Natural Gas Storage +40 Bcf (vs +31 Est) — Salt Drew 11 Bcf and the Surplus Is Shrinking

Fundamentals · 2026-09-10

Working gas +40 Bcf for the week ending 4 September against a +31 Bcf consensus — a 9 Bcf bearish miss — taking total storage to 3,254 Bcf; but the same +40 was 12 Bcf below the 5-year average build of 52 for this week and 29 Bcf below the 69 Bcf injected a year ago, so the print missed the survey and tightened against both historical benchmarks at once; stocks are 148 Bcf (4.8%) above the 5-year average yet 79 Bcf (2.4%) below last year; South Central fell 7 Bcf and Salt drew 11, leaving Salt 19.2% below last year and the only line in the report under its own 5-year average; South Central alone is 99 Bcf below last year while every other region combined is 20 Bcf above; the surplus to the 5-year average has narrowed 50 Bcf in four weeks because this year is injecting at 67% of the 5-year pace; the standard error on the net change is 0.8 Bcf.

What Is This?

• What it is: The EIA's Weekly Natural Gas Storage Report, released Thursdays at 10:30am ET from Form EIA-912, measuring working gas held in underground storage across the Lower 48 in five regions — East, Midwest, Mountain, Pacific and South Central, with South Central split into salt-dome and non-salt facilities. A positive number is an injection, meaning gas going into storage, which is bearish for price; a negative number is a withdrawal, which is bullish.
• Why it matters: Storage is the only visible buffer in a market with no strategic reserve and almost no import flexibility, so the weekly number is the cleanest read on whether supply is outrunning demand. It matters more than usual right now because energy is doing most of the inflation work in this economy — producer prices for final demand energy are up 24.4% over twelve months in this morning's release, and energy for export is up 40.7%, which is the LNG channel that pulls gas out of exactly the region that drew this week.
• How to read it: Never read the headline alone, because it has three reference points and they routinely disagree. Compare the build to the survey, which is what trades on the release; to the 5-year average build for that specific week, which is what tells you whether the season is running tight or loose; and to the year-ago build. Then read the regions, and read salt first. Salt-dome storage cycles fast — it can be filled and emptied repeatedly in a season — so it is where price-responsive demand shows up first, while the non-salt and northern regions fill slowly on a seasonal schedule and tell you far less about the margin.

Want the full explainer? Economic News Events, Explained breaks down this release and every other one we cover, in plain English.

Summary

Working gas in underground storage rose 40 Bcf in the week ending 4 September, to 3,254 Bcf. Consensus was +31 Bcf, so on the number that trades this was a 9 Bcf bearish miss — more gas went into the ground than the market expected, and NGI's own projection of +34 was closer without being right.

That is the entire bearish case, and it lasts about one sentence.

Because the same +40 Bcf was 12 Bcf below the 5-year average injection of 52 Bcf for this week of the year, and 29 Bcf below the 69 Bcf injected in the comparable week of 2025. A print can miss a survey and still be tight against history, and this one did both at once. The survey was simply wrong about how tight.

The consequence shows up immediately in the two headline comparisons, which point in opposite directions. Stocks are 148 Bcf, or 4.8%, above the five-year average of 3,106 Bcf — the surplus that has framed this market all summer. But they are 79 Bcf, or 2.4%, below where they stood a year ago at 3,333 Bcf. EIA's own summary adds that at 3,254 Bcf total working gas is within the five-year historical range, so nothing here is an outlier in either direction.

Both of those gaps moved the same way this week. Injecting 12 Bcf less than the 5-year norm narrowed the surplus by 12. Injecting 29 Bcf less than last year widened the year-over-year deficit by 29, from −50 to −79. One week, two benchmarks, both tightening.

Zoom out four weeks and the erosion is the story. On 13 August this feed reported stocks of 3,153 Bcf against a five-year average of 2,955 — a surplus of 198 Bcf. Today that surplus is 148 Bcf. It has narrowed by 50 Bcf in four weeks, and the arithmetic of why is stark: this year built 101 Bcf over those four weeks while the five-year average built 151 Bcf. This injection season is running at 67% of the normal pace — roughly 25 Bcf a week against a 38 Bcf norm. Over the same four weeks the year-over-year position deteriorated from 25 Bcf below last year to 79 Bcf below.

