AI’s rapid expansion is putting increasingly new pressure on electricity grids as companies build larger data centres to train and operate demanding models.
This pressure is changing how the AI industry plans its infrastructure, and new power generation, grid upgrades, as well as onsite energy storage are part of the response. Companies are also exploring software that can adjust computing demand during periods of grid stress, allowing existing power capacity to be used more efficiently.
It is in this context that Emerald AI has now raised $150 million in an oversubscribed Series A round to expand this approach. The funding values the Washington, D.C. company at $1.05 billion and brings its total capital raised to more than $220 million.
Energize Capital and DCVC co-led the round, with Nvidia, Samsung Ventures, Siemens, GE Vernova, RWE, and other technology and energy companies also investing.
The mix of investors reflects how electricity management is becoming a central part of AI infrastructure planning. Emerald AI will use the funding to expand commercial deployments with AI companies, data centre operators, and electric utilities.
How Emerald Conductor Reduces Grid Pressure
Emerald AI’s Conductor platform links data centres with utilities and responds to changes in grid conditions. The software can briefly slow or pause AI work that does not need to run immediately, move workloads to another region where power is more available, and coordinate onsite energy resources such as batteries.
These changes allow a facility to lower its power draw at the times when the grid needs relief, as Emerald AI says the system works within performance and latency limits designed to protect critical AI workloads.
The company has tested the technology with partners including Nvidia, Oracle, EPRI, Nebius and National Grid. During peak electricity demand on a hot afternoon in Phoenix, Conductor reduced AI power use at an Oracle data centre by 25% for three hours. According to Emerald AI, the results were peer reviewed.
“Our demonstrations around the world proved that data centres can adjust their power use precisely when the grid needs relief, without compromising critical computing workloads,” founder and CEO Varun Sivaram said.
AI’s Power Demand is Rising Quickly
The funding arrives as electricity demand from AI infrastructure grows faster than many power grids can add capacity. The International Energy Agency expects global data centre electricity use to more than double by 2030 to about 945 terawatt-hours, slightly above Japan’s current annual electricity consumption.
In the United States, data centres are expected to account for almost half of the growth in electricity demand through 2030. New grid infrastructure can also take years to plan and build, leaving some data centre projects waiting for power connections.
Emerald AI estimates that flexible data centres could unlock more than 100 gigawatts of capacity on the existing U.S. grid. This figure is the company’s estimate and will depend on how widely the technology is deployed.
Commercial Expansion Comes Next
Emerald AI has moved beyond trials with a full-data-centre deployment in California. It is also working with Silicon Valley Power on a programme that offers data centres expanded grid access in exchange for verified and controllable reductions in electricity use.
The company is also working with Nvidia and Digital Realty on a 96-megawatt Vera Rubin AI Research Factory in Manassas, Virginia, with the power-flexible facility expected to come online later in 2026.
The new funding gives Emerald AI the capital and industry partners to take its system into more data centres. And its commercial rollouts will now show whether flexible computing can ease grid pressure as the global AI buildout continues.
