The Federal Energy Regulatory Commission’s (FERC) landmark action issuing “show-cause” orders to the nation’s regional transmission organizations (RTOs) has injected a dose of pragmatic urgency into the grid expansion debate. Driven by an unprecedented wave of power demand from artificial intelligence and data centers, the federal regulator is forcing grid operators to modernize their rigid, antiquated frameworks by introducing flexible interconnection pathways for large loads.
This regulatory shift has sparked an essential conversation about how the grid integrates massive energy users. Writing in Latitude Media, Duke University’s Jackson Ewing framed this moment perfectly, noting that the grid’s ultimate choice is between “large load flexibility versus defection.” If utilities and RTOs treat every new data center request as an all-or-nothing proposition by requiring fully firm transmission service backed by a decade of multi-million-dollar network upgrades, developers simply won’t wait. They will “defect” from the grid entirely, bypassing the system to construct privately financed, behind-the-meter generation. “Flexible interconnection,”where data centers connect years faster by agreeing to real-time curtailment during grid stress by either modulating their load when called upon or utilizing on-site generation or batteries, is the strategic safety valve that keeps these large customers integrated into the bulk power system. This approach can help insulate existing grid customers from negative reliability impacts the new load could otherwise create, and to enhance system affordability by reducing the total amount of grid upgrades required.
Ewing’s argument is a good one, but it exposes a glaring asymmetry in how we think about grid bottlenecks. If we limit our policy approaches to load alone, we are fighting a losing battle. To truly accommodate demand growth while maintaining affordability and reliability, regulators and grid operators must take this same logic and apply it to the supply side. We must push for flexible interconnection for generation, rather than treating massive, multi-year transmission expansions as a necessary prerequisite for bringing all new power online.
The parallel on the generation side is stark. Right now, the U.S. power grid is choked by a multi-terawatt backlog of generation projects stranded in interconnection queues. The average wait time to bring a new generator online has stretched beyond seven years. The root cause is identical to the load problem: a rigid, outmoded “invest-and-connect” paradigm. Grid planners analyze interconnections based on a hypothetical, worst-case peak hour of the year. If a new clean energy project triggers a minor localized bottleneck for just a handful of hours annually, it is forced to pay for deep network upgrades to ensure “deliverability.”
Instead of treating physical transmission expansion as the only path forward for new energy, we should embrace a “flex and invest” philosophy. Under a flexible generation interconnection framework, similar to the approach successfully utilized by ERCOT in Texas, power suppliers are allowed to plug into the existing transmission system almost immediately. In exchange for rapid energization, the generator accepts the operational and financial risk of conditional curtailment during those rare hours when the local grid is congested. When paired with smart software, grid-enhancing technologies (GETs) like dynamic line rating, and co-located energy storage, this approach allows cheap, reliable electrons to flow into the market roughly 95% of the time, long before the first shovel hits the ground for a major new transmission line.
The consequences of failing to implement flexible generation interconnection mirror the dangers of load defection. While the data center boom threatens “grid defection” via a parallel buildout of private assets, a rigid generation queue creates “grid stagnation.” As more data centers connect to the grid, the system will become increasingly energy constrained. Faster connection for generation and batteries that inject on a non-firm basis would help meet these needs while pushing rates lower for all customers. If generation can’t connect quickly to meet rising demand, utilities will have no choice but to extend the lifespan of aging, expensive, and high-emitting legacy plants. Ratepayers will bear the brunt of rising wholesale energy costs, and the stability of the entire bulk power system will be tested.
FERC’s bold move proved that decades-old regulatory inertia can be overcome when the economic and national security stakes are high enough. But we cannot build a 21st-century economy by updating only one side of the ledger. True grid resilience and energy dominance require flexibility at both ends of the wire. If regulators want to successfully power the AI era without leaving ratepayers holding the bag, they must recognize that flexible interconnection isn’t just a clever tool for managing data centers, it is the modern playbook for generation as well. We don’t just need more wires; we need smarter ways to plug into the wires we already have.