Data point
US Virtual Power Plant Capacity: Current vs. 2030 Potential (GW)
Estimated aggregated capacity of distributed energy resources (rooftop solar, batteries, smart thermostats, EV chargers) coordinated as virtual power plants in the United States, current baseline versus modeled 2030 potential, in gigawatts.
Source: U.S. Department of Energy, Virtual Power Plants: Pathways to Commercial Liftoff (2023) - DOE reports current US VPP capacity as roughly 30 to 60 GW, and models growth to 80 to 160 GW by 2030. The chart uses the low end of the current range and both ends of the 2030 range to show the modeled spread rather than inventing a single midpoint.
The decentralization pitch, restated for 2026
The cleantech decentralization thesis holds that value shifts from large centralized power plants toward millions of small distributed assets, rooftop solar, home batteries, EV chargers, smart thermostats, coordinated in aggregate as virtual power plants (VPPs). The pitch is that this coordination layer, not the underlying hardware, becomes the durable business, because it captures data, dispatch rights and customer relationships across an asset base that keeps growing on its own.
The US Department of Energy’s 2023 Liftoff report on VPPs put real numbers behind this narrative. It estimated the country’s existing VPP capacity at roughly 30 to 60 gigawatts, built mostly from legacy demand-response programs, and modeled that this could grow to 80 to 160 gigawatts by 2030 if enrollment and market rules keep pace. DOE also argued VPPs are one of the cheapest ways to meet peak demand growth, undercutting new gas peaker plants on cost in many regions.
That is a real, source-backed number, not a talking point. It is also a wide range, 30 to 60 GW today and 80 to 160 GW by 2030, which is itself informative: the uncertainty band is roughly as large as the central estimate. A thesis that leans on this data needs to hold two things at once, the direction is credible, and the precision is low.
Where the thesis gets tested
Three things determine whether decentralization actually captures value versus just adding capacity to someone else’s balance sheet.
Interconnection and enrollment friction. VPP capacity only counts if devices are actually enrolled, metered, and dispatchable. FERC Order 2222, finalized in 2020, was meant to let DERs participate directly in wholesale markets, but implementation is grid-operator by grid-operator and has moved slowly in several regions. A capacity estimate is not the same as capacity that clears a market.
Who owns the coordination layer. If utilities, not independent aggregators, end up running the VPP programs, the “software captures the margin” argument weakens considerably, since the incumbent keeps the customer relationship and the aggregator becomes a vendor rather than a platform.
Reliability under stress. VPPs are attractive because they are cheap during normal peak events. Their value during genuine grid emergencies, extended heat waves or cold snaps where consumer devices are needed for many consecutive hours, is less proven at the scale DOE is modeling for 2030.
Countercase
The DOE figures come from a single US government report, not an independent audit, and DOE has an institutional interest in showing VPPs as commercially ready given the Liftoff series exists partly to attract private capital into a policy priority. The 30-to-160 GW range is wide enough that a reader could select either end to support almost any thesis about pace. This note also cannot verify current enrollment numbers against the DOE modeling assumptions, meaning the 2026 real-world trajectory relative to the 2030 target is not independently confirmed here. The thesis further assumes US market structure and FERC rules generalize to other geographies, which is not established. Finally, “decentralization creates a durable moat” is a narrative claim; the data here supports growth in aggregated capacity, not that any particular aggregator, utility or platform captures lasting economic value from that growth.