Go back

The Hidden Cost of AI: The Energy Crisis Behind the Digital Revolution

Artificial intelligence is no longer just a futuristic promise; it’s here, and it's changing our world at a dizzying pace. Chatbots that write articles, image generators that create works of art in seconds, systems that optimize global logistics... the list is endless. But behind this digital magic lies a huge cost, and we're not talking about money, but about terawatt-hours. AI is devouring our planet's energy at a rate that is worrying experts and analysts.

 

AI Data Centers: The New Industrial Giants

According to a report by the International Energy Agency (IEA), data centers—the physical “brains” of AI—consumed about 415 TWh of electricity in 2024, which is 1.5% of the total global consumption. The United States leads this wave with 45% of that consumption, followed by China (25%) and Europe (15%). But these numbers are just the beginning of a much larger, more worrying story.

This isn’t just gradual growth. The IEA warns that electricity demand from data centers is set to more than double by 2030, reaching 1,050 TWh. This figure is higher than the entire current energy consumption of Japan. In the IEA’s base scenario, by 2035, consumption could hit 1,300 TWh. To give you a more concrete idea, a single AI data center can use as much energy as 100,000 homes. The largest ones under construction are projected to use 20 times more, putting them on par with massive industrial plants like aluminum smelters or steel mills.

 

Behind the Scenes: The Engine of AI’s Energy Boom

The engine behind all this is generative AI. Tools like ChatGPT and video generators require unprecedented computing power, consuming far more energy than traditional digital services like streaming or cloud computing. Investments in this infrastructure have exploded, nearly doubling since 2022 to $500 billion in 2024, confirming the market’s direction. Other factors include the need for increasingly powerful AI models and their expansion into crucial sectors like healthcare, transportation, and manufacturing.

 

A Power Struggle: Is the Grid Ready for AI?

All this energy demand poses a critical question: Can our electrical grid handle the load? In many regions, the answer is no. The IEA estimates that about 20% of planned data center projects could face delays due to a lack of grid connection capacity. Wait times for new power lines can be up to 8 years in advanced economies, compounded by equipment shortages and limited space in existing infrastructure.

A close collaboration between data center operators and energy providers is essential. AI companies need to stop building just anywhere and start locating in areas with strong grids and an abundance of clean energy.

 

A Green or Gray Future? The Energy Mix for AI

To meet this growing demand, the future lies in diversification. The IEA predicts that about 50% of the electricity growth for data centers will come from renewables, primarily wind and solar. While clean, these sources aren’t always available, so a mix with natural gas, nuclear, and geothermal will also be necessary. The IEA forecasts a significant rise in generation from gas and nuclear, especially in the U.S., China, and Japan, with new small modular reactors (SMRs) expected to come online around 2030.

 

The Carbon Question: Is AI a Climate Friend or Foe?

There’s concern that AI’s huge electricity consumption could worsen climate change. In the IEA’s base scenario, data center emissions will grow from 220 million tonnes (Mt) in 2024 to 300-320 Mt by 2035.

 

Despite this, AI can also be part of the solution. The very technologies that consume energy can also help optimize and reduce consumption elsewhere. For example, AI systems can:

• Improve power grids by reducing energy losses.
• Optimize building consumption by adjusting heating and cooling.
• Make transportation and industry more efficient by improving logistics and planning.

According to the IEA, widespread adoption of AI solutions could reduce global emissions by up to 5% by 2035, which is more than the emissions produced by data centers themselves. In a sustainable future, AI could be a key ally in the fight against climate change.

The AI revolution is inevitable. The real question is: Can we manage its energy growth responsibly, turning a potential threat into an opportunity for a more efficient and sustainable future?