GridBeyond raises €12M to scale virtual power plant software

GridBeyond has raised a €12 million ($13.8 million) equity round led by Samsung Ventures to expand its virtual power plant software. Virtual power plants are systems that aggregate disparate energy sources to behave as a single power plant, helping to stabilize the electrical grid. The startup plans to use the funding to expand its portfolio of managed assets.

The funding round featured participation from a list of investors:

• ABB
• Act Venture Capital
• Alantra’s Energy Transition Fund
• Constellation Technology Ventures
• EDP
• Energy Impact Partners
• Enterprise Ireland
• Klima
• Mirova
• Yokogawa

The startup installs hardware controllers in batteries, renewable power plants, and large commercial and industrial facilities. It currently manages around 1 gigawatt of solar, batteries, wind, and hydropower capacity. On the demand side, Michael Phelan, co-founder and CEO of GridBeyond, stated that the company manages several gigawatts of flexible load across commercial and industrial facilities. The company operates in Australia, Ireland, Japan, the U.K., and the United States, including managing a 200-megawatt battery installation in California.

GridBeyond originally developed its technology to address grid-balancing challenges in Ireland as the island added wind power. “The problem on the grid is a peak problem. Most of the time you’re okay, you have plenty of power. But in those peak hours you might not have enough,” Phelan said. By utilizing flexible load and battery storage, grid operators can shave peaks off demand during extreme periods, which is cheaper than building new transmission lines or power plants. Batteries also respond quicker to demand than traditional peaking power plants, allowing the company to buy and sell power rapidly in a form of arbitrage.

The company is now targeting the data center market, where facilities require large amounts of electricity to train and operate AI models. These facilities do not always draw power continuously, often peaking during AI training, which can cause oscillations on the grid. Batteries located at data centers can absorb this load and smooth out the facility’s power profile. According to Phelan, having enough stored energy or flexible industrial load allows developers to build hyperscalers—large-scale data center operators—without causing unwanted grid collapses. He added that plugging into these systems makes it easier for facilities to secure grid connections.

Why it matters

GridBeyond’s software helps energy-intensive data centers manage grid load—a critical bottleneck for AI infrastructure—by aggregating distributed energy resources into virtual power plants.