Emerald AI Raises $150 Million to Optimize Data Center Power
Emerald AI raised $150 million in Series A funding to deploy software that allows data centers to adjust energy use based on grid demand.
AI start-up Emerald AI raised $150 million in an oversubscribed Series A funding round co-led by DCVC and Energize Capital, bringing the company's valuation to $1.05 billion. The company develops the Emerald Conductor platform, which uses AI agents to turn data centers into flexible electricity consumers by slowing, pausing, or moving workloads when power grids experience stress.
Founded in 2024 by CEO Varun Sivaram and chief scientist Ayse Coskun, the company aims to unlock over 100 gigawatts of untapped capacity on the U.S. power grid to bypass infrastructure delays. The start-up has completed demonstrations in London, Arizona, Illinois, Virginia, and Oregon, including a Phoenix field test that reduced an Nvidia cluster's power draw by 25%. It has since moved into commercial deployment with a full data center in California.
Emerald AI is currently partnering with Nvidia, Digital Realty, Nebius, and EPRI to launch the first power-flexible AI data center in Manassas, Virginia, later this year. Additionally, the company is collaborating with Silicon Valley Power on the first Flexible Load Interconnection Program in the United States, which grants data centers expanded grid access in exchange for dispatchable power flexibility. Strategic investors in the round included Nvidia, Samsung Ventures, GE Vernova, and Salesforce Ventures.
