Injections Collapse 58% In Three Weeks As Heat Holds Through September 4 — October Target Now Needs 80 Bcf A Week

US natural gas prices rose to $2.90/MMBtu on Thursday, the highest level in a month, extending Wednesday's advance from $2.80 as forecasts for several more weeks of hot weather lifted cooling demand expectations.

The move ran directly into the storage report, and the report validated it. Working gas in storage totaled 3,184 Bcf as of Friday, August 21 — a net increase of just 15 Bcf from the previous week. The EIA Weekly Natural Gas Storage Report released the figure at 10:30 a.m. ET.

That build came in roughly a third smaller than the 23 Bcf the market was projecting and below the 17 Bcf injected during the same week last year. It also marks the second consecutive week of mid-teens injections after three straight weeks in the high twenties to mid thirties.

The immediate driver is heat that will not quit. Well-above-average temperatures are expected across the eastern two-thirds of the United States from August 31 through September 4, likely sustaining gas demand from power generators as air-conditioning use stays elevated. A persistent heat dome across the South has been supporting strong regional cooling demand and slowing the pace of storage injections for weeks.

The offsetting force is supply, and it is enormous. Lower 48 production has averaged a record 111.4 Bcf/d so far in August, up from 110.7 Bcf/d in July. That is the wall every rally in this market has run into all year.

Feedgas to the nine major LNG export facilities has averaged 17.1 Bcf/d this month, slightly below July's 17.2 Bcf/d, though signs of recovery are emerging as Freeport maintenance winds down.

The price context frames how modest this rally actually is. Henry Hub spiked toward $7/MMBtu in January on a polar vortex, then fell below $3.00 by mid-March and has spent the entire summer beneath it. The September NYMEX contract closed down 2.88% on August 20 near $2.73 before recovering.

A one-month high at $2.90 is a recovery within a bear market, not a breakout from one.

The Storage Print: 3,184 Bcf And A Build Half The Expected Size

The number that matters is the shape of the injection season, and it has been deteriorating for four consecutive weeks.

Working gas reached 3,184 Bcf as of August 21 on a 15 Bcf net increase. Against a projection of 23 Bcf, that is a 35% miss on the bearish side of the estimate — a genuinely tight print for late August.

The four-week sequence tells the story:

Week ending July 24: +28 Bcf to 3,084 Bcf, against a five-year average build of 23 Bcf for that week.
Week ending July 31: +33 Bcf to 3,117 Bcf, above a 31 Bcf consensus and well above the 13 Bcf injected a year earlier.
Week ending August 7: +36 Bcf to 3,153 Bcf, above a 31 Bcf consensus but far below the 49 Bcf year-ago build and the 33 Bcf five-year average.
Week ending August 14: +16 Bcf to 3,169 Bcf, below a 19 Bcf consensus, below the 19 Bcf year-ago figure, and dramatically below the 29 Bcf five-year average.
Week ending August 21: +15 Bcf to 3,184 Bcf, against a 23 Bcf projection and a 17 Bcf year-ago build.

Injections have fallen from 36 Bcf to 16 Bcf to 15 Bcf in three weeks — a 58% collapse in the refill rate at exactly the point in the calendar when storage operators need to be filling ahead of winter.

The surplus position is what has kept this from mattering more. At 3,169 Bcf on August 14, stocks sat 6.2% above the five-year average and 0.9% below year-ago levels. Two weeks of undersized builds narrow that surplus but do not eliminate it.

The regional detail from earlier in the summer showed where the tightness sits. In the week ending July 24, Mountain withdrew 2 Bcf, Pacific withdrew 7 Bcf, and South Central withdrew 9 Bcf, with South Central salt alone withdrawing 14 Bcf while nonsalt stocks ran 6.0% below last year.

Salt-dome withdrawals during injection season are the signature of heat-driven power burn overwhelming regional supply. That is the mechanism converting a Texas heat dome into a national storage number.

