California residential rates now sit at nearly twice the national average. PJM wholesale electricity prices jumped 56% in a single year, reaching $80.5 billion. Low and middle-income families are feeling the squeeze.
Behind it all is a regulatory model designed for a different era. Utilities recover costs through rates and earn returns on capital investment. This is a model that worked when demand was predictable, but is now producing rate increases that regulators and customers are less and less willing to absorb.
Utilities are turning to flexible DERs to help. Here are six trends defining that movement in 2026.
1. Data centers are fast becoming the grid’s biggest test
Data centers are now the largest source of load growth for utilities across multiple US regions. Utilities have committed to more than 187GW of data center load, but research suggests that figure drops closer to 69GW when you apply a strict definition of “binding contracts”.
Even that pipeline is uncertain:
- Sightline Climate predicts that between 30 -50% of large data centers scheduled to come online in 2026 could be delayed, with power constraints and equipment shortages being the primary causes
- At least 25 projects were cancelled in 2025 due to local resistance, four times the number cancelled the previous year
In March 2026, the White House framed data center growth as an affordability issue, arguing that households should not bear the cost of infrastructure built to serve private AI and cloud demand.
2. Data center flexibility can become part of the solution
To counter these issues, data centers are starting to serve as grid assets. ACEEE research suggests that curtailing data center load for just 0.25% of their uptime could unlock enough capacity to serve 76 GW of new demand.
Companies like Google are already testing this, signing demand response agreements to shift workloads away from peak periods.
Regulators are catching up on this point too. For example, Texas SB6 requires ERCOT to develop a competitive reliability service procuring demand reductions from large loads (75 MW+) ahead of emergency events. The bill mandates that data centers disclose backup generation capacity that can be directed to curtail during grid stress.
3. Demand flexibility is scaling…
Data centers are the most visible example of a broader opportunity.
Only 6% of US energy consumers enrolled in a retail demand response program in 2024. However, ACEEE estimates 60-200 GW of load flexibility potential is available within the next decade. That exceeds forecasts for data center capacity growth.
Plus, as of 2025, 34 states have some form of virtual power plant or demand flexibility program in place. The DOE estimates the US could reach 80-60 GW of VPP capacity by 2030 (although deployment needs to scale to hit this target). If it does, the DOE predicts that could save around $10 billion annually in grid costs.
Demand-side resources can support the grid in months at a fraction of the cost of new infrastructure. This makes them one of the few solutions capable of responding to pressures from near-term load growth.
4. …and policy is shifting toward DER orchestration
The big constraint slowing this opportunity is how demand response and VPPs are valued.
States are beginning to shift toward DER orchestration, building the frameworks needed to dispatch flexibility at scale:
- California’s CPUC is designing a DSO-led orchestration framework to enable DER dispatch at scale
- Ameren Illinois is developing flexible interconnection and DER orchestration models across multiple value streams
But utilities should factor in an economic layer to direct DER use where it delivers the most value vs. traditional infrastructure. Without that lens, the whole system value that DERs can offer won’t be measured or captured.
5. There’s an unresolved DER ownership debate
Alongside the growth of VPPs sits a harder question: who should own and control DER flexibility? For public-owned utilities, this is less of an issue as many already source, own, and operate DERs as grid or customer assets. This issue is unresolved for investor-owned utilities
In Minnesota, Xcel Energy’s Capacity*Connect program – the first utility-owned VPP of its kind in the US – was approved at around $2,150 per kilowatt. By comparison, an equivalent Xcel aggregator-led program in Colorado costs closer to $625 per kilowatt.
This reflects a regulatory tension. Investor-owned utilities can earn regulated returns on capital investments. This creates stronger incentives for utility-owned infrastructure than for third-party aggregation models.
That tension is shaping policy:
- In Michigan and New York, proposed VPP bills prioritize third-party access and exclude utility-owned DERs
- In California, a 1 GW customer-led program is under budget pressure despite analysis suggesting it could save $550 million annually by 2035
- “Bring your own capacity” models, where data centers fund local DER investment as a condition of connection, are gaining traction as a response to cost and community concerns, with the first launched in 2025
6. Grid utilization is under the spotlight
The US grid was built to serve peak demand. For the rest of the time, significant capacity sits idle. The Brattle Group estimates that better utilization of existing infrastructure could save ratepayers tens of billions of dollars in its “The Untapped Grid” report.
Virginia has become an early test bed, with new legislation requiring utilities to measure and report grid utilization.
Utilization is a good starting point, but not the end goal. More utilization does not automatically mean lower costs. In some cases, pushing more through the system increases losses and raises the cost to run it.
The question is whether flexibility is being deployed in the right places, at the right times, and valued accurately enough to compete with traditional infrastructure on equal terms.
Until utilities have a granular view of where their grid is constrained, where DERs are located, and what those assets can actually deliver at a specific location and time, that comparison can’t be made with confidence.
The opportunity to make better use of the grid will remain only partly realized until the data and valuation frameworks catch up.
