Why four-hour storage is the wrong default for Texas

Where the default came from

Four hours was a capacity-accreditation artifact from markets that predate today's price shapes. In ERCOT, scarcity arrives in narrow, volatile windows, and the revenue is captured by assets that can cycle into them repeatedly — not by duration sitting idle on the balance sheet.

The default survived because it was easy to underwrite: a fixed duration maps neatly onto a capacity payment. But ERCOT is an energy-only market. There is no capacity payment to map onto, and the revenue stack that actually exists rewards different hardware than the one the industry keeps buying.

What the price data says

Look at the distribution of spreads rather than the average day: most of the year's storage revenue in ERCOT concentrates into a small number of hours, and those hours cluster in short, sharp events. An asset that can cycle twice into a volatile evening frequently beats a longer-duration system that committed its state of charge too early.

Sizing for the grid you have

Our systems ship in two- to eight-hour configurations, and the honest answer to "how much duration" is a modelling exercise against the node you sit on. Buying duration you cannot monetise is the most expensive way to feel safe.

When longer duration does win

None of this is an argument that four or eight hours is never right. Transmission-constrained load pockets, resiliency contracts, and certain hedges genuinely need it. The argument is against defaults: duration is a design variable, and treating it as a constant quietly hands the decision to a market that no longer exists.