Across 12 utility clients, 60+ events, and 770,000 thermostats in summer 2026, EnergyHub’s Dynamic Load Shaping (DLS) increased delivered capacity by an average of 10-20% during the second half of events. By sustaining load reduction across the duration of an event, DLS unlocks new opportunities to maximize the grid value of VPPs.

These results came during another demanding summer for the grid. NOAA recorded 2026 as the warmest summer since national records began in 1895, with temperatures averaging 3°F above the 20th-century average. Intense heat and long, uncertain peak windows gave utilities a meaningful test of flexible load performance across multi-hour events.

Consistent load reduction hour after hour is becoming increasingly important as peak windows grow longer and less predictable. Grid operators and resource management teams depend on this consistency for day-ahead resource planning, especially as VPPs scale. These teams also need resources that can sustain performance across multiple hours and return to baseline without creating unintended secondary peaks due to the snapback effect from hundreds of thousands of devices. Doing so requires a combination of sufficient flexible capacity available to meet grid needs, accurate forecasting of when and how that capacity can be delivered, and precise control over how it is dispatched.

Standard DR remains effective when a utility knows when the peak will hit and wants to maximize load reduction immediately in that hour. DLS is more effective when the peak window is wider or less certain, or when preconditioning and recovery load spikes create additional grid impacts.

From pilot to proven performance

Our Dynamic Load Shaping pilot demonstrated up to 77% less hour-over-hour variability in load reduction, validating that DLS delivers consistent performance across multi-hour events. Building on these results, EnergyHub made DLS available to all utility clients on the platform in 2026.

DLS uses a machine-learning-driven optimization engine informed by approximately 3 million devices to automatically stagger device dispatch and deliver consistent load reduction across the entire event window. In standard demand response, all devices are dispatched at once and load shed tapers over the course of an event. DLS instead evaluates thousands of possible dispatch schedules to determine how flexible capacity should be optimized to meet various objectives. 

At its core, DLS optimizes for maximum total load reduction and hour-over-hour consistency, delivering 85–95% of the load reduction of a standard DR event. Utilities can also configure DLS to smooth preconditioning and post-event recovery. Customer protections can then be layered on top, including avoiding dispatch during time-of-use periods or limiting an individual device’s participation to a shorter window than the overall peak period — for example, three hours within a four-hour event.

Fig. 1: A DLS event delivers consistent hour-over-hour load shed across a four-hour event.

How DLS helps utilities maintain consistent performance during the hours that matter most

This summer, utilities used DLS to optimize for consistent load reduction for up to four hours in order to address uncertain peak timing, improve worst-hour performance, reduce exposure to demand charges, and manage post-event recovery at scale.

New England IOU overcomes uncertain peak timing with 97% load shed consistency

One New England IOU delivered 97% load shed consistency across a three-hour DLS event this summer while maintaining 86% of the average load shed of a comparable standard DR event.

As peak windows grow wider, utilities have less certainty about which hour will ultimately become the coincident peak. Because standard DR load shed typically declines 20–30% hour-over-hour, utilities can miss out on grid value when the coincident peak falls in hour two or three.

DLS helps utilities maximize grid value even when the exact timing of the coincident peak is uncertain. By sustaining more consistent load reduction across the event window, DLS helps preserve capacity if the peak arrives later than expected.

Fig. 2: New England IOU maintains 97% load shed consistency across a three-hour event.

Midwestern IOU improves worst-hour performance by 55.2%

During a DLS event on July 17, this Midwestern IOU’s minimum hourly load shed reached 10.77 MW — 55.2% higher than the 6.95 MW minimum delivered during a standard DR event on July 14.

In MISO, where this utility operates, demand-side resources are accredited based on the minimum load shed they expect to deliver across a four-hour demand response event. As a result, declining hour-over-hour performance can force utilities to be conservative in the capacity they nominate in the market to avoid penalties. By raising the performance floor, DLS can support larger, more confident capacity nominations.

Southeastern co-op maintains >90% load shed consistency over four hours

One Southeastern cooperative maintained more than 90% load shed consistency across a four-hour peak window this summer, while still achieving 0.9 kW of average load shed per device.

For distribution cooperatives and municipal utilities, generation and transmission (G&T) demand charges are assessed based on a utility’s demand during the coincident system peak. Utilities may know the broader peak window without knowing which hour will ultimately become the coincident peak. If load reduction has already declined by the time that hour arrives, underperformance can result in higher demand charges for the utility and its members.

By maintaining a flatter load-reduction profile across the full peak window, DLS helps co-ops and municipal utilities reduce the risk that a later coincident peak results in higher G&T costs for the utility and its members.

Fig. 3: Southeastern co-op achieves 92% load shed consistency across a four-hour DLS event.

Canadian ISO reduces post-event snapback by 99% across 300,000+ devices

This Canadian ISO used DLS to reduce post-event snapback by 99% across more than 300,000 participating devices. The same event also delivered 91% hour-over-hour load shed consistency and 780 MWh of total energy shifted, while maintaining 94% of the average load shed of a standard DR event.

As VPPs scale, post-event recovery becomes part of the load-shaping challenge. If hundreds of thousands of devices increase their load at roughly the same time, they can create an unintended secondary peak. By staggering device transitions during recovery, DLS smooths the return to baseline and mitigates snapback.

Fig. 4: Canadian ISO reduces post-event snapback by 99% with DLS.

What comes next for DLS

Taken together, these results show how more consistent load reduction can create value across very different grid and market conditions — from preserving capacity through uncertain coincident peak windows to improving worst-hour performance and smoothing recovery.

In 2027, EnergyHub plans to extend DLS to batteries, enabling more consistent battery discharge across events of up to four hours and the ability to co-optimize battery dispatch alongside thermostats during DLS events. Combining the complementary flexibility of multiple DER types will enable utilities to deliver increasingly targeted, configurable load shapes.

See how Dynamic Load Shaping can perform in your portfolio. Contact EnergyHub to start a conversation.

200+ utilities | 3M devices | 4 GW dispatchable flexibility