The five-year storage average stopped describing the system you're trading in.

The five-year storage average is the most-cited number in weekly gas commentary. It also stopped describing the system it was supposed to help you evaluate.

As of July 3, 2026, working gas in storage sat at 2,983 Bcf, 185 Bcf (6.6%) above the five-year average (currently defined as 2021–2025). The window itself is dominated by the export era. U.S. LNG exports averaged effectively zero in 2015, around 10.6 Bcf/d in 2022, near 15 Bcf/d in 2025, and are forecast above 18 Bcf/d by 2027. The benchmark has largely assimilated the regime change it was supposed to help evaluate. "Average" used to mean normal for a pre-LNG system. It now means normal for a structurally tighter one.

The architecture problem cuts the other way. The benchmark treats every Bcf as equivalent. The system doesn't. Depleted oil-and-gas reservoirs hold roughly 80% of total working gas capacity but cycle slowly. Salt caverns hold on the order of 7% of capacity but provide about 23% of daily deliverability, per EIA's long-standing figures. LNG feedgas swings, intraday power-burn spikes, and basis volatility around Gulf liquefaction need fast-cycle deliverability, which lives almost entirely in salt. A Bcf in an Illinois aquifer and a Bcf in a Mont Belvieu salt dome aren't the same Bcf.

Capital is starting to vote on the deliverability side. The first material salt storage build-out in roughly fifteen years is underway in the Gulf Coast and East Texas corridor. Trinity Gas Storage Phase II in East Texas adds at least 13 Bcf, targeted for service August 1; Golden Triangle Storage in Beaumont is building about 14 Bcf more toward 2027, roughly 90% contracted at sanction. Both LNG- and power-adjacent. Roughly 27 Bcf combined won't move the capacity headline, but it will move the deliverability mix.

What replaces the five-year average? Trailing three-year for current-regime normalcy. Pre-2022 baseline for showing how much tighter the system runs. Days-of-forward-demand for adequacy: 2016's record peak covered on the order of 40–45 days of winter demand and exports; the same inventory today would cover fewer days because the denominator grew, which sets where Henry Hub trades to defend winter paths. Salt-cavern-weighted deliverability for stress-event response. Each tells you something the five-year average doesn't.

When Thursday's print lands, the question isn't whether storage is above or below the five-year average. It's whether the five-year average still describes the system you're trading in.