The growing energy demands of artificial intelligence and its associated data centers, alongside soaring fuel prices exacerbated by global conflicts, are significantly influencing the discussions at New York Climate Week. Experts and activists note that these two factors are changing the conversation from abstract global warming concerns to the tangible economic pressures felt by households and businesses. The UN climate chief, Simon Stiell, highlighted that these developments are driving up food prices and wider inflation, with fossil fuel dependence hitting consumers hard at the fuel pump. For instance, energy shortages from the Iran war alone have cost American consumers over $100 billion in higher fuel prices.
AI's environmental footprint is a major concern, as the data center boom threatens to increase reliance on coal, oil, and gas, even though these centers could theoretically be powered by zero-emission fuels. Stiell warned that "Energy-guzzling artificial intelligence is driving up planet-heating pollution from coal, oil and gas, while ratcheting up energy costs for households and businesses." He added that AI leaders are on "thin ice" regarding public support due to these impacts. Tufts University analyst Jennifer Morgan noted that AI is, in some ways, becoming more vilified in climate talks than the fossil fuel industry itself, partly because its impact on local pollution and electricity bills makes the issue very tangible to the public.
The scale of AI's energy consumption is substantial and rapidly increasing. While AI currently uses about 1.5% of global electricity, this is projected to reach 3% by the end of the decade. The International Energy Agency (IEA) stated that data centers accounted for 1.5% of global electricity consumption in 2024, growing at 15% annually over the prior five years. The IEA forecasts data center consumption to more than double from 415 terawatt hours (TWh) in 2024 to 945 TWh by 2030, though McKinsey estimates it could reach 1,400 TWh by 2030, or 4% of the world's consumption. This rapid expansion, particularly in the US and China which account for about 70% of global data center energy consumption, is creating tension with climate goals, especially in regions with strict environmental targets like Europe, where relying on fossil fuels to compensate for renewable energy intermittency is not an option.
The buildout of data centers is being driven by the development and use of AI, with worldwide installed data center capacity expected to reach 220 gigawatts by 2030, a sixfold increase from 2020. The energy consumed by a single data center can be equivalent to 100,000 households, with larger ones under construction expected to be 20 times bigger. While some optimists, like MIT's Evelyn Wang, believe data centers might not contribute to climate problems in about a decade, the immediate reality involves a significant increase in natural gas power plants to meet current demand. These plants, once built, have a decades-long operational life, further locking in fossil fuel reliance and sparking public pushback over increased pollution and noise.
