Nvidia thinks AI can solve electrical grid problems brought on by AI

Nvidia announced Thursday it’s partnering with EPRI, an influence industry R&D organization, to make use of AI to resolve problems facing the electrical grid. Perhaps sarcastically, the problems are largely brought on by rising power demand from AI itself.

The Open Power AI Consortium, which incorporates plenty of electrical utilities and tech firms, says it’s going to use what are often known as domain-specific AI models to plot latest ways to tackle problems that the facility industry is predicted to face in the approaching years. The models can be open sourced and available to researchers across academia and industry.

The facility industry is facing surging demand from data centers in the US and elsewhere as AI ramps up the necessity for computing power. Electricity demand is anticipated to grow by 4% annually in the approaching years, according to the International Energy Agency, nearly double over 2023 figures. 

Along with Nvidia and EPRI, the consortium includes PG&E, Con Edison, Constellation Energy, Duke Energy, the Tennessee Valley Authority, and ENOWA, NEOM’s energy and water company. On the tech side, Microsoft and Oracle are each members.

In an try and stay ahead of the trend, tech firms have been racing to secure generating capability as power has transformed from an easy line item to a competitive advantage.

During the last 12 months or so, tech firms have been consistently inking latest contracts. They’ve largely been spread across renewable energy projects, spurred mostly by solar’s low price, modularity, and the speed at which it could possibly be deployed.

Microsoft, for instance, recently added 475 megawatts of solar energy to its sizable renewable portfolio. Last 12 months, it became an anchor investor in a $9 billion renewable development project run by Acadia, and earlier within the 12 months it said it was working with Brookfield asset management to deploy 10.5 gigawatts of renewable power within the U.S. and Europe, all of which is anticipated to come back online by 2030.

But though latest power sources would be the most evident answer to losing power shortages, they aren’t the just one. 

One recent study found that by curtailing use when demand on the grid peaks, including shifting tasks that aren’t time sensitive to periods when demand is low, a further 76 GB of capability might be unlocked within the U.S. It’s a not insignificant amount, making up roughly 10% of peak demand within the U.S. 

It’s likely those are the styles of solutions, amongst others, that this latest consortium can be exploring.