For decades, electric utilities have planned around a relatively predictable relationship between population growth, economic activity, weather and electricity consumption. Demand increased gradually, infrastructure investments followed long-term forecasts and utilities generally had time to prepare for changes in where and how electricity was consumed.
That planning model is being tested. The U.S. Energy Information Administration (EIA) projected that commercial electricity sales would surpass residential sales in 2026, as overall U.S. electricity consumption reaches new highs. The rapid expansion of artificial intelligence (AI) data centers, growth in other large commercial facilities and continued electrification are reshaping the electricity landscape.
COMMENTARY
But the significance of this crossover extends well beyond which customer class consumes the most electricity. It signals a fundamental change in the scale, location and concentration of demand, one that challenges the assumptions utilities have relied on for decades.
A single large commercial customer can introduce hundreds of megawatts of new demand into a relatively small geographic area, sometimes on timelines that outpace the planning, permitting and construction of traditional grid infrastructure. At the same time, distributed energy resources (DERs), electric vehicles (EVs) and behind-the-meter technologies are making electricity flows increasingly dynamic.
The question is no longer simply how many additional megawatts the grid will need. Utilities must also understand where those megawatts are needed, when demand will materialize and how much capacity can be made available through better use of existing infrastructure.
Not All Grid Capacity Is Created Equal
One of the biggest challenges with commercial-led load growth is that available grid capacity is not evenly distributed. A utility may have sufficient generation resources or capacity across its broader service territory while facing significant constraints at the substations, feeders or transmission facilities serving a proposed commercial development. In other words, having enough electricity available in aggregate does not necessarily mean it can be delivered where a customer needs it.
Consider the scale of emerging AI infrastructure. Projects such as the Stargate-associated Crusoe data center campus in Abilene, Texas, illustrate the magnitude of the challenge, with development plans at approximately 1 gigawatt (GW). At that scale, accommodating new demand becomes a major transmission, generation and regional infrastructure planning consideration. While projects of this magnitude create significant transmission and regional planning challenges, the same fundamental questions about location, timing and available capacity also apply at the distribution level, where much smaller loads can create highly localized constraints.
But utilities do not need a gigawatt-scale project to encounter serious capacity constraints. Consider a commercial facility proposing to add 20 MW of demand within an existing distribution service area. Whether the utility can accommodate that load depends on the configuration and capacity of the surrounding network, including feeder and substation capabilities, transformer loading, voltage conditions, existing customer demand and the facility’s operating profile.
A feeder that appears to have available capacity during much of the year may approach its operating limits during a handful of summer afternoons. Another feeder may have more consistent headroom but require significant upgrades to support the customer’s full requested load. These distinctions matter because the infrastructure required to resolve each situation could be very different.
The first scenario might benefit from targeted load management or flexibility during constrained periods. The second may require physical upgrades regardless of how effectively existing resources are managed. Utilities must therefore move beyond simply estimating aggregate demand growth and toward understanding where and when new loads intersect with actual system constraints.
From Static Forecasting to Locational Grid Intelligence
Traditional load forecasting remains essential to utility planning. Long-term projections guide investments in generation, transmission, substations and other infrastructure that can take years to develop. However, the emerging load environment requires utilities to connect those long-range forecasts with a more granular understanding of distribution-system conditions.
Advanced metering infrastructure (AMI), grid-edge sensors, distribution automation systems and other operational technologies can provide utilities with a more detailed understanding of conditions across the distribution network. When combined with network models and analytics, these data sources can help utilities evaluate transformer and feeder loading, voltage conditions, localized demand peaks and the effects of distributed energy resources.
This information becomes particularly valuable when utilities connect it with advanced analytics and more dynamic forecasting models. Return to the example of a proposed 20 MW commercial load. A conventional planning approach might evaluate the customer’s requested peak demand against the capacity of the surrounding system. A more adaptive approach would examine how that demand is expected to develop and interact with existing grid conditions.
Will the facility require the full 20 MW immediately, or will its demand increase over several years? Will its peak coincide with the system’s existing summer peak? Are nearby customers adding EV charging, rooftop solar or battery storage? How will those resources affect the local load profile? Most importantly, is the resulting constraint persistent, or does it occur only during a limited number of high-stress hours?
