strain on the grid, the report contends that data centers, especially AI-focused ones, could use spare server capacity or back-up generation more flexibly. In addition, the IEA posits that grid operators could also incentivize data center operators to site their operations in areas with fewer grid constraints.
Currently, more than 50 percent of data centers in the US are in pre-existing large clusters. The move away from the cluster model could be aided by the growth of off-grid solutions for data centers in areas with energy abundance, such as Texas, where there is an excess of available energy.
The role of AI in improving energy efficiency is also explored in the report. The IEA contends that AI, if deployed correctly, could optimize energy and mineral supply, electricity generation and transmission, and energy consumption. However, the report also argues that the energy sector has so far failed to reap the benefits of AI, due to inadequate access to data and digital infrastructure and skills, as well as persistent digital and physical security concerns, which often trump potential efficiency gains.
The report also considers the impacts the growth of AI data centers will have on global supply chains. Materials such as gallium, which is a crucial metal used in computing chips and power electronics, are expected to skyrocket in demand, reaching more than 10 percent of the current supply by 2030.
With China accounting for 99 percent of global refined gallium supply, there is an increasing risk that the supply of the metal could become even more constrained, leading to delays for data center developers worldwide.
Finally, the IEA reflects on concerns that AI could accelerate climate change due to its significant power requirements. In the report, emissions from electricity use by data centres grows from 180 million tons (Mt） today to 300Mt in the Base Case by 2035, and up to 500Mt in the Lift-Off Case.
However, the report also notes that a widespread adoption of existing AI applications could lead to large emissions reductions that outweigh the overall emissions from the data center market. The IEA estimates that such a broad application could be equivalent to five percent of energy-related emissions in 2035