Now the regional detail, which is where the report stops being a summer surplus story. Two regions did essentially all the building. East added 20 Bcf and Midwest 18 — together 38 of the 40. Mountain added 3 and Pacific 5. And South Central fell 7 Bcf.

Inside South Central, salt-dome storage drew 11 Bcf while non-salt added 4. Salt now sits at 227 Bcf, down 19.2% from a year ago and — this is the line to take away — 3.8% below its own five-year average of 236 Bcf. It is the only figure in the entire report that is under its five-year benchmark. Every other region and sub-region is above: Pacific by 10.3%, Mountain by 8.6%, East by 5.9%, Midwest by 4.4%, non-salt by 3.7%.

The year-over-year table makes the same point more forcefully. The national deficit is 79 Bcf. South Central alone is 99 Bcf below last year. East is 19 Bcf above, Midwest 22 above, Pacific 3 below and Mountain 18 below — every other region combined is 20 Bcf above last year. In other words the whole national storage deficit, and then some, sits in one region, and it is the Gulf Coast.

That is not a coincidence, and this morning's producer price report supplies the mechanism. Energy for export rose 40.7% over twelve months, the fastest-inflating energy category in the release, against 24.4% for final demand energy overall and 36.7% for government-purchased energy. LNG liquefaction capacity sits on the Gulf Coast, it draws from South Central storage, and salt domes are what it draws from because they cycle fastest. Export prices running 16 percentage points hotter than domestic energy and salt storage 19.2% below last year are two readings of one thing.

Two smaller notes worth making. The northern regions filling comfortably — East 5.9% and Midwest 4.4% above their five-year averages — is what actually matters for winter heating reliability, and on that measure the system is in good shape going into the cold season. And the statistics here are unusually clean: the standard error on the 40 Bcf net change is 0.8 Bcf, and the coefficient of variation on the total stock figure is 0.4%. Pacific's is 0.0%. Of the four releases this desk has read today, this is the one whose headline number carries the least measurement doubt.

The seasonal clock is the last thing to hold. The injection season conventionally runs to the end of October, so roughly eight weeks remain. At the current 25 Bcf weekly pace the market adds around 200 Bcf more; at the 38 Bcf five-year pace it would add 300. The difference between those two numbers is the difference between entering winter with a comfortable cushion and entering it with the surplus gone.

The Internals

The headline against its three reference points, which is the only way to read this release:

Comparison · This week · The benchmark · Gap · Read

Against the survey · Up 40 Bcf · Up 31 Bcf consensus · 9 Bcf more than expected · Bearish, and this is what trades

Against the 5-year average build · Up 40 Bcf · Up 52 Bcf · 12 Bcf less than normal · Bullish, the surplus narrowed

Against the year-ago build · Up 40 Bcf · Up 69 Bcf · 29 Bcf less than last year · Bullish, the deficit widened

Against NGI's projection · Up 40 Bcf · Up 34 Bcf · 6 Bcf more · Bearish

Total stocks, and the two gaps that frame the market:

Measure · Level · Difference · Change %

Working gas, 4 September 2026 · 3,254 Bcf · Up 40 Bcf on the week · Not applicable

Working gas, 28 August 2026 · 3,214 Bcf · The prior week · Not applicable

Year ago, 4 September 2025 · 3,333 Bcf · 79 Bcf below last year · Down 2.4%

Five-year average, 2021-25 · 3,106 Bcf · 148 Bcf above the average · Up 4.8%

EIA's own note · Not applicable · Within the five-year historical range · Not applicable

How the surplus has eroded since this feed last covered the series:

Date · Working gas · Five-year average · Surplus · Versus year ago

Week ending 7 August 2026 · 3,153 Bcf · 2,955 Bcf · Up 198 Bcf · 25 Bcf below

Week ending 4 September 2026 · 3,254 Bcf · 3,106 Bcf · Up 148 Bcf · 79 Bcf below

Change over four weeks · Up 101 Bcf · Up 151 Bcf · Narrowed 50 Bcf · Widened 54 Bcf

Implied weekly pace · 25.2 Bcf per week · 37.8 Bcf per week · 67% of normal · Not applicable

Sampling variability, which EIA publishes and almost nobody reads:

