Emerald AI Alliance Targets Grid Space for New Data Centers

Emerald AI, a grid-software startup, has formed the AI Energy Management Alliance (AEMA) with Google, Nvidia, Anthropic and utilities AES, Constellation, National Grid and NRG Energy. The alliance plans to use Emerald AI’s platform to coordinate demand-response actions—temporarily pausing non-critical workloads or shifting compute to locations with spare capacity—so that additional data-center capacity can be added without overloading the grid. The startup recently secured a $150 million Series A round led by Energize Capital and DCVC to expand rollout.

TechCrunch

Demand response, long used by large industrial users, lets utilities pay participants to reduce load during peak periods. Traditionally, factories or diesel-generator-backed sites have provided this flexibility. Emerald AI’s software links utility requests directly to data-center operators, enabling rapid, battery-like responses without resorting to diesel generators. Google has been developing similar tools, and Enel X offers UPS-based peak-shaving, but Emerald AI claims its direct utility-to-data-center connection is a distinct advantage.

TechCrunch

Emerald AI estimates that coordinated demand response could free up roughly 100 gigawatts of grid capacity for new data-center connections. A Goldman Sachs study cited by the alliance suggests that limiting grid usage to 90 % for a few hours could release about 76 gigawatts of capacity. These figures indicate that, if widely adopted, demand-response programs could substantially increase the amount of compute power that can be supported without building new generation assets.

TechCrunch

The coalition also intends to assist tech firms and utilities in identifying suitable sites for future data centers, a task that has proven challenging for both parties. While the approach promises to ease grid strain, analysts note that more sophisticated demand-response alone will not solve all current grid reliability issues.

TechCrunch