In a time when AI data centers are scrambling for power and utilities are overwhelmed with connection delays, a stealthy startup called Gridcare has found a surprising solution. There’s unused energy hiding right inside the grid.
Founder and CEO Amit Narayan, a Stanford researcher-turned-tech entrepreneur, told TechCrunch that data centers are so desperate for electricity that many are building their private power plants. But Narayan insists that a better answer has been overlooked: “We believe more than 100 gigawatts of capacity is available in the grid, it just hasn’t been discovered yet.”
AI Mapping + Energy Strategy = Matchmaking Magic
Gridcare, fresh off a $13.5 million seed round led by Temasek’s Xora, is taking a bold new approach. The company uses AI to study the entire power grid, layering in thousands of variables — from weather to fiber optics, water access, natural gas, permits, and even local opinions about development.
“There are over 200,000 possible scenarios we evaluate,” Narayan said, adding that they cross-reference results with federal energy regulations to avoid missteps.
Once potential hot spots are found, Gridcare speaks directly with utilities to verify the opportunities. At the same time, they talk to hyperscalers and data center developers to learn where new sites are needed. When all the pieces align, Gridcare makes the match.
“We’ll find out where the maximum bang for the buck is,” said Narayan.
Unlocking Energy, Selling Megawatts
Gridcare makes money by charging data centers based on how many megawatts it can unlock for them. According to Narayan, the cost is tiny compared to the value it brings to clients.
Some deals might involve small trade-offs, like shutting off grid power during peak hours and switching to backup. Others might trigger a new battery project or energy upgrade in the area. Utilities are even considering auctioning off this newly found capacity.
For now, Gridcare’s biggest win is proving that there is enough power for AI’s growing appetite — it just takes smarter planning.
“We don’t have to solve nuclear fusion to do this,” Narayan said.
