Your Grid May Have Found Its Next Power Plant: In Your Garage

The next power plant feeding your grid may not have a smokestack or a turbine hall. It may be hanging on the wall in your garage — and in a few hundred garages down the street.

Home batteries are usually sold as insurance against an outage. They charge, keep a house running when power is not available and store electricity from rooftop solar.

But link enough of them together, and the same devices can begin to behave like a utility-scale plant: taking power off the grid when supply is plentiful, then returning it during the evening peak, a heat wave or an emergency.

Texas is now moving closer to treating those batteries that way.

At a meeting on Aug. 17 of ERCOT‘s Demand Side Working Group, staff proposed a fourth phase of ERCOT’s Aggregate Distributed Energy Resource pilot. The new category has a dense name — Aggregate Distributed Energy Storage Resource, or ADESR — but the idea is simple. Hundreds or thousands of household batteries could be registered, modeled and dispatched as one grid resource.

The discussion now at ERCOT showed why the proposal is arriving now. The pilot has grown to 249 megawatts of qualified capacity, nearly half of its 500-megawatt energy limit, across nine qualified aggregations. The growth has been driven largely by residential batteries.

The current model, however, treats an aggregation as a load resource and dispatches it using broad load-zone assumptions. ERCOT cannot see its state of charge, and the model does not fully represent its ability to withdraw power from the grid and inject it back.

Phase 4 would make that resource more visible and more precise. Batteries grouped under the same substation transformer would be mapped to the corresponding point in ERCOT’s network model. A qualified scheduling entity would register the aggregation.

The local distribution utility would verify its location. Device-level data would be rolled up into a single telemetry stream, and ERCOT’s dispatch software could send the fleet a target based on the grid conditions at that location.

In other words, ERCOT would not need to know what every battery in every garage was doing. It would need to know what the fleet was capable of doing, where it was connected and whether it followed instructions.

That last part is no longer theoretical. Base Power, an Austin-based startup that installs large home batteries at little upfront cost in exchange for multiyear retail-electricity contracts and then dispatches the aggregated batteries into ERCOT’s markets, said it had added approximately 250 megawatts over the past year and expected more than 100 megawatts to be approved for pilot participation soon. “They follow set points within acceptable Unit Dispatch System Performance tolerance bands and behave like a resource,” a Base Power representative told the working group at the Aug. 17 meeting.

The proposed framework is deliberately incomplete. ERCOT wants to dispatch the batteries using nodal shift factors — the more precise signal used for resources at specific grid locations — while initially settling them at the broader load-zone price. 

The new aggregations would not be allowed to provide ancillary services at first, even though existing ADERs can qualify for some of them. ERCOT described that restriction as a way to reduce complexity, not a permanent judgment about what the batteries can do.

“The proposal is to start simple so we can move quickly,” an ERCOT staffer said during the committee’s meeting.

The initial stage would require no new metering and could begin after approval of the pilot amendment. Nodal dispatch and nodal settlement — sending the fleet instructions and paying it based on conditions at its specific location on the grid, rather than at a broad load-zone average — would require coordination with distribution utilities, possible changes to load-zone price calculations and a regulatory review. ERCOT is targeting a board decision in December.

Metering may become the fault line. “If it is not practical to meter at the battery, does the entire construct fail?” an engineer from Oncor asked during the discussion.

“No, not for the initial stage, which uses nodal dispatch and zonal settlement,” an ERCOT staffer replied.

That answer captures the pilot’s strategy: prove that the fleet can act like a plant before solving every question about how to pay it like one.

The Rocky Mountain Institute has argued that the rest of the power industry needs to catch up. In its report, RMI describes virtual power plants (VPPs) as aggregations of batteries, electric vehicles, thermostats and other connected devices that can provide “utility-scale and utility-grade services.”

The scale is already comparable to conventional generation. RMI noted that an average U.S. combustion turbine was 180 megawatts in 2024, while several VPP programs have met or exceeded that figure.

Those resources have also performed when the grid was under stress. During an eastern U.S. heat dome, Sunrun, the largest U.S. residential solar and battery company, dispatched more than 340 megawatts of customer batteries, while EnergyHub, a Brooklyn-based software firm that helps utilities manage distributed energy resources, shed 900 megawatts of peak demand and shifted 3.5 gigawatt-hours to other hours. A California test simultaneously discharged behind-the-meter batteries across the state and flattened the evening net peak.

The problem, RMI found, is that utility plans often mention VPPs without giving them the same modeling treatment as power plants. Conventional options arrive in planning models with costs, reliability contributions, operating limits and build schedules. VPPs are frequently treated as fixed reductions to a load forecast, which prevents the model from selecting more of them when they are the least-cost option.

RMI said they should be “evaluated on an even playing field against conventional resources.” That means counting not only peak relief, but also avoided generation, deferred transmission and distribution upgrades, local capacity, frequency regulation and resilience during extreme weather.

That is also the policy problem the Energy Policy Design Institute, a Denver-based nonprofit focused on energy policy, described to MISO — the Midcontinent Independent System Operator, the grid operator that runs the high-voltage system and wholesale power market across 15 states in the central U.S. — earlier this year.

The institute’s VPP Convergence Project, developed with NARUC, RMI, SEPA and others, laid out a shared vision in which dispatchable distributed resources are paid for supporting reliability, affordability, resilience and clean energy. It paired those goals with grid services including local capacity, peak shaving, backup power and renewable integration.

The institute’s message was that the technology is only one building block. Regulators and grid operators also need rules for registering devices and aggregators, interconnection, data access, cybersecurity, operational coordination, communications, dispatch overrides, metering, settlement and dual participation in retail programs and wholesale markets.

It proposed policy-readiness levels and a set of tools — a policy map, a common dictionary, a design-questions library and a regulatory playbook — to help states move from pilots to durable programs.

ERCOT’s proposal is that framework arriving in operational form. It is deciding what must be registered, who verifies the batteries, how the fleet communicates, where it appears in the network model and how it gets paid. It is also preserving room to learn. Mixed fleets of batteries and thermostats can remain in the existing program; the new ADESR category would be reserved for batteries that meet tighter location rules.

The payoff is not merely more megawatts. It is megawatts in the right place. A battery fleet under a congested substation can create room for new demand, reduce stress on transmission and potentially postpone an expensive upgrade. The wires may not care whether the relief came from one large battery or 600 small ones.

Base Power told ERCOT the scale could move quickly. “Today, we are discussing hundreds of megawatts; in the future, the locational value of aggregated distributed resources could reach the gigawatt scale.”

That is the larger bet behind ERCOT’s pilot. The hardware is already being installed for household backup. The grid’s next job is to recognize that, at the right moment and under the right rules, those batteries are not just appliances. For the people who own them, they remain backup power. Together, they are infrastructure.