The headline number is designed to stop you cold: $6 billion. That is what the Solar Energy Industries Association says solar generation and battery storage saved wholesale electricity buyers in California and Texas during a single summer — roughly $2.5 billion in the California Independent System Operator’s territory and $3.55 billion in the Electric Reliability Council of Texas — comparing June through August 2026 with the same three months in 2021. The analysis, released October 22, arrives as affordability dominates the political conversation ahead of the midterm elections, and its sponsor is not a neutral party: SEIA is the trade association for the solar and storage industries, and its chief executive, Tim Pawlenty, framed the findings explicitly as proof that solar and storage growth is driving down prices in red and blue states alike.
Strip away the framing and the underlying result is still worth taking seriously, because it is built on market data rather than modeling. SEIA compared weighted-average hourly day-ahead prices — locational marginal prices averaged across hubs and load zones and weighted by generation in each hour — for the two summers, with 2021 prices adjusted for inflation into 2026 dollars. The underlying hourly generation and price data come from the Energy Information Administration’s Hourly Electric Grid Monitor and GridStatus.io, pulled in September 2026. This is not a simulation of what prices would have been without solar and storage. It is a comparison of what prices were, in two summers five years apart, across a period in which the two fleets transformed.
Two markets, one direction
Start with the California numbers, because they are the more striking on their face. Average wholesale prices in CAISO fell 51 percent, from $77.54/MWh in summer 2021 to $38.32/MWh in summer 2026, even as electricity consumption rose 2.5 percent. The peak hourly price fell further still, from $141.75/MWh to $59.69/MWh. Over the same five years, solar generation in California rose 90 percent, and installed battery storage grew from 8.1 GWh to 62.0 GWh through the first half of 2026. Batteries now supply 6 to 7 percent of daily energy use in CAISO — a share that would have sounded fanciful in 2021, when the fleet was barely out of its pilot phase.
Texas tells the same story with a compressed timeline and a much bigger demand shock. ERCOT’s average wholesale price fell 42 percent, from $54.97/MWh to $31.77/MWh, while total load surged 30.1 percent — the kind of demand growth that, in the old textbook, should have pushed prices up, not down. The peak hourly price fell from $110.96/MWh to $57.30/MWh. Solar generation rose 421 percent, and battery storage expanded more than 26-fold, from 1.0 GWh to 33.1 GWh. Storage now provides more than 1.3 percent of the state’s energy — small in absolute terms, but the marginal hours are where prices are set, and storage’s contribution is concentrated exactly there.
The mechanism SEIA describes is the one this series has traced before: solar and storage increased the number of hours in which they displaced expensive peaker generation, and the number of hours in which they set the market-clearing price themselves. Midday solar floods the market and pushes the marginal unit down the cost stack; evening batteries discharge into the ramp and bid against the gas turbines that used to set those prices. It is the same dynamic behind the October story this desk published last week, when California absorbed a near-summer peak with barely a price ripple — but this is a different measurement. That piece examined real-time price formation on individual days. This analysis measures day-ahead seasonal averages, the prices at which most energy is actually scheduled and settled. The two lenses agree, which is itself informative: the flattening shows up whether you look at the extreme hours or the average ones.
The peak-hour figures are the cleanest evidence for the mechanism. Averages can hide a multitude of stories — a mild summer, a favorable gas market, a few very cheap weeks dragging the mean down. Peak prices are harder to explain away, because they are set in the tightest hours, when the system leans hardest on its most expensive capacity. In both markets the summer peak price fell by more than half: from $141.75 to $59.69/MWh in CAISO, from $110.96 to $57.30/MWh in ERCOT. That is exactly what a fleet built to shave peaks should do.
It is also worth noting what the analysis does not measure. Day-ahead energy is only one revenue stream and one cost component. It excludes capacity payments in CAISO’s resource adequacy framework, ancillary services in both markets, and the uplift and congestion charges that show up elsewhere in settlement. A complete accounting of what solar and storage did to total wholesale costs would need those pieces too — and in ERCOT, where scarcity pricing substitutes for a capacity market, the energy price carries weight that CAISO spreads across multiple products. The $6 billion is a day-ahead energy figure, not a total system cost figure, and it should be quoted that way.
Why two different designs produced a similar result
The more interesting question is why the result looks so similar in two markets designed to do different things. CAISO operates with a resource adequacy framework and a $2,000/MWh energy offer cap; ERCOT runs the energy-only model, relying on scarcity pricing through its operating reserve demand curve to pay for reliability. In the usual telling, these are opposite philosophies. Yet both saw average wholesale prices fall by roughly half over five years of solar and storage growth.
The common variable is the shape of the net load curve, not the market design. In both systems, the expensive hours are the same hours: late afternoon and early evening, when solar fades and demand stays high. Solar attacks the first half of the problem by covering more of the daytime, and storage attacks the second half by moving midday energy into the evening. Neither technology cares whether the market pays for capacity separately or through scarcity prices. Both reduce the hours in which a gas peaker is the marginal unit, and that is where the price effect lives.