Region · Coefficient of variation on stocks · Standard error on the net change

Total · 0.4% · 0.8 Bcf

East · 0.6% · 0.2 Bcf

Midwest · 0.8% · 0.2 Bcf

Mountain · 2.8% · 0.5 Bcf

Pacific · 0.0% · 0.0 Bcf

South Central · 0.8% · 0.6 Bcf

Salt · 1.2% · 0.6 Bcf

Nonsalt · 1.0% · 0.3 Bcf

The Regional Split

Every region, with both benchmarks. Note that only one line in this table is below its five-year average:

Region · 4 September · 28 August · Net change · Versus year ago · Year-ago % · Versus 5-year average · 5-year %

East · 773 Bcf · 753 Bcf · Up 20 Bcf · Up 19 Bcf · Up 2.5% · Up 43 Bcf · Up 5.9%

Midwest · 908 Bcf · 890 Bcf · Up 18 Bcf · Up 22 Bcf · Up 2.5% · Up 38 Bcf · Up 4.4%

Mountain · 239 Bcf · 236 Bcf · Up 3 Bcf · Down 18 Bcf · Down 7.0% · Up 19 Bcf · Up 8.6%

Pacific · 290 Bcf · 285 Bcf · Up 5 Bcf · Down 3 Bcf · Down 1.0% · Up 27 Bcf · Up 10.3%

South Central · 1,044 Bcf · 1,051 Bcf · Down 7 Bcf · Down 99 Bcf · Down 8.7% · Up 20 Bcf · Up 2.0%

South Central, salt · 227 Bcf · 238 Bcf · Down 11 Bcf · Down 54 Bcf · Down 19.2% · Down 9 Bcf · Down 3.8%

South Central, nonsalt · 817 Bcf · 813 Bcf · Up 4 Bcf · Down 45 Bcf · Down 5.2% · Up 29 Bcf · Up 3.7%

Total, Lower 48 · 3,254 Bcf · 3,214 Bcf · Up 40 Bcf · Down 79 Bcf · Down 2.4% · Up 148 Bcf · Up 4.8%

The year-over-year arithmetic, isolated, because it is the sharpest thing in the release:

Grouping · Versus year ago · Share of the national deficit

South Central · Down 99 Bcf · More than the entire national figure

East, Midwest, Mountain and Pacific combined · Up 20 Bcf · An offset, not a contributor

National total · Down 79 Bcf · 100%

And the same regions ranked by how comfortable they look against their own five-year benchmark:

Rank · Region · Versus 5-year average · Interpretation

1 · Pacific · Up 10.3% · The most comfortable in the country

2 · Mountain · Up 8.6% · Comfortable

3 · East · Up 5.9% · Comfortable, and it matters most for winter

4 · Midwest · Up 4.4% · Comfortable

5 · South Central, nonsalt · Up 3.7% · Adequate

6 · South Central, total · Up 2.0% · Thin, and thinning

7 · South Central, salt · Down 3.8% · The only deficit in the report

Impact on Natural Gas (NG)

• Mixed, lean bearish — the 9 Bcf miss against a 31 Bcf consensus is the tradeable event on the release, and a build is a build.
• Everything structural argues the other way: 12 Bcf below the 5-year norm for the week, 29 Bcf below last year, and a surplus that has shrunk 50 Bcf in four weeks.
• Salt is the tell. Down 11 Bcf on the week, 19.2% below last year, and the only figure in the report below its own five-year average.
• The Gulf Coast carries the whole national deficit — South Central 99 Bcf below last year while every other region combined is 20 Bcf above.
• Eight weeks of injection season remain. At 25 Bcf a week the surplus keeps eroding; at the 38 Bcf norm it rebuilds. That is the position, not the print.

Impact on USD

• Slight bullish — marginal, through the inflation channel rather than anything direct.
• Energy is doing the inflation work: final demand energy up 24.4% over twelve months in this morning's producer report, and a storage picture that is tightening rather than loosening does not reverse that.
• Against it, a 4.8% surplus to the five-year average is not a shortage, and EIA notes stocks are inside the historical range.
• No trade in it. A weekly storage number has never moved the dollar and will not five days before an FOMC.

