Natural Gas Storage Builds Only 16 Bcf as Salt Caverns Draw 18

Fundamentals · 2026-08-20

Working gas +16 Bcf to 3,169 Bcf for the week ending August 14, below the +19 Bcf consensus and less than half the prior week's +36 Bcf; only two of five regions built — East +15 to 708, Midwest +19 to 848 — while Mountain -2 to 237, Pacific -4 to 296 and South Central -13 to 1,080 all drew; South Central salt -18, nonsalt +5; total stocks 28 Bcf (-0.9 percent) below last year but 185 Bcf (+6.2 percent) above the 2021-2025 five-year average of 2,984 Bcf; EIA's August STEO still projects 3,985 Bcf by end-October, the highest pre-winter level since 2016.

What Is This?

• What it is: The EIA's Weekly Natural Gas Storage Report, released Thursdays at 10:30am ET, measuring the net change in working gas held in underground storage across five regions. Working gas is the portion that can actually be withdrawn and sold, as distinct from base gas needed to maintain reservoir pressure. The injection season runs roughly April through October; the withdrawal season November through March.
• Why it matters: Storage is the single cleanest weekly read on whether US gas supply is outrunning demand, and it is the primary price driver for Henry Hub and NG futures. It also feeds the inflation picture from the cost side — power burn and heating costs flow through to CPI energy services with a lag.
• How to read it: The headline build matters less than where it happened. South Central salt-cavern storage is the fast-cycling capacity in Texas and Louisiana that can inject or withdraw within days; it responds immediately to power burn, heat and LNG feedgas demand. Salt drawing in August is a demand signal, not a supply problem. Note also that the EIA rounds each region independently, so the regional figures here sum to +15 against a reported +16 total — that gap is rounding, not an error.

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 storage rose 16 Bcf to 3,169 Bcf for the week ending August 14, coming in under the 19 Bcf consensus and less than half the prior week's 36 Bcf injection. Two weeks ago the build was 33 Bcf. The trajectory over three weeks is 33, 36, 16 — the injection season is decelerating hard into the back half of August.

The regional composition is the entire story and the headline hides it. Only two of five regions actually added gas. The East built 15 Bcf to 708 and the Midwest added 19 Bcf to 848 — between them more than the national total. Everything west and south of that drew down: Mountain fell 2 Bcf to 237, Pacific fell 4 Bcf to 296, and South Central dropped 13 Bcf to 1,080.

Inside South Central sits the number that matters most. Salt-cavern storage fell 18 Bcf while nonsalt rose 5 Bcf. Salt caverns in Texas and Louisiana are the swing capacity of the US gas market — they can inject and withdraw in days rather than weeks, which makes them the first place Gulf Coast demand shows up. An 18 Bcf salt draw in mid-August means heat-driven power burn plus LNG feedgas demand on the Gulf Coast comfortably exceeded what was flowing in. That is a tight physical balance in the region that sets the marginal price, and it is what turned a routine build into a near-flat national number.

On the aggregate positioning, the picture is genuinely two-sided. Total stocks sit 28 Bcf below where they were a year ago, a 0.9 percent deficit — the year-over-year cushion is now gone. But they remain 185 Bcf above the 2021-2025 five-year average of 2,984 Bcf, a 6.2 percent surplus, and the EIA notes total working gas is within the five-year historical range. So inventories are comfortable against the medium-term norm while having converged all the way back to last year's level.

Then there is the arithmetic problem, and it is worth spelling out because nobody in the coverage did. In its August 11 Short-Term Energy Outlook the EIA projected end-of-October inventories of 3,985 Bcf — the highest level heading into winter since 2016. "More natural gas in inventories in the fall season provides a cushion for increased heating-related consumption during the winter," said EIA Administrator Tristan Abbey. Getting from 3,169 Bcf on August 14 to 3,985 Bcf by October 31 requires roughly 816 Bcf across about eleven weeks — an average of about 74 Bcf per week. This week delivered 16. Shoulder-season builds do routinely accelerate into the 70-100 Bcf range once cooling demand collapses in late September, so the forecast is not unreachable. But it requires a near-fivefold acceleration from here, and every week that runs like this one makes that projection harder to hit.