Record Production At 111.4 Bcf/d Is The Wall

Every bullish input in this market has to get past the same obstacle, and the obstacle keeps getting bigger.

Lower 48 dry gas production has averaged a record 111.4 Bcf/d so far in August, up from 110.7 Bcf/d in July. That is a 0.7 Bcf/d monthly increase, or roughly 4.9 Bcf of incremental weekly supply — which is by itself larger than the shortfall in this week's storage build against the projection.

The production growth is driven primarily by activity in the Permian Basin, where gas comes out as an associated byproduct of oil drilling. That is the structurally difficult part of the supply picture: Permian gas is not price-responsive in the way Haynesville or Marcellus production is, because the drilling decision is made on oil economics.

With WTI at $81.36 and US crude output approaching a record 14 million barrels per day, with active rigs at their highest level of the year and completion crews at a 16-year high, associated gas keeps arriving regardless of what Henry Hub does.

The consequence for the balance is that gas has been able to absorb a record heat dome, record power burn, salt-dome withdrawals in South Central, and two weeks of mid-teens injections while still holding a 6%-plus surplus to the five-year average.

Production has occasionally dipped and the market has noticed. Futures pressed higher earlier this week as enduring summer heat intersected with a dip in production, and squeezed out a small gain Monday as sweltering Texas heat and reduced output kept bulls afloat after an overnight rally faded.

Those are one-day supply wobbles against a structural record. The direction of the trend has been up all year and there is no forecast calling for it to reverse.

The one genuine crack is regional rather than national. Appalachian spot prices have sunk to levels that could trigger Marcellus Shale supply curtailments this fall — the first sign that the production wall has a price beneath which it stops building.

Curtailments would be the mechanism that finally tightens the balance. They have not happened yet.

LNG At 17.1 Bcf/d And The Freeport Restart

The export channel is the swing factor for the fourth quarter and it is about to change.

Average gas flows to the nine major LNG export facilities have run 17.1 Bcf/d so far this month, marginally below July's 17.2 Bcf/d, with signs emerging of a recovery in demand.

The reason feedgas has been soft is maintenance rather than weak international demand. Freeport LNG began maintenance on July 10, removing roughly 2.0 Bcf/d of nominal export capacity, with completion expected in late August. That work is finishing right now.

Third-quarter LNG exports are forecast to average 16.5 Bcf/d, revised down 0.2 Bcf/d from the prior estimate.

The restart matters arithmetically. Two Bcf/d of incremental feedgas demand equals 14 Bcf of weekly draw on the domestic balance — roughly the size of this week's entire storage injection. If Freeport returns fully and flows normalize, the injection season effectively ends two weeks early.

The constraint on that scenario is that even with Freeport fully operational, exports would remain limited by slow growth in additional export capacity, despite US price spreads to Europe and Asia remaining wide.

That is the structural frustration of this market. Henry Hub near $2.90 while European and Asian prices sit substantially higher is an arbitrage that physical infrastructure cannot execute at scale. The terminals are running near capacity and new capacity arrives on multi-year timelines.

International prices rose in July to levels last reached in early April as LNG vessel traffic through the Strait of Hormuz slowed considerably after strikes on vessels resumed on July 7. Qatari cargoes displaced from the Gulf tightened the Atlantic basin without lifting Henry Hub, because the US export machine was already maxed out.

European natural gas has since jumped to its highest level since 2023, with euro-area energy inflation accelerating to 10.0% in July and lower-than-normal EU storage adding tightness. Diesel prices across the continent are rising in parallel.

None of it reaches Louisiana at a meaningful scale until new liquefaction capacity comes online.

The Weather Trade: August 31 Through September 4

The next seven days are the last genuine cooling-demand event of the season and the market is pricing it.

Well-above-average temperatures are forecast across the eastern two-thirds of the United States from August 31 through September 4. That window captures the Labor Day holiday and the transition into September — a period when cooling demand normally fades hard and injections normally accelerate.

