For decades, capacity planning for the electric grid has operated on a conservative guess. To maintain reliability, utilities plan on how to operate if any two major system components fail simultaneously, in a process called N-2 contingency criterion. On a system with ten thousand parts, that creates a hundred million possible combinations, a mathematical burden traditional utility planning could not actually calculate.
So, engineers conservatively built the grid for the absolute worst-case scenario. That abundance of caution is why the American grid runs at only about a third of its capacity for most of the year, leaving infrastructure idle.
The core issue holding back clean energy deployment in the United States is not a lack of physical power generation but an acute underutilization problem, suggested an episode of the podcast Energy Empire hosted by Jigar Shah.
Shifting from a legacy, conservative planning model to software-driven monitoring can safely unlock 300 GW of capacity on existing infrastructure within three to five years without requiring new transmission lines or physical power plants, said Amit Narayan, founder and CEO of GridCARE as featured guest in Shah’s podcast.
Data from a Duke University analysis led by Tyler Norris reveals the leverage hidden in idle capacity. If large loads accept curtailment for just 0.25% of the year, roughly two hours at a time during the most stressed peaks, the existing grid can absorb nearly 100 GW of new demand.
For utility-scale solar and energy storage developers, this software-driven visibility is a critical lifeline. Interconnection queues across the country have bloated to multi-year waits, where multi-million-dollar network upgrade costs routinely kill otherwise viable solar projects.
Startups like Narayan’s GridCARE are shifting this from a theoretical model to operating rules by running quadrillions of scenarios to locate near-term capacity. Once those specific, highly unlikely constraint hours are mapped, utilities can allow immediate interconnections backed by flexible, interruptible tariffs or paired storage dispatch. Instead of forcing a solar project to wait a decade for a new substation, software identifies exactly when and where the current system has room to breathe.
The software exposes a structural disconnect inside utility headquarters. Right now, long-range planning departments calculate available capacity using static models that completely ignore flexible grid resources. Yet, in the very same company, real-time operations teams routinely dispatch batteries, demand response, and virtual power plants to keep the system stable. Teaching the planning side to account for the tools the operations side already uses is a massive part of the unlock.
The efficiency gains are showing up in weeks rather than years:
- National Grid (New York): GridCARE identified more than 650 MW of previously invisible capacity.
- Portland General Electric (Oregon): The software unlocked over 400 MW in Hillsboro, allowing PGE to connect six data centers years ahead of schedule.
- PJM Interconnection: The nation’s largest grid operator used an AI system to review 811 generation applications representing 220 GW of capacity in under an hour, a process that normally takes weeks.
This software will not eliminate the need for heavy construction. Connecting massive new gigawatt-scale data centers or regional solar hubs will still require building physical substations and heavy transmission lines. But squeezing more value out of the grid ratepayers already paid for lowers the shared cost of the entire system. When you run more electricity through existing wires, the cost to maintain those wires is spread out across a much larger volume of power, which ultimately helps keep electricity bills down for everyday consumers.
The immediate hurdle is institutional trust, said Narayan. Activating this latent capacity requires utilities, regulators, and developers to align on the same data.
The 90-day timebox on grid assessments forces conservative utilities to make a choice: either identify an immediate path to interconnect new generation, or produce hard data proving exactly where the physical wires are constrained. While long-term electrification goals still require tripling the physical footprint of the American transmission network, advanced software provides an immediate bridge out of multi-year queue backlogs today.
For the full broadcast discussion detailing how software uncovers this hidden capacity, check out the Energy Empire Grid Capacity Episode. The interview breaks down how advanced modeling can find immediate room on the existing transmission network.