The price context frames why this is a low-drama print despite tight internals. The EIA's STEO has Henry Hub averaging $2.87/MMBtu in the third quarter, $3.44 for full-year 2026, and $3.31 in 2027, with LNG exports at 16.5 Bcf/d in the third quarter — held down by Freeport maintenance — rising to a 17 Bcf/d full-year average and 19 Bcf/d in 2027. Cheap gas with structurally growing export demand is the setup, and that rising LNG call is precisely what keeps pulling on South Central salt.

The Internals

Regional stocks and net change, week ending August 14, 2026 (Bcf):

Region · Stocks · Net change · Read

East · 708 · +15 · Carried the national build

Midwest · 848 · +19 · Largest regional injection

Mountain · 237 · -2 · Small draw

Pacific · 296 · -4 · Small draw

South Central total · 1,080 · -13 · The offset that halved the headline

South Central salt · 268 · -18 · Fast-cycle capacity, heavy August draw

South Central nonsalt · 812 · +5 · Slow-cycle capacity, still injecting

Total lower 48 · 3,169 · +16 · Below the +19 consensus

Regional figures are rounded independently and sum to +15 against the reported +16 total.

Where total stocks sit:

Comparison · Level · Difference · Percent

Current, week ending Aug 14 2026 · 3,169 Bcf · n/a · n/a

Same week last year · 3,197 Bcf · -28 Bcf · -0.9 percent

Five-year average, 2021-2025 · 2,984 Bcf · +185 Bcf · +6.2 percent

The recent build trend:

Week ending · Build · Consensus · Result

July 31, 2026 · +33 Bcf · n/a · n/a

August 7, 2026 · +36 Bcf · +31 Bcf · Bigger than expected

August 14, 2026 · +16 Bcf · +19 Bcf · Smaller than expected

The EIA's own end-of-October projection, against what this week delivered:

Measure · Value

Current stocks, August 14 · 3,169 Bcf

EIA STEO end-October projection · 3,985 Bcf

Implied total injection required · about 816 Bcf

Weeks remaining to October 31 · about 11

Implied average weekly build required · about 74 Bcf

This week's actual build · 16 Bcf

Supply, Demand and the LNG Call

The EIA's August 11 Short-Term Energy Outlook sets the backdrop this report is being measured against:

Forecast · Value · Note

End-October 2026 inventories · 3,985 Bcf · Highest pre-winter level since 2016

Henry Hub, 3Q 2026 · 2.87 dollars per MMBtu · Cheap gas through the summer

Henry Hub, full-year 2026 · 3.44 dollars per MMBtu · Full-year average

Henry Hub, 2027 · 3.31 dollars per MMBtu · Modest expected easing

LNG exports, 3Q 2026 · 16.5 Bcf per day · Held down by Freeport maintenance

LNG exports, full-year 2026 · 17 Bcf per day · Average

LNG exports, 2027 · 19 Bcf per day · Structural growth in the export call

Two forces are pulling in opposite directions on the same storage number. Rising LNG feedgas demand concentrated on the Gulf Coast draws directly on South Central salt, which is exactly what showed up this week. Against that, Freeport maintenance is temporarily suppressing third-quarter export volumes, so the 16.5 Bcf/d running rate understates where the structural call is heading — 19 Bcf/d by 2027. When that maintenance rolls off, the pull on salt gets stronger, not weaker.

The counterweight is production. The EIA is confident enough in supply to forecast the largest pre-winter inventory since 2016 despite that export growth, which implies it expects dry-gas output to keep outrunning the export call through the shoulder season. This week's print is one data point against that thesis, not a refutation of it.