If that heat verifies, injections stay suppressed through the first week of September, and the storage surplus narrows further exactly when the market starts pricing winter risk.

The recent history shows how sharply this market trades weather forecasts. On August 20, the September contract closed down 2.88% on forecasts for cooler US temperatures. On August 21, it closed up 1.46% as forecasts flipped hotter. That is a 4.3% swing across two sessions driven entirely by model runs.

The regional dispersion has been extreme. Record-breaking heat across the southern tier lifted Gulf Coast and Western hubs while Appalachia sank toward curtailment-triggering levels. Spot prices for weekend and Monday delivery gave up ground overall as milder weather signals in the northern Midwest and East pushed buyers to the sidelines.

That split — South scorching, North moderating — is why total demand and Henry Hub prices have been muted despite the heat dome. Texas power burn is enormous but it is one region against a national balance carrying a 6% storage surplus.

The seasonal math is unforgiving after this window closes. Cooling degree days collapse through September. Heating demand does not begin meaningfully until late October. That six-to-eight-week shoulder period is when storage fills fastest and prices typically bottom.

The market has roughly ten weeks of injection season remaining. At the current 15 Bcf pace, that adds 150 Bcf and takes storage to approximately 3,334 Bcf. At the 30 Bcf pace that prevailed in July, it adds 300 Bcf to about 3,484 Bcf.

Both figures land well below the 3,985 Bcf end-October target in the current official forecast — which is the most interesting discrepancy in this entire market.

The El Niño Question And The Winter Setup

The variable that could reprice this entire curve is a Pacific Ocean anomaly that nobody can handicap yet.

A potentially historic El Niño is increasingly reshaping natural gas weather risks heading into fall and winter, with characterizations running as far as a "Godzilla El Niño" sending the market into unknown territory.

The conventional El Niño relationship for North American gas is bearish: warmer-than-normal winters across the northern tier, reduced heating demand, weaker withdrawals. That is the pattern the market defaults to pricing.

The complication is that a historically strong event does not necessarily behave like a moderate one. Extreme El Niño winters have produced sharply divergent regional outcomes, including cold intrusions into the southern US where the gas system has the least winterization and the highest price sensitivity.

The precedent is fresh. Last winter delivered a polar vortex that spiked Henry Hub toward $7/MMBtu in January and produced record storage withdrawals of 2,020 Bcf across the season — 4% above the five-year average. Winter Storm Fern drove temporary price spikes that influenced the February forecast of $4.31/MMBtu for the full year.

That forecast is now $3.44. The market has spent six months unwinding a winter premium.

The current curve reflects the standard El Niño assumption. The December 2026 contract trades above $4/MMBtu, and the mid-range of the forward channel sits near $4/MMBtu, which represents the market's central expectation for winter 2026-27. The upper boundary near $5/MMBtu is described as reachable only under a significantly colder-than-normal winter.

That structure — front months below $3.00, December above $4.00 — is a $1.10-plus contango that pays storage operators to inject and hold. It is also why producers have no incentive to curtail: hedging the December contract locks in a price 38% above spot.

The risk to the bearish case is that a historic El Niño produces a non-standard outcome and the market has to reprice a $4 December contract that it currently treats as the ceiling.

EIA Cut Its 2026 Forecast To $3.44 And The Trajectory Says More

The official price outlook has been revised down all year and the direction has been consistent.

The August Short-Term Energy Outlook lowered the 2026 Henry Hub spot price forecast by more than 6%, from $3.67 to $3.44/MMBtu. That followed a July revision that had actually raised the number from $3.60 to $3.67.

The full-year trajectory is the more telling number. The annual spot price forecast has fallen more than 20% since the February estimate of $4.31/MMBtu, which was inflated by sustained heating demand, record storage withdrawals and the Winter Storm Fern spikes.

