Climate commitments are colliding with the rise of Artificial Intelligence, and so far, AI is winning. Energy demands from AI infrastructure are surging to levels that make net-zero targets look increasingly unrealistic, yet countries continue competing to host data center projects. A single large data center now consumes as much electricity as 50,000 homes use in a year. As countries weigh the economic promise of this infrastructure against their pledges to transition toward clean energy, which future will they prioritize?
The rapid growth of AI infrastructure carries significant environmental costs that extend far beyond electricity bills. Data centers generate mounting electronic waste laden with hazardous substances like mercury and lead. They consume vast amounts of water during construction and throughout their operational lives to cool electrical components, even as water supplies become scarce in many regions. The microchips powering these facilities depend on rare earth elements, often extracted through mining practices that devastate local ecosystems. The most pressing concern, however, remains their electricity consumption, which continues to drive up greenhouse gas emissions at an alarming rate.
The International Energy Agency projects that global electricity demand from data centers will more than double by 2030, reaching around 945 terawatt-hours. That’s slightly more than Japan’s entire current electricity consumption. Training GPT-4 alone consumed over 50 gigawatt-hours of electricity, 50 times what was needed for its predecessor, GPT-3. The water demands are equally staggering: a single data center can use more than 500,000 gallons daily, nearly as much as an Olympic-size swimming pool.
As AI spreads into manufacturing, logistics, and other energy-intensive industries, each new application demands more experimentation, more data, and more power. Companies are betting heavily on this future, preparing to invest nearly $7 trillion in data center infrastructure by the end of the decade.
These projections put them on a collision course with international climate commitments. The Paris Agreement calls for cutting carbon emissions by 45% by 2030 and reaching net zero by 2050. The United Nations Environment Programme has urged countries to require greener data centers, powered by renewable energy, and to integrate AI policies into broader environmental regulations. The implementation of these recommendations will face systemic challenges: in a race to capture AI investment these countries may prioritize economic opportunity over environmental commitments.
We already see the effects of this tradeoff in Ireland where massive new data centers are emerging for companies like Meta, Google, Amazon, and TikTok, fueled by burning fossil fuels. A fifth of all electricity used in Ireland is spent powering the country’s data centres, more than is used by its urban homes. With one data centre for every 42,000 people, Ireland has one of the highest per-person concentrations of computing power in the world. As Dublin prepares for the next wave of Artificial Intelligence, power demands have surged considerably, leading Irish households to pay some of the highest electricity bills in Europe. This has sparked criticism and protests from residents who see their electricity bills climbing to subsidize tech giants’ infrastructure. Meanwhile, Ireland’s investments in wind energy remain far from sufficient to meet the soaring demands of its AI economy.
Many other countries with ambitious plans for developing computing power and hosting data centers face similar challenges. Malaysia has invested more than $55 billion USD in growing its data center industry, but the environmental costs are mounting. With 81% of Malaysia’s electricity still generated from fossil fuels, surging energy demands from data centers directly conflict with the government’s goal to increase renewable capacity to 70% by 2050.
The pattern is global. Tech giants are actively courting developing nations to host AI infrastructure, drawn by cheap labor, government subsidies, and lower operational costs. Earlier this year, OpenAI announced plans to support data center construction abroad under the banner of spreading “democratic AI.” These initiatives are financially lucrative but have the potential to cause major delays in countries meeting their decarbonization and renewable energy targets.
Tech companies are increasingly looking beyond traditional markets to countries in Southeast Asia, Africa, and Latin America, but these are also some of the most vulnerable regions to climate change. These areas are already experiencing rising temperatures, severe droughts, deforestation, and water shortages. As the World Meteorological Organization notes, the past ten years have been the warmest on record, with 2024 seeing unprecedented land and sea surface temperatures. These countries must decide whether the promise of AI-driven development justifies worsening the climate impacts already threatening their communities.
Proponents of Artificial Intelligence argue that the technology could eventually aid the energy transition by optimizing grids and improving efficiency, which remains to be seen. What seems certain is that countries won’t voluntarily sit out the AI race while competitors surge ahead. Whether they can balance that ambition with climate commitments that are already slipping further out of reach is the challenge that will define the true cost of the AI revolution.