Impact on US Indices (ES / NQ / YM)

• Mixed — bearish for gas-consuming utilities and chemicals, bullish for the export complex.
• LNG exporters are the clearest beneficiary: energy for export up 40.7% over twelve months, the fastest-inflating energy line in this morning's release, and salt storage drawn down alongside it.
• Utilities and industrial gas consumers face the other side of that, with power burn competing directly against export demand for the same Gulf Coast molecules.
• Chemicals wholesalers already show it — sales up 17.1% over twelve months in today's wholesale trade report on inventories up only 5.8%.
• The winter-reliability question sits with East and Midwest, both comfortably above their five-year averages, so this is a price story rather than a supply-security one.

Impact on Gold

• Slight bullish — one more reading of the energy inflation that has run through every release this fortnight.
• The transmission is the inflation-persistence argument rather than real yields; nothing in a weekly storage report moves rates.
• Against it, a national surplus of 148 Bcf is a comfortable absolute position and argues for no energy shock.
• Forward conditional: if the injection pace stays at 67% of normal through October, the surplus is gone before winter and the inflation argument gets a lot louder.

What To Watch

• Consumer prices — Friday, 11 September, 8:30am ET. Household energy is a direct component, and this morning's producer report already showed residential electric power down 0.5% against final demand energy up 4.2%.
• FOMC decision and dot plot — Tuesday and Wednesday, 15-16 September. Three dissents for a hike in July, with energy still the largest single contributor to the inflation the committee is arguing about.
• Next storage report — Thursday, 17 September, 10:30am ET, covering the week ending 11 September.
• Whether salt draws again. One weekly draw is a heat-and-export artifact; a second turns a summer surplus into a Gulf Coast deficit heading into the cycling season.
• The injection pace against 38 Bcf. Eight weeks remain to the conventional end of injection season on 31 October. This year is running at 25.
• The gap between the two benchmarks. Stocks are 4.8% above the five-year average and 2.4% below last year at the same time. Whichever one the market decides to trade is what sets price into winter.

TLDR

EIA Weekly Natural Gas Storage, week ending 4 September 2026 (released 10 September, 10:30am ET):

• Working gas +40 Bcf against a +31 Bcf consensus — a 9 Bcf bearish miss; NGI had projected +34
• But 12 Bcf below the 5-year average build of 52 for this week and 29 Bcf below the 69 Bcf injected a year ago
• Total stocks 3,254 Bcf — 148 Bcf (4.8%) above the five-year average but 79 Bcf (2.4%) below last year
• EIA notes total working gas is within the five-year historical range
• Both gaps tightened this week: the surplus narrowed 12 Bcf, the year-over-year deficit widened 29
• The surplus has shrunk 50 Bcf in four weeks, from 198 on 7 August to 148 now
• This season is injecting at 67% of the 5-year pace — 25 Bcf a week against a 38 Bcf norm
• East +20 and Midwest +18 made 38 of the 40 Bcf; Mountain +3, Pacific +5
• South Central fell 7 Bcf, and salt drew 11 — salt is 19.2% below last year and 3.8% below its own five-year average, the only deficit in the report
• South Central alone is 99 Bcf below last year while every other region combined is 20 Bcf above — the entire national deficit is one region
• The mechanism is in this morning's producer prices: energy for export up 40.7% over twelve months, against 24.4% for final demand energy
• Winter reliability looks fine: East 5.9% and Midwest 4.4% above their five-year averages
• Cleanest statistics of the day — standard error on the net change 0.8 Bcf, coefficient of variation on total stocks 0.4%

A bearish miss that tightened the market. The 40 Bcf build beat the 31 Bcf survey by nine, which is what the tape reacts to, but it was twelve below the five-year norm for the week and twenty-nine below last year — so the print was loose against expectations and tight against history at the same time, and the survey was simply wrong about how tight. The four-week picture is the one to carry: the surplus to the five-year average has fallen from 198 Bcf to 148 because this season is injecting at 67% of the normal pace. Underneath the national number, the whole story is the Gulf Coast. South Central fell 7 Bcf, salt drew 11, and salt is now 19.2% below last year and the only line in the report under its own five-year average, while South Central alone is 99 Bcf below last year against every other region combined being 20 Bcf above. This morning's producer prices put energy for export up 40.7% over twelve months, which is what is pulling those molecules. Eight weeks of injection season remain. Consumer prices tomorrow at 8:30am ET; the FOMC decides 16 September; next storage 17 September.

_For informational purposes only. Not investment advice._