Distributed Storage Costs Eight Times More to Operate Than Utility-Scale as UK Installer Data Shows Attach Rates Falling to 12.5 Percent

Vish Kulkarni of Madison Energy Infrastructure published operating-cost benchmarks this week that separate distributed storage from utility-scale by roughly a factor of eight: $30 to $50 per kilowatt-hour a year for distributed-generation projects, against $3.75 to $7.75 a year for utility-scale four-hour systems.

In the same week, POWER Magazine ran an analysis by Manan Shah of the UK installer Solar4Good, drawn from 2,610 installations completed since February 2021. Sorted by job value, storage attached to 91.3 percent of projects between £10,000 and £20,000 (1,123 of 1,230), 55.9 percent between £20,000 and £30,000, and 12.5 percent above £30,000 (2 of 16).

The two datasets sit on different continents and measure different things. Both land on the same segment of the market, and neither one measures a cell price.

The spread. Kulkarni’s benchmarks are per kilowatt-hour of installed capacity per year, so the gap compounds across a 20-year asset life rather than appearing once at commissioning. The op-ed also gives the warranty envelope that operating cost has to survive: modern LFP guarantees commonly run 20 to 25 years or 10,000 to 12,000 cycles, with annual degradation of 1 to 2.5 percent and standard utilization of 300 to 350 cycles a year. At 2 percent degradation, usable capacity lands near 65 percent of nameplate after 20 years.

A four-hour utility-scale plant spreads a service visit across tens of megawatt-hours of capacity. A 200-kilowatt-hour battery in a retail back room spreads the same visit across almost nothing. That arithmetic is interpretation rather than a line item the op-ed itemizes, but it is the only mechanism that plausibly produces an eight-fold divergence in a cost category dominated by labor and site access.

What LCOS omits. The same piece notes that demand-charge savings do not appear in standard levelized cost of storage math. Demand charges are billed against peak kilowatts rather than delivered kilowatt-hours, so a battery that shaves a handful of peak intervals a month produces savings the metric was never built to capture.

The standard measure therefore misses a major revenue line in a commercial building project while also understating the operating cost line in that same project. Both errors run in the direction of making distributed storage look more like utility-scale storage than it is.

Where the UK pipeline stalls. Shah locates the failure in capital approval rather than engineering, describing the mismatch as days of engineering against weeks of finance. His worked example is a care facility that required 14 months of operational data before it approved a phase-two battery.

The UK number carries a caveat. The attach-rate collapse from 91.3 percent to 12.5 percent is one installer’s measurement of one national market, with its own revenue mechanisms and its own grant history. What transfers is the shape of the failure: a capital-approval process whose duration scales with project value while the engineering work does not.

Hardware deflation stops before the installed price. Wood Mackenzie and SEIA’s Q2 2026 Solar Market Insight, released September 24, records commercial solar systems at $1.77 per watt, up 5.6 percent year over year, against residential at $3.36 per watt, down 1.4 percent, utility-scale fixed-tilt at $0.95 per watt, up 0.9 percent, and single-axis tracking at $1.06 per watt, up 2.0 percent. Distributed-generation modules fell 16 percent to $0.37 per watt over the same period.

Module costs moved in the opposite direction to installed costs, which places the increase in labor, interconnection and logistics. Those categories rose roughly 15 percent across all segments.

That is solar pricing, not storage pricing. It shares an installation labor pool, a freight network and a permitting queue with commercial battery projects, which makes it a reasonable read on the same cost stack. Commercial is the only segment in the dataset where installed price inflation outpaced both residential and utility-scale, and the increase sits outside the hardware.

The operating budget. A BOMA study released this week measures 35.4 billion square feet of office, retail and industrial property across 79 markets, generating $609.9 billion in annual economic output, $344.4 billion of GDP contribution and 3.9 million jobs. The operating-expenditure base itself is $274.9 billion a year.

That $274.9 billion is the budget line a storage service contract competes inside, and the one a demand-charge saving flows back into. It is also the pool from which a $30-to-$50-per-kilowatt-hour annual service cost gets paid, in a building whose facilities team already has a maintenance schedule and no line item for a battery.

What the evidence supports. Three observable facts now point the same direction: operating cost for distributed storage runs roughly eight times the utility-scale rate on a per-kilowatt-hour basis, installed-price inflation in the commercial segment is concentrated in labor and logistics, and attach rates in at least one installer’s national dataset fall by a factor of seven as project values reach commercial scale.

What follows is a modelling problem. Underwriting for building-sited storage typically treats capital cost as the hard number and service as a residual assumption. If Kulkarni’s range holds anywhere near the middle, service is a first-order term over a 20-year warranty, comparable in magnitude to the savings it is meant to protect. No distributed-storage operating-cost benchmark disaggregated by site type, access method or enclosure appears to have been published, which is the dataset that would settle the question.

Neither of the quantities these datasets measure moves when a cell price falls.

Sources