Finally, the right time for Distributed Energy Resources

The distinction that carries the piece sits inside the category itself. “Distributed energy resource” was always a deliberately broad tent. Lubershane says he does not know where the term originated, but recalls it being promoted in the early 2010s because the industry needed an umbrella for a set of new products and their vendors wanted to be seen as resources rather than nuisances, which is how rooftop solar was starting to be perceived, in many cases rightly. Only a subset, though, can be switched on and off on demand. Because supply and demand for electricity must balance continuously, what a grid operator actually values is a button: distributed generation, distributed storage and flexible load can be dispatched, while efficiency and rooftop solar can only be planned around. Lubershane proposes calling the dispatchable subset distributed capacity resources, notes the term is not yet in regular use, and credits Sparkfund, a portfolio company at his firm, with pushing it. A virtual power plant is what you get when many of those are aggregated with software and a single virtual button.

His account of the failed decade has two causes, and both are needed. The first is demand. He describes asking a group of utility engineers where distributed batteries could cost-effectively substitute for grid investment and being told there were practically no such places; he came to accept they had searched in good faith, because there were so few places where demand was growing fast enough to necessitate big, costly upgrades, and capacity markets across North America were mostly satisfied by existing large-scale resources. The second is cost, and the point is counterintuitive: solar panels, batteries and connected devices all got steadily cheaper, but the soft costs did not follow. Selling unfamiliar, expensive equipment to customers who did not need it stayed expensive once the early adopters ran out, and signing customers up for a utility program that might interrupt their home or business made it harder still. Nor was dispatching them collectively simple. He is careful about why: incompatible vendor interfaces and patchwork utility systems were real speedbumps but surmountable, and the intractable obstacles were commercial, regulatory and organizational, including unresolved questions about where DER management sits inside a utility, whether aggregators should bid into wholesale markets or stay on the demand side, and how to avoid counting the same resource twice.

Both conditions have reversed over roughly two years. On demand, he points to what he has elsewhere called the electricity gauntlet: transmission construction slowing, prices for transformers, switchgear and lines rising, local opposition to wind and solar spreading, and AI load that planners cannot size. He does not re-derive that evidence here, referring back to his own earlier posts for each piece of it. The consequence is that electricity prices are rising, which means savings that were not worth chasing five years ago now are. His strongest evidence is utilization rather than deployment, and it is the one number in the piece that behaves like a trend break: at the demand response aggregator Voltus, capacity under management roughly doubled since the summer of 2023 while the number of times per year grid operators actually called on those resources rose more than sixfold. Historically, he says, these resources were often a nice-to-have, demanded as much by regulatory fiat as by real system value, and this is the best signal he has encountered that the position has changed.

On cost the picture is mixed, and he says so. Residential solar costs including soft costs have drifted upward. Batteries are the category that matters, because they can be dispatched instantly, arbitrage cheap power against expensive power and do it daily without the site host noticing, and the cheap electric vehicle battery packs took years to show up in installed household system prices. That has now started, and the mechanism he identifies is circular: falling prices draw more customers, more customers spread acquisition and installation costs further, which lowers prices again, and once enough batteries exist to aggregate, the aggregator can hand some grid-services revenue back as a discount. The evidence that the loop is turning is scale: Sunrun dispatched its California fleet in concert with several other aggregators on a hot summer evening, and the combined dispatch supplied 535 megawatts when the system operator most needed it. At last, he says, there is an aggregation that merits the title, because it has reached the scale of an actual power plant. His closing bet is that economies of scale and learning curves are the two most reliable forces in a modern economy and that virtual power plants are finally tapping both, which is why he is cautiously optimistic that nearly every utility will have one by the end of the decade.