New York's Hidden Megawatts

GridCARE and National Grid are using AI to slash large-load connection times from years to months in upstate New York

14 Apr 2026

Connecting a data center or semiconductor plant to the US power grid has long been an exercise in patience. The traditional process can take anywhere from three to seven years, a timeline that has become a serious liability as demand for electricity accelerates across the country. A new partnership between GridCARE and National Grid, announced March 25, aims to change that, at least in upstate New York.

The core idea is deceptively simple. Rather than building new infrastructure, GridCARE's Energize platform uses generative AI and physics-based simulations to find capacity that already exists within the grid but goes unused. Stanford University research suggests US networks operate at roughly a third of their potential, which means the power needed to serve a new data center or manufacturing facility may already be sitting there, waiting to be unlocked.

National Grid serves more than 20 million customers across New York and Massachusetts. Its upstate New York territory has attracted growing interest from data center developers and large manufacturers, all of them facing the same chokepoint: grid connection timelines that can stretch nearly a decade. GridCARE's approach, which incorporates battery storage and distributed energy assets to release latent capacity, targets that bottleneck directly while keeping customer rates protected.

The results elsewhere have been hard to argue with. GridCARE has already brought hundreds of megawatts online ahead of schedule, generating more than $10 billion in economic value across its deployments. The National Grid collaboration brings that track record into one of the country's most competitive and capacity-constrained markets.

The broader stakes are not lost on anyone in the industry. As AI infrastructure demands ever-greater amounts of power, the utilities that can move quickly and intelligently stand to shape where that growth lands. Software that finds hidden grid capacity, without the cost or the wait of new builds, is fast becoming one of the most consequential tools in energy planning.