GridBeyond raises €12M led by Samsung to grid-power AI
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- GridBeyond raised a €12 million ($13.8 million) equity round led by Samsung Ventures, with ABB, Act Venture Capital, Alantra's Energy Transition Fund, Constellation Technology Ventures, EDP, Energy Impact Partners, Enterprise Ireland, Klima, Mirova, and Yokogawa also participating.
- GridBeyond already manages around 1 gigawatt of solar, batteries, wind, and hydropower on the supply side and "several gigawatts" of demand-side flexibility across commercial and industrial facilities, co-founder and CEO Michael Phelan told TechCrunch.
- The grid's core problem is a peak problem, Phelan said — most hours are fine, but peak demand can outstrip supply, a constraint now hitting data center developers and tech companies powering AI training and inference hardest.
- GridBeyond operates a 200-megawatt battery in California and argues that on-site or nearby batteries let data centers absorb AI training spikes, smoothing the oscillations on the grid that Phelan likened to the event that collapsed the Spanish grid.
- Batteries also enable rapid arbitrage because they respond far faster than traditional peaker plants, which can take minutes to come online — a speed advantage GridBeyond is selling to hyperscalers trying to secure grid connections.
- Dublin-headquartered GridBeyond traces its origins to Ireland's island-grid balancing challenges as wind power scaled, and now has hardware controllers installed in Australia, Ireland, Japan, the U.K., and the United States.
Why it matters: AI and data center buildouts are colliding with a grid that has plenty of average capacity but not enough peak capacity. GridBeyond is selling hyperscalers a way to get connected faster by turning flexible industrial loads, renewables, and batteries into dispatchable virtual power — and Samsung Ventures' lead check signals that the world's biggest memory and chip players see grid-edge software as strategic infrastructure for the AI buildout.