The near-term path in the August outlook is explicit: the Henry Hub spot price is expected to remain below $3.00/MMBtu until November and average $3.03/MMBtu over the remaining five months of the year — nearly 50 cents lower than the prior month's forecast. Full detail publishes through the Short-Term Energy Outlook.

Futures show the same pattern, with contracts through September 2026 remaining below $3.00/MMBtu.

At $2.90 today, spot sits within a dime of that ceiling. The forecast and the market are aligned, and both are describing a commodity that cannot clear $3.00 until the heating season arrives.

The quarterly detail from the prior outlook — since revised lower — had Q3 2026 at $3.37 and Q4 at $3.57, with 2027 running $3.83 in Q1, $2.99 in Q2, $3.36 in Q3 and $3.78 in Q4.

The longer-horizon story is more constructive and it has been consistent across every revision. Supply growth outpaces demand growth by 0.5 Bcf/d in 2026 but falls behind by 1.6 Bcf/d in 2027, putting upward pressure on prices. The driver is feed gas demand from LNG export facilities reducing gas in storage.

That is a 2027 story requiring liquefaction capacity that does not yet exist to come online. The January outlook projected 2027 averaging just under $4.60/MMBtu — a 33% annual increase — before subsequent revisions brought it toward $3.49.

The dispersion between those two 2027 numbers is the honest measure of how uncertain this forecast is.

The 3,985 Bcf October Target And The Math That Does Not Work

The single most interesting discrepancy in the natural gas market right now is between the official storage projection and the observed injection pace.

The August outlook forecasts natural gas inventories at a record 3,985 Bcf at the end of October 2026 — an increase of 19 Bcf from the July estimate and 5% above the five-year average. The third-quarter working gas outlook was raised to more than 3.6 Tcf, a 0.6% gain over the prior forecast.

Storage stood at 3,184 Bcf on August 21. Reaching 3,985 Bcf by October 31 requires 801 Bcf of net injection across roughly ten weeks — an average of 80 Bcf per week.

The last two weekly builds were 15 Bcf and 16 Bcf.

Even at the strongest injection pace of this summer — 36 Bcf in the week ending August 7 — ten weeks delivers 360 Bcf and takes storage to approximately 3,544 Bcf. That is 441 Bcf short of the target.

The reconciliation is seasonal. Injections accelerate dramatically once cooling demand collapses in mid-September, and the shoulder season regularly produces 80 to 100 Bcf weekly builds when power burn falls away and production stays flat. The forecast is not obviously wrong; it is dependent on a step-change in the injection rate within three weeks.

But it assumes cooling demand ends on schedule, production holds at 111.4 Bcf/d, and Freeport's 2.0 Bcf/d restart does not materially draw down the domestic balance.

Three assumptions, and the current weather forecast already puts above-normal temperatures across two-thirds of the country through September 4.

The reason it matters for price is direct. High storage heading into winter is the stated basis for the sub-$3.00 forecast through November. Inventories moving closer to or below the five-year average is what lifts the forecast price — storage levels remain the key indicator of market balance and price formation.

If the injection pace stays at 15 to 30 Bcf rather than stepping to 80, storage enters November closer to 3,500 Bcf than 3,985 Bcf, the surplus to the five-year average narrows sharply, and the sub-$3.00 assumption breaks.

The next outlook revision publishes September 9.

Appalachia, Marcellus Curtailments And The Basis Problem

The regional price collapse in the Northeast is the mechanism that would rebalance this market, and it is close to triggering.

Appalachian spot prices have sunk to levels that could trigger Marcellus Shale supply curtailments this fall. That is producers shutting in wells because the netback no longer covers operating and transport costs — the price-responsive supply reduction that Permian associated gas cannot deliver.

The dynamic is straightforward. Marcellus and Utica production is constrained by pipeline takeaway capacity out of Appalachia. When national demand softens and pipes fill, in-basin prices disconnect from Henry Hub and fall to whatever level clears local supply. Producers then face a choice between selling at a loss and shutting in.

