From 5 to 95: Why Load Forecasting Needs a New Approach

During my time as a Senior Director of Load Forecasting for one of the largest multi-jurisdictional utilities in the United States, my team would typically incorporate around five major new-load projects into a planning-cycle forecast.

Then, in one cycle, that number became approximately 95.

The rapid growth of data centers and other large-load opportunities meant we were suddenly evaluating projects at very different stages of development, each with its own proposed in-service date, ramp-up schedule, expected load factor and likelihood of moving forward.

This was not simply a matter of processing more projects. It fundamentally changed the forecasting challenge.

Some projects could be delayed, downsized, or accelerated. Others might never materialize. Even small changes in timing or load assumptions could materially alter the demand outlook and, in turn, the resource decisions being considered by the utility.

Our established forecasting process was no longer enough.

Moving beyond one forecast

Utilities have always dealt with uncertainty. What has changed is how quickly those factors are evolving and how strongly they now interact.

Electrification, distributed energy resources, and large-load growth are changing both the scale and shape of demand. Data centers can introduce city-scale load over a relatively short period, often with considerable uncertainty around timing.

That creates a direct challenge for Integrated Resource Planning (IRP). Decisions around generation, transmission, storage and demand-side programs all depend on a view of future customer demand.

Historically, utilities often centered planning on one “most likely” forecast, supported by high- and low-growth sensitivities. That approach is increasingly difficult to defend when a relatively small number of large projects can materially reshape the outlook.

In the case of the 95 potential projects my team was evaluating, we developed low-, medium- and high-load outcomes based on different assumptions around project timing, customer readiness, ramp rates, and expected demand.

The aim was not to create an unlimited number of possible futures. It was to give planners a credible and manageable range of outcomes against which they could test major decisions.

Scenario planning cannot remove uncertainty, but it can make that uncertainty visible and easier to plan around.

Keeping the forecast current and defensible

Scenarios are only useful when the assumptions behind them can be reviewed and refreshed as conditions change.

Large-load projects move through different stages of development. Customer plans change, new information becomes available, and economic conditions shift. A planning forecast therefore cannot remain a point-in-time exercise that is revisited only during the next formal planning cycle.

In our case, each major project required its assumptions to be researched, documented, and incorporated into the broader demand outlook. The aggregate impact then had to be reconciled with economic forecasts, customer trends, transmission plans, and resource-planning assumptions.

That experience reinforced an important lesson: the value of a forecast depends not only on the resulting number, but on how clearly its assumptions can be updated and explained.

Utility planners, executives, and regulators need to understand why a forecast has changed, which assumptions are driving the result and how alternative scenarios were developed.

Technology can make scenarios easier to update and assumptions easier to trace, but it must be supported by close collaboration between forecasting teams and the utility functions responsible for customer growth, system planning, and resource planning.

A strategic planning capability

Load forecasting has evolved from a largely analytical function into a strategic capability that shapes utility investment, affordability, regulatory confidence, and long-term reliability.

My experience taught me that the goal is not to predict one future with certainty. It is to give planners a credible range of outcomes, keep the assumptions behind them current and clearly explain why the outlook has changed.

Now, at GridX, I see utilities working through those same challenges at greater scale and speed. Our role is to help them bring their assumptions and scenario work into a more dynamic and consistent forecasting process, with the granularity and traceability needed to support sound planning decisions.

In an environment where the load outlook can change materially during the planning process itself, understanding more than one possible future is becoming essential.

Michael Pirro is Senior Director of Rate Design at GridX. He has more than 30 years of utility-industry experience across leadership roles at Duke Energy, PG&E, Florida Power & Light, and National Grid. His background spans load forecasting, rate design, resource planning and regulatory engagement, giving him a practical perspective on how utilities can respond to accelerating load growth and increasing planning uncertainty.