Google, Nvidia and Anthropic want Emerald AI to find space on the grid for more data centers

Google, Nvidia, Anthropic, and a consortium of utilities have united under the AI Energy Management Alliance (AEMA) to leverage Emerald AI's software in securing space on the grid for additional data centers. The aim is to incorporate demand response into the development of data centers, which involves temporarily reducing electricity consumption from noncritical tasks and shifting some compute loads. This could potentially accommodate an extra 100 gigawatts of data center infrastructure on the grid.

Demand response, an established practice in the utility sector for decades, capitalizes on the fact that the grid is designed with peak loads in mind. During peak periods, utilities offer compensation to large electricity users like factories to reduce usage by pausing production or switching to backup generators. Data centers can participate in these programs through the use of Emerald AI's software, which coordinates requests from utilities with data centers.

This coordination could involve pausing noncritical tasks or redistributing loads to other data centers that have available capacity.

Emerald AI's software differs from traditional demand response approaches by directly connecting utilities to data centers. This direct connection enables data centers to respond swiftly to utility requests, similar to the rapid response capabilities of batteries. The startup recently secured $150 million in Series A funding led by Energize Capital and DCVC, providing the necessary resources to expand the technology's reach.

Emerald AI's solution has several advantages over alternative methods. By connecting utilities directly to data centers, it allows for quicker responses to requests, akin to the immediate response of batteries. Moreover, the company recently raised substantial funding, underscoring the potential for widespread adoption of its technology.

Given the nature of AI compute loads, which tend to be unpredictable and require rapid scaling up and down, the formation of AEMA is a timely development. The coalition also intends to facilitate the discovery of new data center sites, a process that has been challenging for both tech companies and utilities.

Despite its potential, AEMA's demand response strategies are unlikely to resolve all the challenges facing the power grid today. Ayse Coskun, Emerald AI's chief scientist, acknowledges that while her company's technology could alleviate the industry's reliance on new generating sources, it is not a panacea.