These questions can produce very different investment decisions. For example, a feeder that exceeds its operating threshold for a few hours each year presents a different planning problem than one that is consistently overloaded. Both conditions require attention, but they do not necessarily require the same solution. More granular forecasting can help utilities distinguish between those situations, evaluate potential scenarios and identify where infrastructure investments will deliver the greatest benefit.
This is where grid-edge intelligence becomes especially important. Rather than relying exclusively on historical averages or system-level assumptions, utilities can use more detailed operational information to understand how their infrastructure is performing under actual conditions. That visibility does not create physical capacity on its own. It does, however, provide a stronger foundation for determining where additional investment is genuinely necessary.
Flexibility Becomes Part of the Capacity Equation
Once utilities understand where and when capacity constraints are likely to occur, another opportunity emerges: determining whether some portion of those constraints can be addressed using flexible resources already connected to the grid.
For years, demand response programs largely focused on reducing systemwide peak demand. Today, the growing number of distributed resources creates opportunities to apply flexibility more precisely, matching available resources with specific local grid needs. Residential batteries, smart thermostats, EV chargers, electric water heaters, distributed generation and controllable commercial loads can potentially reduce, shift or otherwise manage demand during constrained periods.
Consider a distribution feeder that experiences capacity limitations during several late-afternoon hours in August. If a utility can reliably coordinate nearby batteries, EV charging and other controllable loads during those periods, it may be able to reduce stress on the feeder without immediately rebuilding every affected asset.
That does not mean flexibility can replace traditional infrastructure. It cannot resolve every thermal limitation, voltage problem or reliability requirement, and a persistent capacity shortage will often demand physical upgrades.
But when the constraint is temporary, flexibility can become a valuable operational tool. The challenge is making that flexibility dependable enough to factor into planning and operational decisions. Five megawatts of enrolled flexible capacity sounds promising. But that number has limited value if the resources are not located behind the relevant constraint, cannot respond during the required hours or cannot sustain their response long enough to address the problem.
Utilities need to know not just how much flexibility exists, but how quickly it can respond and whether it performs as expected. This is where distributed energy resource management systems (DERMS) are becoming increasingly important. By coordinating distributed resources across customer programs and grid operations, DERMS can help utilities identify available flexibility, dispatch resources against specific system needs and measure their performance.
The distinction is important. A resource that participates in a broad demand response program may help reduce overall system peaks. A resource that can be reliably dispatched at a particular location and time can potentially help address a specific distribution constraint.
Moving from the first model to the second requires closer coordination between utility planning departments, grid operators and customer-facing energy programs. It also requires utilities to rethink how they value distributed resources. Rather than treating them exclusively as customer-owned technologies or program participation numbers, utilities can begin evaluating their contribution to system capacity, reliability and operational flexibility.
Build Where Necessary. Optimize Wherever Possible
One of the biggest challenges in planning for commercial load growth is uncertainty. Large commercial projects may be announced years before they reach full operation, and development timelines, energy requirements and even project viability can change along the way.
That creates a difficult balancing act for utilities. Investing too early in infrastructure for demand that is delayed or never materializes can place unnecessary costs on the system. Waiting too long, however, can create capacity constraints that slow economic development and limit the ability to serve new customers.
A more adaptive approach allows utilities to align infrastructure investments with how demand actually develops. This may include phased upgrades, closer coordination with large commercial customers and the strategic use of flexible resources to manage constraints while longer-term infrastructure is planned and built. The objective is not to avoid new construction, but to make more informed decisions about where, when and how much capacity to add.
Every Megawatt Has to Work Harder
The shift toward commercial-led electricity demand growth is changing more than how much capacity utilities need. It is changing how they must evaluate and plan for that capacity.
As large loads emerge in concentrated locations and on uncertain timelines, the ability to distinguish between immediate constraints, future requirements and opportunities to use existing infrastructure more effectively will become increasingly important.
For utilities, the challenge is no longer simply preparing for more megawatts. It is making sure every investment, and every available megawatt, delivers the greatest possible value as demand evolves.
—Nick Tumilowicz is director of Product Management, Distributed Energy Management Solutions, at Itron.
