The Summer Discount on Natural Gas Is a Storage Bill, Not a Bargain

The price that looks like a discount against the January contract is not a seasonal sale. It is the price of holding gas underground for the five or six months until the heating demand arrives.

The calendar is the market

Working natural-gas inventories in the Lower 48 obey the same rhythm every year. From April through October, when air-conditioning load is spotty and industrial demand lacks the heat-driven spike, fields inject gas into reservoirs, and the reported stockpile swells. From November through March, as the nation burns through its underground stash to keep warm, the numbers reverse. The EIA defines injection season as April 1 through October 31 and withdrawal season as November 1 through March 31.

The weekly storage report puts a number on the swing. Published by the EIA for the entire United States and five separate regions, it shows working gas in the ground, measured against a five-year average and a five-year historical range. The average is not a forecast band—it is a statistical comparison, computed with a daily interpolated approach that the EIA has maintained since the report for the week ended January 7, 2005. That chunk of history tells you whether storage is running heavy or light, but it does not tell you what the weather will do next month.

What the weekly storage report actually measures

The report tracks working gas, not every molecule in the reservoir. Some gas stays put to maintain pressure; only the working inventory—the gas that operators can pull out during withdrawal season—matters for prices. When the July quote sits below the January quote, the gap is largely a function of what it will cost to wheel that working gas through the injection-and-withdrawal cycle: space in the facility, fuel to compress it into the rock, and the time value of money until the gas is finally sold.

The EIA’s five-year historical range is presented alongside the average, and it can lull a casual reader into treating the high and low bounds as a prediction cone. They are not. They are simply the maximum and minimum levels recorded in the comparison years. The weekly report is a measurement, not a forecast—a point that matters when the market is deciding whether the summer-winter spread will widen or shrink.

Recent winters and refill seasons

Two back-to-back injection seasons show how fast the inventory picture can shift. At the end of March 2024, working gas in the Lower 48 sat at 2,282 billion cubic feet. That was 25% more than at the same point in 2023, and 40% above the five-year average for March. The going-in cushion was plush.

The injection season that followed, from April through October 2024, added a net 1,640 Bcf. That was 21% less than the five-year average injection. Even so, because the starting point was so high, storage ended the season at 3,922 Bcf—6% above the five-year average, the most gas entering a winter since 2016.

The 2025 injection season flipped the script. The Lower 48 began it with 1,890 Bcf, only 3% above the five-year average. Then net injections hit 2,105 Bcf, 11% more than the average. By October 31, 2025, working gas was 4% above the five-year average—a comfortable start to the heating season, but not the glut of the year before.

Why the spread moves—and what it doesn’t mean

When storage enters withdrawal season unusually full, the market has less reason to bid up winter deliveries. The buyer who would have paid a premium for scarce January supply no longer needs to: the cushion is there, and the refill burden next spring looks smaller. The summer-winter spread compresses. When inventory is tight, the opposite happens. The deferred price has to climb to pull enough gas out of the ground early and store it. That is the storage-carry logic, and it is visible in the numbers without ever opening a weather forecast.

The EIA’s Short-Term Energy Outlook said inventories could reach 3,850 Bcf at the end of October 2026, which would be 2% above the five-year average. Should that hold, the market would enter next winter with a roughly normal cushion—no acute scarcity, no oversupply. The summer-winter spread would reflect the plain cost of storage: lease fees, compression fuel, and the interest bill on capital tied up for half a year.

None of this is a prediction of January temperatures. The weekly report and the STEO do not carry an Arctic blast or a mild stretch. A reader who treats the storage surplus as a guarantee of a soft winter bill is reading a level gauge as a weather map. Equally, the Henry Hub quotes that show the seasonal shape are wholesale prices at a Louisiana pipeline junction, not a residential utility rate. Distribution charges, taxes, and local transmission mark-ups sit on top of the commodity price, and those layers move on their own schedule. The EIA data explain what it costs to hold a physical commodity in the ground; they do not rewrite a utility rate case.

The quote-screen takeaway

The EIA’s injection-season calendar turns again on April 1 every year. By the end of October, the industry will have filled storage to whatever level the summer’s production and demand have allowed. The January contract trading above the July contract is simply the market’s way of billing for the service of carrying that gas across the withdrawal-season boundary. The spread is a storage bill, paid in dollars per million Btu, not a July bargain.