(The Center Square) – As electricity demand rises amid concerns about future supply, officials are pushing to modernize the region’s energy mix as reliability and affordability hang on the line.
New Jersey lawmakers are among those in several states trying to add oversight to load forecasting – the process utilities and grid operators use to predict future electricity demand and plan accordingly.
Accurate forecasts matter as AI-powered data centers drive rapid demand growth. Underestimating demand can strain reliability, while overestimating it can lead to unnecessary spending that ultimately shows up in customers’ bills.
Bill A6156 recently advanced unanimously from the Assembly Telecommunications and Utilities Committee and was referred to the Appropriations Committee. The bill would require the Board of Public Utilities, or BPU, to collect and analyze load forecasting information and coordinate with regional grid operator PJM Interconnection to improve accuracy and prevent duplication.
Sponsors – Assembly Democrats Tennille McCoy-Hamilton Square, Dave Bailey, Jr.-Glassboro, and Mitchelle Drulis-Flemington – say the measure responds to concerns raised during legislative energy affordability hearings, including the risk that large projects like data centers can be overcounted when developers submit requests to multiple utilities, creating so-called “ghost projects.”
“Accurate load forecasting is critical to keeping the lights on and energy costs affordable,” McCoy said in a press release.
Baily added, “Our residents should not be paying higher electric bills because of flawed or duplicative data. By ensuring the BPU has full visibility into load-forecasting information, we can reduce inefficiencies and protect consumers from unnecessary costs.”
Likewise, Senate Bill 4938, titled the “Load Forecast Accountability Act,” would also require the BPU to review electricity load forecasting.
Todd Snitchler, president and CEO of the Electric Power Supply Association, told The Center Square New Jersey’s bills are very similar to Pennsylvania’s Load Forecast Accountability Act, or House Bill 1924 – of which some components were rolled into the state budget in November.
They seek to do virtually the same thing, he said, which is to “add some teeth to the load forecasting question and make sure we’re aiming at a target that makes sense.
Snitchler said he thought making data-driven decisions that will inform policymakers is a good step, noting newly elected Gov. Mikie Sherrill’s commitment to minimizing the impact on customers.
PJM, he said, has figured out how to apply more stringent criteria to separate likely-to-be-built projects from speculative ones, which should improve planning and clarify whether the region was truly short of capacity in the last auction.
“You have to give credit where it’s due,” Snitchler said. “We haven’t had the lights go out. It may not be as artful as some would like, but we have managed to ensure a reliable system that continues to deliver results that I think are valuable to consumers of all types.”
Jeffrey Shields, PJM spokesman, told The Center Square, they support all viable efforts to improve load forecasting and that over the past few years, as requests to connect large-load data centers have increased, the grid operator has made a number of improvements to its forecasting process.
PJM relies on its load-serving entities – utilities and retail suppliers – who actually connect projects like data centers to the grid, because they see the interconnection and service requests firsthand.
Annually, in July, those entities report new large load requests, which PJM uses as a key variable in its long-term load forecast.
An analysis of requests submitted for the 2026 Long Term Load Forecast, published back in November, confirmed that electricity demand from data center growth is projected to reach approximately 30 GW between 2025 and 2030 – even as they scaled back some near-term projections, based on several factors. That much power could supply as many as 26 million homes.
Shields added that requests are evaluated and only those projects with “firm” commitments are allowed to impact costs.
Additionally, PJM considers that new data centers won’t typically draw their full power right away. It reviews “ramp rates,” imposing a three-year minimum to reach full demand and assumes 70% utilization unless otherwise supported. Economic data in the load forecast model is also reviewed to prevent double-counting load growth.
“This year, we added the lens of national projections in our evaluation,” Shields said. “Based on available data, PJM is home to approximately 40% of the country’s data center load. The large quantity of “firm” projects indicates this share may increase.
The PJM Board is expected to outline its determination of a path forward on the Critical Fast Path, CIFP, issues in the next few weeks.
