The most important near-term question in the wake of FERC’s large load rulemaking is not jurisdictional; it is how utilities, RTOs, and regulators will implement flexible interconnection.
FERC’s ruling emphasizes getting more from existing infrastructure broadly and gives runway for non-firm interconnection options specifically. States and transmission organizations should prioritize these pathways. If they fail to do so, it will likely encourage using behind-the-meter gas generation operating as primary supply rather than reliability backstop.
Treating large-load requests as though the only acceptable outcome is fully firm service supported by required network upgrades is misaligned with both system realities and customer preferences. A growing share of prospective data center load can accept some combination of curtailment risk, staged service rights, on-site generation support, battery-backed flexibility, or other operational constraints in exchange for earlier energization. But in many regions, interconnection frameworks remain binary: Either the system is prepared to serve peak demand under all modeled conditions, or the customer waits for studies and upgrades to run their course.
If policymakers get this wrong, they risk more widespread grid defection.
Consider two hypothetical 500-megawatt facilities. Under a flexible interconnection framework, the first facility receives service years before all network upgrades are completed. It accepts curtailment during limited constrained hours each year, potentially supported by on-site batteries or supplemental generation. The overwhelming majority of its energy consumption still comes from the grid. The facility remains part of the planning framework, contributes to system utilization, and helps drive transmission and generation investment.
The second facility reaches a different conclusion. Faced with lengthy queue timelines and uncertainty regarding service availability, it develops dedicated gas generation sized to meet most or all of its expected demand. The grid itself becomes a backup resource, rather than the facility’s primary source of power. In effect, the project exits the traditional utility planning paradigm altogether.
From a reliability perspective, both facilities may ultimately secure adequate power. From a public interest perspective, they are profoundly different outcomes. The first leverages existing infrastructure while creating time for transmission and generation expansion to catch up with load growth. The second accelerates the development of parallel energy infrastructure that sits largely outside traditional planning constructs, and is often accompanied by serious social and environmental externalities.
This is why flexible interconnection deserves to be viewed not as a queue management tool, but rather as a strategic response to the risk of grid defection.
The concept isn’t radical. Non-firm transmission service has existed for decades. Demand response programs compensate customers for accepting curtailment, and interruptible service tariffs are common. On the supply side, generation developers frequently operate within frameworks that allow projects to interconnect and accept congestion risk before every long-term transmission need has been resolved.
ERCOT’s connect-and-manage approach illustrates the principle. It doesn’t eliminate all constraints for new generation before allowing participation, but rather creates a framework in which resources can connect, operate, and respond to system conditions while infrastructure develops.
Large loads should be viewed through a similar lens.
That could take multiple forms. Some customers may accept curtailment rights tied to local transmission constraints. Others may agree to staged service arrangements under which load increases as upgrades enter service. Still others may combine grid service with battery storage, flexible computing architectures, or limited on-site generation for reliability events and system peaks. The specific mechanism matters less than the principle: service products that allow earlier interconnection in exchange for operational flexibility.
Not every data center will accept such arrangements. Nor should regulators assume that all load is equally flexible. The objective should be to expand the menu of options rather than preserve a framework in which full firm service is the only meaningful pathway to interconnection.
The alternative is clear, and worrisome. A growing number of large-load proposals are being paired with dedicated gas generation. Proponents often characterize these projects as reducing pressure on the grid, but the reality is that many are simply a response to limitations in the interconnection process. If obtaining interconnection rights requires waiting six, eight, or 10 years while a self-supplied configuration can be deployed sooner, many will choose the latter.
Meanwhile, many of these gas technologies were not designed to operate as the primary energy source for large, continuously operating facilities. More importantly, widespread deployment of self-supplied thermal generation would create additional demand for fuel supply, transportation infrastructure, and pipeline capacity, while simultaneously reducing opportunities to improve utilization of existing electric infrastructure.
The result could be increasing pressure on gas markets alongside a less efficient trajectory for electric system development. That will raise rates. Utilities, regulators, and consumers would ultimately face many of the costs associated with rapid load growth — while capturing fewer of the associated benefits.
More facilities relying primarily on their own generation would also squander the most significant opportunities for grid investment and modernization in decades. Large, creditworthy customers can support transmission expansion, improve utilization of existing assets, contribute to system cost recovery, and strengthen the economics of new generation development. Those benefits are maximized when load remains integrated with the bulk power system, rather than migrating toward self-supply.
For that reason, the emerging debate over large-load policy should not be framed primarily as a question of jurisdiction, cost allocation, or study procedures. Those issues matter, but they are downstream of a more fundamental choice: between flexibility and defection.
Large loads are going to find power. The capital that has already been committed to AI infrastructure makes that virtually certain. The question is whether energy regulatory frameworks encourage those loads to develop with the grid while infrastructure catches up, or else incentivize a parallel buildout of privately financed generation outside it. If policymakers want the data center boom to strengthen rather than circumvent the electric system, flexible interconnection may be the most important policy tool currently available.
Jackson Ewing, PhD, is the director of energy and climate policy at Duke University’s Nicholas Institute for Energy, Environment & Sustainability. The opinions represented in this contributed article are solely those of the author, and do not reflect the views of Latitude Media or any of its staff.
