Unlock Editor's Digest Lock for Free
FT editor Roula Khalaf will select your favorite stories in this weekly newsletter.
Big technology is getting more energy. Data centers for artificial intelligence applications require additional power supply, making energy transitions a more burdensome movement. But even so, AI's climate benefits could ultimately outweigh its costs.
The immediate issue is clear. The AI boom is rushing to build gas-fired power plants. The intermittent renewable energy means it's difficult to really turn data centers green. This helps explain why Google is still turning to theoretical nuclear fusion.
However, AI-related demand is a new addition to the energy transition plan, but from an absolute standpoint it becomes a relatively small part of the overall system. In the context, the data center consumes today's 415 terawatt hours of electricity. This amounts to around 1.5% of global demand, the International Energy Agency said. The IEA predicts this will more than double by 2030, but electric vehicles and air conditioners will greatly contribute to growing demand.
Some content could not be loaded. Please check your internet connection or browser settings.
On the other hand, AI's reach can be greater than their energy footprint. This technology is expected to improve almost everything we do. Of course, reducing CO2 today for the purposes of climate change is more valuable than reducing it tomorrow. But looking at the numbers at the crisis, if AI even promotes modest savings in overall electricity usage, it is net positive for energy transitions.
That's even more likely, as some pockets of the energy system are surprisingly inefficient. According to Consultant Sander, the material value chain, which produces 6 billion tonnes of steel, glass, hydrogen, ammonia and copper, uses the minimum energy required for the related chemical reactions.
It is a massive warrant opportunity. Finding new materials, catalysts, or processes that can produce things more efficiently is like a “needle in the haystack” problem where AI is ideally suited and can help solve in the biotech sector.
In the battery space, races are chasing breakthroughs with solid state devices that need to be smaller and lighter, and can be used to store energy for longer distance transport. As Microsoft and the US government labs are working together to screen tens of millions of new solid electrolytes for lithium-based batteries, IEA's Simon Bennett narrows it down to 23 potentially viable candidates.
There are also many other things that use AI models to sift through the database. And it's just a drooping fruit. With “smart” appliances and sensors multiplying, AI should help reduce waste such as energy production, transportation, and more. Among many hazy use cases, this can be a power of good.
camilla.palladino@ft.com
