Emerald AI has raised a $150 million Series A at a $1.05 billion valuation, the Washington, D.C. company announced on August 25. The oversubscribed round was co-led by Energize Capital and DCVC, and ESG Today reports the investor list runs to more than twenty names, including Nvidia, Siemens, Samsung Ventures, Salesforce Ventures, Aramco Ventures, GE Vernova, RWE, JERA Ventures, In-Q-Tel, Radical Ventures, Energy Impact Partners, Lowercarbon Capital, Emerson Collective, Earthshot Ventures, and individually John Doerr and Tom Steyer. The company is led by founder and CEO Dr. Varun Sivaram.
What the software actually does
Emerald's product, the Emerald Conductor platform, coordinates a data center's AI computing workloads alongside its onsite energy resources — batteries, generators — so the facility can cut its draw from the grid during periods of stress, then resume. The pitch is that this is done selectively: latency-tolerant training and batch jobs get shifted or slowed while time-critical inference workloads keep running at full speed.
"Our demonstrations around the world proved that data centers can adjust their power use precisely when the grid needs relief, without compromising critical computing workloads," Sivaram said in the announcement.
The company has run five commercial data-center demonstrations — in Arizona, Illinois, Virginia, Oregon and London — with partners including Nvidia, Oracle, Nebius, EPRI, National Grid, and regional utilities and grid operators. Its next deployment is in Manassas, Virginia, where Emerald is working with Digital Realty and Nvidia on the roughly 100-megawatt Vera Rubin AI Research Factory, alongside EPRI, Dominion and grid operator PJM Interconnection. That facility is expected online later in 2026.
The 100 GW claim
Emerald's central claim is a capacity number: applying flexible-load management across the AI data-center build-out could unlock more than 100 GW on the existing US power grid — capacity available years before new transmission and generation could be built.
That constraint is visible across the buildout Uristocrat has tracked. Developers have gone hunting for power at enormous scale, from Nvidia's talks to guarantee $250 billion for OpenAI's Ohio data center to Anthropic's $10 billion, six-year compute deal for a Norway facility, and into new generation itself, with Valar Atomics raising $1 billion at a $6 billion valuation to build reactors.
Data centers increasingly draw local opposition over what they take from a shared grid — a fight Uristocrat covered when Fisk University's $400 million campus data center drew community pushback.
DCVC framed the round around that shift. "Emerald AI's compute workload orchestration platform makes flexibility a permanent feature of how data centers are powered," said Zachary Bogue, co-founder and managing partner at DCVC. "This turns data centers into grid-responsive assets instead of energy-hogging liabilities."
Strategic and industrial investors turned up again days later, when Gatik raised a $200 million Series D co-led by the Qatar Investment Authority and Koch Disruptive Technologies.
Source: ESG Today
Comments