AI power demand to grow tenfold by 2030, new report says
Add Axios as your preferred source to
see more of our stories on Google.
Data centers built for AI may demand 10 times more power over the next five years, according to a new report from global risk firm DNV.
Why it matters: How much energy the world will need to power artificial intelligence is a hotly debated topic.
- It has giant implications for the billions of dollars going into AI startups and electricity infrastructure across the U.S. and the world.
Driving the news: The firm forecasts that North America — especially the U.S. — will lead global demand among data-center energy consumers over the coming decades.
- DNV estimates that by 2040, U.S. and Canadian data centers will account for 16% of electricity use, with 12% of that tied to AI applications.
Reality check: The report argues that even with this growth, AI's share of global electricity is unlikely to exceed 3% by 2040, staying below demand from sectors such as electric vehicle charging and building cooling.
The big picture: The report from Norway‑based DNV covers a range of topics.
- Its topline projection finds that even sweeping policy reversals under President Trump would have only a "marginal impact" on the broader global energy transition.
- Much like an International Energy Agency's report out Tuesday, this study finds that within the U.S., Trump's moves do "slow that nation's transition markedly," though they do not reverse it.
- Emission reductions are delayed by about five years, it finds.
The bottom line: DNV analyzed over 50 publications with global data center energy estimates. It finds itself on the lower end of such projections — so maybe they will be an understatement.
