Emerald AI has raised $150 million in a Series A funding round co-led by Energize Capital and DCVC, valuing the energy technology startup at $1.05 billion.
The round drew substantial participation from strategic and corporate venture arms across the semiconductor, utility, and industrial sectors, including NVIDIA, Samsung Ventures, Siemens, GE Vernova, Aramco Ventures, Salesforce Ventures, RWE, JERA Ventures, and In-Q-Tel. Additional participants include Radical Ventures, Energy Impact Partners, Lowercarbon Capital, Emerson Collective, Tom Steyer, and John Doerr.
The capital injection targets a critical infrastructure bottleneck: the widening gap between surging AI data center electricity demand and transmission grid interconnection capacity.
Power-Flexible Compute Orchestration
Founded in 2024 by Chief Executive Officer Varun Sivaram, Emerald AI develops the Conductor AI software platform. The system coordinates AI training and inference cluster power draws in real time based on electrical grid frequency, wholesale spot power prices, and local utility demand response signals.
Rather than treating data center power as a static baseline load, the software dynamically modulates GPU power states and batch execution schedules to curtail energy use during peak grid stress without interrupting active model training checkpoints.
Industry forecasts indicate that while up to 50 gigawatts of new data center capacity is planned across the United States over the coming decade, regional transmission constraints and substation backlogs threaten to delay commercial interconnection for years. Dynamic power flexibility allows operators to secure grid connection approvals faster by agreeing to throttle non-urgent compute workloads during grid emergencies.
Commercial Scale and Deployment Partnerships
Emerald AI has initiated commercial pilots with utility and infrastructure partners including Silicon Valley Power, Digital Realty, the Electric Power Research Institute (EPRI), and regional grid operator PJM Interconnection.
The startup is also collaborating with NVIDIA on its DSX OS architecture to enable automated power-responsive telemetry across high-density AI clusters, including planned commercial deployments in Texas (under ERCOT's flexible load frameworks) and Virginia later this year.