What the $6 billion does not prove
Here is where the honest reading has to slow down, because the $6 billion figure invites a stronger conclusion than the data supports. First, the comparison spans a 2021 baseline shaped by conditions that no longer hold. Summer 2021 was a different gas market, a different scarcity environment, and a different fleet — California was still absorbing the lessons of the August 2020 heat storm, and Texas was months past the February 2021 winter storm that had reshaped its market’s risk perceptions. Some portion of the price decline between the two summers reflects those baseline conditions normalizing, not solar and storage displacing peakers. SEIA’s analysis does not decompose the decline into its causes, and no honest reading should attribute all of it to the two technologies.
Second, the savings figure is an accounting of the price difference applied to volumes, not a settlement record. It tells you wholesale buyers paid less per megawatt-hour in summer 2026 than in summer 2021, multiplied across the energy traded. It does not tell you who captured the difference, or how much flowed to any particular customer class. In markets where most load is hedged forward — and both California utilities and Texas retailers hedge heavily — the day-ahead averages SEIA measured are not the prices most customers actually paid. The hedged cost, built from forward purchases made months or years earlier, moves much more slowly than the day-ahead average. The savings are real at the market level and diffuse at the customer level, arriving with the lag hedging always imposes.
Third, and most importantly, wholesale prices are not retail bills. SEIA says so itself, and its own retail breakdown is the most sobering part of the analysis. In California, the generation portion of residential bills fell 10 percent between the two summers — but transmission costs rose 12 percent, distribution costs rose 51 percent, and other costs, including wildfire management, rose 90 percent. The net effect, per SEIA’s data from the Heatmap project, was a 22 percent increase in overall residential bill prices. In Texas, generation costs fell 4 percent, while transmission rose 17 percent and distribution rose 23 percent, for a net overall increase of 3 percent. The wholesale savings were real, and bills still went up, because the wires and the wildfire ledger moved in the opposite direction.
The retail gap is the story, not a footnote
That gap deserves more attention than it usually gets, because it explains the political paradox of the moment: wholesale markets delivering their cheapest summer in years while affordability dominates the election. The generation commodity — the part of the bill that solar and storage actually touch — is typically less than half of a residential bill. The rest is the cost of the network and the policy obligations layered onto it, recovered slowly through regulated rates. When distribution costs rise 51 percent in five years in California, no plausible decline in the energy commodity offsets it.
This is not an argument that the wholesale result doesn’t matter. A 10 percent decline in the generation portion of California bills and a 4 percent decline in Texas are meaningful, and they would have been larger declines in the other direction had the solar and storage fleets not been built — the counterfactual in which 30 percent more Texas load met a 2021-vintage fleet is not a pleasant one to contemplate. But it is an argument about where the affordability problem now lives. The binding constraint on household bills in both states is no longer the marginal megawatt-hour. It is the wires, the wildfire costs, and the capital programs that regulators approve in rate cases. Anyone promising that more solar and storage alone will lower your bill is selling a wholesale solution to a retail problem.
There is a second, quieter detail in SEIA’s data that points in the same direction. Rooftop solar generation rose 172 percent in Texas and 107 percent in California over the five years, and SEIA notes that customer-sited solar and batteries reduce stress on the grid and offset the need for centralized generation assets. That is the distributed half of the story, and it matters for the distribution side of the bill — the very side rising fastest. But the analysis does not quantify how much distribution investment rooftop solar avoided, and that avoided-cost question is precisely where the next affordability debate will be fought: whether distributed resources defer wires spending or merely shift its recovery onto fewer customers.
What to watch from here
Three things will determine whether this summer’s result repeats or fades. The first is gas prices. If the 2021-to-2026 comparison flattered solar and storage by starting from an expensive baseline, then a sustained period of low gas prices would test how much of the decline was structural and how much was cyclical. The second is the storage revenue question this desk has raised before: as fleets grow, the spreads that pay for batteries compress, and the next wave of projects will need capacity accreditation, ancillary revenues, or nodal price separation to pencil out. The third is load growth, above all in Texas, where a 30 percent demand surge was absorbed this summer without a price spike — but where the next 30 percent will arrive on top of a fleet that has already harvested the easiest price-suppression hours.
For market designers, the SEIA analysis is best read as a progress report with a warning label attached. It confirms that solar and storage, at scale, do what their advocates claimed: they push down wholesale prices, including in the hours that used to be expensive, and they do it in very different market designs. It also confirms, in SEIA’s own numbers, that wholesale success and retail affordability are now different problems, governed by different cost drivers. The $6 billion is real. It is also, for most households, somewhere else — upstream of the bill, behind the hedge ladder, and dwarfed by the wires. That is not a reason to dismiss it. It is a reason to be precise about what it is.
One final caveat belongs in any honest treatment of this analysis, and it concerns the source. SEIA represents the industries whose products the analysis credits. That does not invalidate the work — the price data come from public grid-operator and EIA sources, and the methodology is disclosed in enough detail to be checked — but it does mean the analysis is advocacy-adjacent, not audited market settlement data, and it was released into an election-season affordability debate by an association with a direct stake in the outcome. The right response is not to discount the numbers but to hold them to the standard this piece has tried to apply: attribute them to SEIA, note what they measure and what they leave out, and let the retail data in SEIA’s own report supply the counterweight.