As the AI boom fuels demand for more computing power, hyperscalers and data center developers are running into a growing problem: the grid can’t keep up. Utilities are overwhelmed, and wait times for new grid connections can stretch into years. That’s bad news for an industry racing to build next-generation infrastructure—fast. “A lot of AI data centers just can’t get connected,” said Amit Narayan, founder and CEO of Gridcare. “They’re desperate. The timelines they need—five years or less—are almost impossible using traditional utility processes.”
In response, many data center operators are pursuing expensive alternatives like building their own power plants, known in the industry as “behind-the-meter” sources. But Narayan believes there’s a smarter, more scalable solution.
Unlocking Hidden Capacity with AI and Grid Mapping
Gridcare has been operating in stealth mode until recently, but its mission is now public: to identify and unlock untapped grid capacity and match it with the surging energy demands of data centers.
Backed by years of grid research and AI modeling, Gridcare has mapped large portions of the electrical grid to find overlooked slack in the system—capacity that utilities themselves might not even know exists.
“There’s plenty of available energy out there,” Narayan said. “The key is knowing where and how to find it.”
Gridcare recently raised a $13.5 million oversubscribed seed round led by Xora, Temasek’s deep tech venture arm. Investors include Acclimate Ventures, Aina Climate AI Ventures, Breakthrough Energy Discovery, Clearvision, Clocktower Ventures, Overture Ventures, Sherpalo Ventures, and WovenEarth.
The startup uses generative AI to forecast how energy demand and availability might shift over time, layering in a wide range of variables—fiber optic access, gas and water infrastructure, extreme weather risk, permitting requirements, and even community sentiment around data center construction.
“We run over 200,000 scenarios for every grid analysis,” Narayan said. “Then we verify that data against federal grid regulations before bringing it to utilities.”
Playing Matchmaker Between Data Centers and Utilities
Once it finds a promising site, Gridcare starts conversations with the relevant utility to validate its findings. At the same time, it works with data center developers to identify ideal expansion zones based on power availability and strategic priorities.
“We know what developers are looking for. We know what utilities can offer. Our job is to connect the dots,” Narayan explained.
Gridcare charges data center operators based on how much megawatt capacity it can unlock. While the fee is meaningful to Gridcare’s bottom line, Narayan says it’s a negligible cost for the data centers—especially given how difficult it is to secure reliable power today.
Sometimes, this means asking data centers to briefly forgo grid power and rely on backup systems. Other times, Gridcare helps developers make the case for grid-scale battery installations or new power infrastructure that benefits both sides.
Interestingly, utilities are now exploring the idea of auctioning off access to this newfound capacity. Gridcare could play a key role in helping those auctions become reality.
Narayan believes Gridcare’s strategy could unlock over 100 gigawatts of grid capacity—enough to power thousands of AI data centers.
“We don’t need a miracle like nuclear fusion,” he said. “We just need better data, smarter planning, and collaboration between tech and utilities.”
As AI workloads continue to explode, solving the AI data center power shortage may come down to innovators like Gridcare—who can turn complexity into opportunity and capacity into action.
