Tech Giants Form Coalition to Use AI Software for Smarter Data Center Power Management

Google, Nvidia, and Anthropic have launched a strategic partnership around Emerald AI, a software startup focused on optimizing how data centers interact with electrical grids. The collaboration, known as the AI Energy Management Alliance (AEMA), brings together major tech companies and major utility providers in an effort to address one of the sector's most pressing constraints: finding enough grid capacity to power new computing infrastructure.

How Demand Response Could Unlock More Capacity#

The core idea relies on demand response—a time-tested practice where large power consumers voluntarily reduce electricity usage during peak periods. Instead of building new power plants, AEMA's approach has data centers temporarily pause less critical computing tasks or redirect workloads to facilities where grid capacity remains available. The coalition estimates this strategy could enable as much as 100 gigawatts of additional data centers to connect to existing grids.

Utilities have long used demand response with industrial users such as factories, which would cut production or switch to backup generators in exchange for compensation. Data centers, which can power down non-essential services nearly instantaneously, are well-suited to this model.

Why Emerald AI Has an Edge#

Emerald AI's competitive advantage lies in its ability to serve as a direct bridge between utilities and data center operators. Its software coordinates power-reduction requests in near-real time, enabling data centers to respond as quickly as battery systems would. Competitors like Google (which has built its own tools) and Enel X (which taps data center backup power supplies) are also pursuing similar solutions, but Emerald AI's dedicated platform approach offers a streamlined pathway.

The startup recently raised $150 million in Series A funding, backing from firms like Energize Capital and DCVC, which positions it to scale deployment across the industry.

The Broader Energy Picture#

Research by Goldman Sachs demonstrated that simply capping grid usage at 90% for brief windows could unlock as much as 76 gigawatts of latent capacity. Yet despite these promising numbers, Emerald AI's own chief scientist acknowledged that smarter demand response will supplement—not replace—the need for new power generation capacity. For data center growth and AI deployment to continue accelerating, the electricity sector will ultimately need both efficiency improvements and fresh generating sources.