Google, Nvidia, and Anthropic back Emerald AI's grid coalition

Grid software unicorn Emerald AI has formed a coalition with Google and Nvidia that will use its technology to secure space on the electrical grid for more data centers. The group, called the AI Energy Management Alliance (AEMA), wants to make demand response — a practice in which data centers temporarily reduce their electricity use — a standard part of data center development. By pausing noncritical tasks and shifting some compute loads, AEMA says, the approach could allow an additional 100 gigawatts of data centers to connect to the grid.

The founding trio is joined by Anthropic and a range of utilities, including AES, Constellation, National Grid, and NRG Energy. Demand response itself is an old idea that utilities have used for decades: because the grid is built to handle peak loads, demand sits well below the maximum the grid can support for much of the year. Large electricity users like factories have traditionally agreed to cut usage during high-demand periods — often by pausing production or switching to backup generators — in exchange for payment from the utility, sometimes a substantial one. Data centers now tap those same programs too, usually by switching on backup generators of their own.

Emerald AI is pitching an alternative to those diesel generators: software that coordinates requests from utilities with data centers, which can pause noncritical tasks or shift workloads to other facilities where the grid has spare capacity. It isn’t alone in the space — Google has been building its own tools, and Enel X lets data centers tap their uninterruptible power supplies to shave peak demand. A Goldman Sachs study published last year found that capping grid usage at 90% for a few hours at a time could free up 76 gigawatts of capacity. Emerald AI’s edge, though, is that its software connects utilities directly to data centers, letting them respond to requests almost as quickly as batteries can. The startup also recently raised a $150 million Series A led by Energize Capital and DCVC, funding it plans to use to roll the technology out more broadly.

Given how peaky AI compute loads are, and how quickly data centers can ramp usage up or down, a coalition like AEMA has been a long time coming; it also aims to help tech companies and utilities identify new sites for data centers, something that has proven difficult for both sides. Even so, Emerald AI chief scientist Ayse Coskun told TechCrunch that more sophisticated demand response won’t solve every problem facing the grid today — it can blunt the industry’s need for new generating capacity, but it won’t eliminate that need entirely.

Why it matters

As AI compute demand collides with a grid built for steadier loads, tools that let data centers flex their usage in real time could determine how fast new capacity — and the power-hungry AI models that need it — can actually come online.