Physical natural gas prices were mixed earlier this week precisely on this split — record heat across the southern tier lifting Gulf Coast and Western hubs while Appalachia sank.

If curtailments materialize, they remove supply at the margin and tighten the national balance without requiring any change in demand. Historically, Appalachian shut-ins have removed 1 to 3 Bcf/d during severe basis blowouts, which is a meaningful fraction of the 111.4 Bcf/d total.

The timing would be favorable. Curtailments arriving in October, as the injection window closes and Freeport returns to full rates, would compress the storage surplus into the heating season.

The counterargument is that these producers are heavily hedged. With December 2026 futures above $4/MMBtu and the 12-month strip well above spot, a producer with hedges on has no economic reason to shut in — they deliver against the hedge and take the basis loss.

The basis structure also explains why national storage looks comfortable while regional markets are stressed. Working gas at 3,184 Bcf is an aggregate. South Central salt withdrew 14 Bcf in a single week in July while national inventories built 28 Bcf.

Aggregate numbers hide the regional tightness that actually sets marginal prices during a winter event.

The International Spread That Cannot Reach Henry Hub

The most extreme dislocation in global energy is between US and international gas, and it has been persistent enough to stop being news.

Henry Hub trades $2.90/MMBtu. European and Asian prices remain substantially above that, with European gas at its highest level since 2023 and euro-area energy inflation accelerating to 10.0% in July from 8.5% in June — a component running at five times the ECB's headline target and the single largest reason the central bank is preparing to hike on September 10.

International prices rose in July to levels last reached in early April as LNG vessel traffic through the Strait of Hormuz slowed considerably after strikes on vessels resumed on July 7. Qatar is the world's largest LNG exporter and its cargoes transit Hormuz.

Lower-than-normal EU storage inventories add another layer of tightness.

None of it transmits to Henry Hub because the bottleneck is liquefaction, not price. US LNG feedgas runs 17.1 Bcf/d against nine major facilities operating near capacity. Even with Freeport fully restored, exports remain limited by slow growth in additional export capacity despite the wide spreads.

The comparison that captures it: European buyers have paid multiples of the US benchmark throughout 2026 and US producers have been unable to arbitrage it at scale.

That is the structural story underneath every 2027 forecast. A growing LNG export base is steadily linking US domestic gas markets to global demand, placing a firmer floor under Henry Hub than existed two years ago. Feed gas demand from new projects — Plaquemines and Golden Pass among them — is projected to push flows above 16.5 Bcf/d, with AI and data centre power burn adding a cumulative 2 to 3 Bcf/d.

Storage deficits are expected to re-emerge over winter 2026-27 on that combination.

The Hormuz situation adds a wildcard running the other direction. Iran and Oman have reached agreement on territorial waters and revenue sharing in the strait, with Brent falling to $86.93 on a fourth consecutive down session. If LNG traffic normalizes, international prices ease, the spread narrows, and the pull on US exports weakens marginally.

That would be bearish for Henry Hub at the margin, though the binding constraint remains capacity rather than demand.

The Curve: December Above $4 And What The Market Is Pricing

The forward structure is where the real information sits, and it is telling a different story from spot.

Front-month gas at $2.90 against a December 2026 contract above $4.00/MMBtu is a contango exceeding $1.10, or 38%. That is an unusually steep seasonal spread and it carries three implications.

First, it pays storage. Any operator with capacity can buy September gas, inject it, and sell December forward at a $1.10 premium — far above the cost of carry. That mechanism is what keeps injections happening even at depressed spot prices, and it is why the market has been able to build toward record inventories despite prices below $3.00.

Second, it removes the incentive for producers to curtail. A Marcellus producer facing negative in-basin economics today can hedge December at $4 and continue producing.

Third, it means the winter recovery is already priced. The mid-range of the forward channel sits near $4/MMBtu as the most likely destination for prices heading into winter 2026-27. The upper boundary near $5/MMBtu is reachable only under a significantly colder-than-normal winter.

