During the last week of January, the PJM Interconnection was operating in the immediate aftermath of the January 23–27 winter storm Fern, which occurred within the broader January–February 2026 North American cold wave. The storm brought widespread snowfall and prolonged below average temperatures across much of PJM’s area, affecting the Midwest, Ohio Valley, and Mid‑Atlantic regions. Major load centers, including Chicago (ComEd zone), Columbus, Pittsburgh, Philadelphia, Baltimore, and Washington, D.C., experienced heavy snow accumulations and sustained cold conditions that persisted well after the storm’s peak. For power markets, this was a classic stress test: prolonged cold, rising net demand, and widening exposure for anyone leaning too heavily on ISO forecasts.
From an operational standpoint, severe cold creates additional stress on generation services, and transmission systems. While PJM maintained heightened situational awareness throughout the last week of January, forecast accuracy — not just operational readiness — became the key differentiator during this event.
PJM total system demand increased as expected. Between Monday, January 25 and Friday, January 30, PJM recorded some of the highest load values of the winter. Demand remained elevated on Monday, though still below 128 GW, but ramped significantly on Tuesday, January 26, reaching an evening peak of 134 GW. The following days saw continued increases, with morning peaks climbing above 136 GW on both Tuesday and Wednesday. On Thursday, PJM registered its highest load of the current winter: 139.047 GW at 8 AM HE. Demand eased slightly on Friday (still exceeding 138 GW in the morning) before beginning a gradual decline into the weekend. For context, PJM’s all‑time winter peak remains 143,700 MW, set on January 22, 2025.
During this volatile period, our day‑ahead load forecast outperformed PJM’s. PJM’s daily forecasts consistently came in much higher than expected, while our model remained more conservative, and substantially closer to the actual demand (see Figure 1). In Figure 2, the blue bars illustrate PJM’s forecast bias; the green bars represent PRT and are tighter, leaner, and far more accurate.
Quantitatively, our day‑ahead MAPE was 1.83%, compared to 3.89% for the ISO. During the On‑Peak period, our model delivered 1.77% MAPE, while the ISO posted 4.08%. For market participants, this forecast accuracy directly translated into better positioning and reduced exposure. The Western Hub (WHUB) printed a +$242.82 DA/RT spread from Monday through Friday. Trusting the PRT curve would have significantly reduced exposure to oversized day‑ahead premiums. In this setup, leaning into our load forecast was not only the more accurate call, but it was the best trade.
In summary, the last week of January in PJM was shaped by the lingering impacts of a major winter storm. The result was a much higher demand and persistent operational strain. In this scenario, our load forecast proved exceptionally valuable for customers, offering both better accuracy and better market outcomes. As winter volatility persists, events like this reinforce that conservative and accurate net demand forecasting becomes a competitive advantage in high stress conditions.