Impact on Natural Gas (NG)

• Slight bullish — a below-consensus build with an 18 Bcf salt draw points to a tighter physical balance than the headline number suggests.
• The three-week trajectory of 33, 36, 16 Bcf is the sharpest deceleration of the injection season so far.
• The year-over-year surplus has fully eroded, from a cushion to a 28 Bcf deficit, which removes a bearish talking point.
• Capping the move: stocks remain 185 Bcf and 6.2 percent above the five-year average, and the EIA still projects the highest pre-winter inventory since 2016.
• The real question is whether the salt draw is weather noise that reverses with the first cool front, or the LNG call showing up structurally — Freeport maintenance rolling off is the tell.

Impact on USD

• Slight bullish — energy-cost pressure feeds the inflation side of the mandate at a moment when the Fed has three members dissenting for a hike.
• Second-order at best: a 3 Bcf miss on a weekly storage number is not a currency driver in isolation.
• The channel that matters is CPI energy services, and it works with a lag of months, not days.

Impact on US Indices (ES / NQ / YM)

• Mixed, lean bullish — Henry Hub near $2.87 in the third quarter is a genuine input-cost tailwind for utilities, chemicals and heavy industry.
• Cheap gas is quietly load-bearing for the AI-datacenter story, where power cost is the binding constraint on buildout economics — the same capex theme the Conference Board flagged this morning.
• Energy producers take the opposite side: sub-$3 gas compresses upstream margins even as volumes grow.
• Tightening physical balances into winter are the risk to that tailwind, and salt draws in August are how that risk first announces itself.

Impact on Gold

• Slight bullish — marginal, via the energy-inflation channel only.
• A tighter gas balance into winter supports the sticky-inflation case that underpins the stagflation hedge.
• Realistically negligible for the level — CPI on September 11 and the dot plot dominate everything here.

What To Watch

• Next EIA storage report — Thursday, August 27, 10:30am ET. Does salt keep drawing, or was this a heat-driven one-off?
• The weekly build run rate. Anything sustained below 50 Bcf puts the EIA's 3,985 Bcf end-October projection out of reach and re-rates the winter curve.
• Freeport LNG maintenance rolling off. Third-quarter exports of 16.5 Bcf/d are artificially suppressed; the return to 17 and then 19 Bcf/d increases the structural pull on South Central salt.
• The year-over-year line. Stocks just crossed from surplus to a 28 Bcf deficit versus 2025; a widening deficit into September changes the winter setup materially.
• September STEO. Watch whether the EIA revises that end-October inventory projection down, which would be the cleanest admission that the injection season is running short.

TLDR

EIA Weekly Natural Gas Storage Report (week ending August 14, released August 20):

• Net change: +16 Bcf (vs +19 est, +36 prior) — a miss, and less than half the prior week
• Total working gas: 3,169 Bcf
• Versus last year: -28 Bcf, -0.9 percent — the year-over-year cushion is gone
• Versus five-year average: +185 Bcf, +6.2 percent — still a comfortable surplus
• East: +15 to 708; Midwest: +19 to 848 — the only two regions that built
• Mountain: -2 to 237; Pacific: -4 to 296; South Central: -13 to 1,080 — all drew
• South Central salt: -18 Bcf, the single number that halved the national build
• South Central nonsalt: +5 Bcf — slow-cycle capacity still injecting
• Three-week build trend: 33, 36, 16 Bcf — sharp deceleration
• EIA STEO still projects 3,985 Bcf by end-October, requiring roughly 74 Bcf per week from here

A 16 Bcf build reads boring until you see that only two regions built and salt caverns drew 18 Bcf in the middle of August — Gulf Coast power burn and LNG feedgas are eating the injection season from the inside. Stocks are still 6.2 percent above the five-year average, but the year-over-year cushion has gone to a 28 Bcf deficit, and the EIA's own 3,985 Bcf end-October forecast now needs roughly 74 Bcf a week when this week delivered 16. Watch the August 27 report for whether salt keeps drawing, Freeport maintenance rolling off, and the September STEO for a revision to that end-October number.

_For informational purposes only. Not investment advice._