How a utility’s recalibration of data center power demand reveals the importance of disciplined load forecasting to protect consumers and strengthen reliability.
At a Glance:
- An investor-owned utility recently cut projected data center energy demand from 30 GW to 13 GW after requiring information that validates proposed projects.
- Speculative forecasts can inflate costs for consumers; making sure the forecast is as accurate as possible is critical.
- Competitive power markets manage risk and promote efficient investment.
- EPSA advocates for transparent, disciplined forecasting to ensure reliability and protect ratepayers.
In October, an investor-owned utility announced it was cutting its projected data center electricity demand in half — from 30 gigawatts to 13 gigawatts — after implementing new tariff rules requiring binding financial commitments from developers. Under these provisions, projects must submit non-refundable study fees and demonstrate firm contractual obligations through signed Letters of Agreement or Energy Service Agreements, which can span up to 12 years with minimum demand commitments exceeding 80 percent. Only projects meeting these thresholds are now included in the forecast.
The move revealed how speculative projects can distort load forecasts, leading to inflated expectations, potentially costly planning errors, and could lead to bad policy outcomes. This recalibration underscores a growing challenge for planners and policymakers: distinguishing between speculative demand and real projects with certain benchmark criteria to demonstrate they are real projects.
For decades, population growth, economic activity, and weather patterns formed the big three of electric load forecasting. As energy planners now account for emerging technologies, electrification, and booming data center growth, forecasting has become significantly more complex — and increasingly prone to error if the inputs are not correct. The result? Forecasts that swing widely from scarcity to surplus, driving concerning headlines and ever louder rhetoric for policy reactions that don’t match reality.
Market Explainer: What Is Load Forecasting?
Load forecasting is the process of predicting future electricity demand based on population growth, economic activity, weather patterns, and emerging technologies. It helps utilities, grid operators, and policymakers plan how much power will be needed in the future — and when. Accurate forecasts are critical to avoid two costly extremes: underbuilding, which risks blackouts and could cause reliability issues, and overbuilding, which can saddle consumers with unnecessary infrastructure costs and potentially stranded costs. As new loads like data centers and electrification increase demand potentially significantly, improving the accuracy and transparency of load forecasting is more important than ever.
Understanding the Forecasting Challenge
Across the PJM Interconnection footprint and the entire U.S., energy planners are working to model the impact of data center expansion on electricity demand. Because many projects are still in the early stages of development, their final location and power needs can be difficult to predict, and the double- and triple- counting phenomenon routinely occurs. That uncertainty can lead to short-term volatility in forecasts, even when long-term growth remains strong.
By improving communication between developers, utilities, and grid operators — and by ensuring that assumptions are based on credible commitments — planners can more accurately match infrastructure development with real needs. The recent utility adjustment shows the value of using updated, verifiable data to strengthen planning and maintain public confidence.
Learning From the Past
Accounting for unexpected rapid growth is nothing new for grid operators. In the early 2000s, when data center electricity use surged nearly 90% between 2000 and 2005, followed by another 24% jump through 2009, many predicted an era of runaway demand. Yet from 2010 to 2018, global data center electricity use flatlined. These cycles exposed a key flaw: forecasts that overestimate demand don’t lead to an abundance of idle energy resources — they lead to misallocated investment and higher costs.
When demand projections overshoot reality, it can misdirect investment. In competitive power markets, independent power producers — not customers — bear the financial risk of those planning errors. But in vertically integrated utility territories, ratepayers can directly shoulder the cost of generation built for demand that never materializes. That’s why transparent, market-based forecasting remains so important.
Why Accurate Forecasting Matters Now
A recent Rystad Energy review identified over 100 gigawatts of projected U.S. data center demand by 2035 — enough to power the entire country today. Meanwhile, the Electric Power Research Institute found a wide range of potential data center load outcomes, from 4.6% to 9.1% of total U.S. consumption — a 200 terawatt-hour gap, roughly the usage of 11 million homes annually. Within PJM, peak load forecasts for 2026/2027 jumped by over 5,400 MW, driven largely by expected data center growth.
One of the biggest common denominators in inaccurate forecasting models is transmission. Load forecast assumptions are hindered by transmission, and any amount of additional load coming onto the grid needs the requisite transmission infrastructure. The record load forecast delivery depends on massive expansions of our transmission grid, but capacity additions are falling while costs to add new resources continue to increase.
Even more burdensome, forecasts tied to speculative projects can misdirect where new transmission is built, resulting in bad investments and making it harder to respond to actual needs as they emerge.
Competitive Markets Help Manage Forecasting Risk
Competitive electricity markets play a crucial role in managing the uncertainty of future load growth. Competitive procurement of power introduces discipline and helps ensure that investment follows actual market signals rather than speculative projections. In a competitive market, independent power producers shoulder the financial risk of overestimating demand, protecting consumers from stranded costs.
By contrast, in regulated monopoly systems, those costs are typically passed directly to ratepayers. Competitive procurement not only delivers cost discipline and innovation but also provides flexibility to scale generation efficiently as verified demand emerges. These market-driven solutions help align planning with real-world conditions while encouraging private investment in reliability and resource diversity.
EPSA’s Perspective: Getting It Right the First Time
Accurate load forecasting is foundational to ensuring that the U.S. meets rising electricity demand without wasting capital on infrastructure that may never be used. The AEP Ohio case provides a model for responsible forecasting practices: requiring real financial commitments from developers helps align planning with actual need. This approach protects consumers, supports reliability, and promotes efficient investment.
EPSA continues to engage with grid operators, utilities, and regulators to improve the rigor of load forecasts. By promoting transparent assumptions and focusing on real, financed projects, EPSA advocates for policies that encourage both investment in new generation and ratepayer protection.
The Bottom Line
Accurate information is critical to reaching the right outcomes for all stakeholders. Load forecasting errors can lead to costly overbuilds that ratepayers ultimately fund, and policy overreactions – while realistic, data-driven planning ensures reliability for the 67 million customers in PJM’s footprint. The true solution to meeting real demand is continued investment through competitive electricity markets in new and existing power generation—guided by disciplined, transparent forecasting that reflects the projects most likely to materialize.
