The central question in energy and climate discussions is the extent to which AI will increase power consumption in the coming years. Many analysts predict a surge in AI-driven electricity use within five years due to its expanding applications for consumers and businesses. However, some analysts advise caution, suggesting that not all planned data centers will be realized, and AI could become more energy-efficient.
While this debate is critical, the complete picture is more complex. Even if AI’s energy demands are less than currently anticipated, various other factors will drive electricity demand. A recent report from the International Energy Agency (IEA) indicates that global electricity demand rose by 4.3% last year due to factors beyond data centers, including the electrification of heating and cooking, increased air conditioning use due to record temperatures, and more. According to Fatih Birol, the IEA’s executive director, electricity growth is a very clear trend.
Broadening the discussion around future power demand helps prevent over-fixation on a single technology, no matter how transformative. It also helps avoid overlooking the complex interaction of factors that will shape our energy future, along with the related challenges and opportunities.
To understand the nature of electricity demand growth, it’s useful to analyze it geographically, starting with emerging markets and developing countries. Electricity demand in China increased by 7%, and in other emerging and developing economies by over 4%, while the European Union saw a growth of about 1.5%, according to the IEA.
Some of this growth results simply from expanding economies. Greater wealth leads to increased electricity consumption. However, a significant portion of the rising demand stemmed from high temperatures, particularly heat waves in India and China, which led consumers and businesses to increase air conditioning usage.
U.S. electricity usage also grew by about 2% last year, according to government . The primary driver was new data centers, but other factors also contributed. Federal programs enacted during the Biden Administration, such as the , have supported the growth of advanced manufacturing in the country. This translates to new facilities requiring electricity on previously undeveloped land or in renovated vacant buildings. Furthermore, modern manufacturing plants rely more on electricity than in the past, when they might have used on-site fossil fuel power.
Electric vehicles are another factor. Despite perceptions of challenges in the U.S. EV market, with sales not meeting expectations, the IEA reports that sales continued to grow at a rate exceeding 10%. Increased EV adoption will inevitably lead to greater power demand in the coming years.
Several lessons can be drawn from this information. Relying solely on AI efficiency improvements to avert an electricity crisis overlooks the broader picture. While developments like more energy-efficient language models could reduce AI’s future electricity demand, they won’t address all other demand sources.
There are also positive aspects to note. The IEA reports that 80% of new electricity generation globally last year came from renewable energy or nuclear power. This means that emissions grew at a slower pace than the economy. Birol notes that there is a continuous decoupling of economic growth from emissions growth, offering a silver lining.
While this decoupling is encouraging, emissions still need to decline to avoid the worst impacts of climate change. This imminent power crunch could force us to address the full range of available solutions. Power companies are actively building natural gas infrastructure to meet demand, but they are also investing in solar power and battery storage. Furthermore, the value of nuclear power is being rediscovered. Demand-reducing technologies like smart grids and demand response are gaining importance, and companies are incentivized to find more energy-efficient operational methods. This trend should continue to grow.
To get this story in your inbox, subscribe to the TIME CO2 Leadership Report newsletter .