As Meta, Amazon, Google, and OpenAI race to build bigger; data center AI is expected to be the “largest driver of electricity consumption” in North America over the next five years, according to a new energy report.
These data centers are packed with thousands of computers to handle everything from training AI models to processing ChatGPT, Gemini, and Sora requests. megawatt power It also requires millions of gallons of water and thousands of acres of land.
According to a report from DNV, an international assurance and risk management company, by 2040, overall energy usage in data centers around the world, including both general-purpose (such as cloud storage and video streaming) and AI-centric data centers, is expected to increase fivefold to 5% of total electricity usage, with AI accounting for more than half of that. But that's the world average. According to DNV, AI data centers alone will account for 12% of electricity consumption in North America in 2040.
Looking broadly at the global energy transition to cleaner power sources, DNV said the world remains “too slow to meet the goals of the Paris Agreement” on climate change to reach net-zero emissions and halt dangerous warming this century.
Don't miss our unbiased technical content and lab-based reviews. Add CNET As a preferred source on Google.
The Trump administration released what it called the “American AI Action Plan” in July, urging the construction of data centers to be expedited, putting regulatory concerns on the back burner. “The U.S. environmental permitting system and other regulations make it nearly impossible to build this infrastructure in the United States at the speed needed.”
But despite rapid advances in AI and the Trump administration rolling back U.S. environmental regulations, the company predicts global emissions will fall by 63% by 2060. The report estimates that what is happening in the U.S. will have a larger domestic impact and will have less of an impact on global clean energy goals because emissions cuts will be brought forward by about five years.
“The United States accounts for one-seventh of the world's primary energy use and has some influence on the overall picture,” the DNV report states. “However, large-scale decarbonization of China's economy continues, coupled with low-cost electrical technology exports from China to other regions.”
Many countries are receiving these resources, and the report states that “China's clean technology exports continue to drive the transition to the rest of the world.”
Meanwhile, the rapid increase in energy usage due to AI is expected to slow. “We find that the initial exponential increase in electricity demand due to AI gives way to a more linear pattern over time,” the report states.
DNV predicts that even in 2040, AI power demand will remain a small share of global power usage compared to EV charging and cooling.