Henry Hub prices are expected to remain subdued through the remainder of summer at roughly $2.80 to $3.00/MMBtu before firming into the fourth quarter as heating season approaches and LNG feedgas demand peaks.

The timing and magnitude of that recovery depend almost entirely on weather.

The historical range this year illustrates how wide the distribution is. January spiked toward $7/MMBtu on a polar vortex with record 2,020 Bcf seasonal withdrawals. Prices fell below $3.00 by mid-March as mild spring weather, healthy injections and rising LNG capacity from Golden Pass and Corpus Christi Stage 3 reshaped the balance. The injection season opened at 1,829 Bcf on March 27 with Henry Hub at $2.90 — the same price it trades today, five months and 1,355 Bcf later.

That symmetry is worth sitting with. Storage has built 74% and the price is unchanged. The entire summer's supply build has been absorbed by the curve rather than by the spot market.

Technical Structure: $2.74 Below, $2.90 Now, $3.00 Overhead

The levels on the front-month contract are tightly compressed and the ceiling is a round number that also happens to be a forecast.

Current price at $2.90 marks the highest level in a month. Above it, $3.00 is the psychological line, the level futures contracts through September have failed to clear, and the explicit ceiling in the current official forecast, which projects spot remaining below $3.00 until November.

That is a triple confluence at a round number, and clearing it would require the market to reject the storage surplus thesis outright.

Below spot, the recent range is well defined. The September contract closed down 2.88% on August 20 near $2.73 on a cooler weather outlook, then recovered 1.46% on August 21 as forecasts flipped. That $2.73 to $2.74 area is the base of the current move and the level that would need to break for the one-month high to be dismissed as noise.

Beneath that, the summer has produced repeated tests near $2.80, which is where the market traded as recently as Wednesday.

The wider frame runs from the January spike toward $7.00/MMBtu to the sub-$3.00 collapse by mid-March, with the market ranging between roughly $2.70 and $3.10 for the entire summer.

Momentum readings turned constructive during the recovery. Relative strength crossed back above 30% after the market bounced from oversold territory, with the standard warning about mean reversion attached.

The practical read: this is a range-bound market with a hard ceiling at $3.00 and a base at $2.74, currently trading in the upper third of that band on a genuine fundamental input — two consecutive undersized storage builds and a heat forecast extending through September 4.

A close above $3.00 changes the character of the market and puts $3.20 to $3.30 in play, which would be the first real challenge to the sub-$3.00-until-November thesis. Failure at $3.00 returns price to the $2.74 to $2.90 band it has occupied all summer.

The December contract above $4.00 is the level to watch for evidence that winter expectations are shifting. It has not moved on this week's storage data.

What Would Break The Range In Either Direction

Upside requires three things and two are already in motion.

The injection pace stays at 15 to 30 Bcf per week rather than stepping to the 80 Bcf average required to reach the 3,985 Bcf end-October target. Two consecutive prints at 16 and 15 Bcf have already established the pattern, and the August 31 to September 4 heat forecast extends it at least one more week.

Freeport LNG completes maintenance and 2.0 Bcf/d of feedgas returns, adding roughly 14 Bcf of weekly demand — larger than this week's entire injection.

Appalachian basis stays depressed and Marcellus curtailments materialize, removing 1 to 3 Bcf/d of supply into the shoulder season.

That combination takes storage into November near 3,500 Bcf rather than 3,985 Bcf, compresses the surplus to the five-year average from 6% toward zero, and breaks the sub-$3.00 assumption that anchors the current forecast. A close above $3.00 opens $3.20 to $3.30, with the $3.44 full-year forecast as the extension target.

Downside requires only reversion to seasonal norms.

Cooling demand collapses after September 4 on schedule. Production holds at the record 111.4 Bcf/d or extends higher on continued Permian associated gas. Injections step to 60 to 90 Bcf as the shoulder season arrives, and storage tracks toward the 3,985 Bcf record. A historic El Niño delivers the standard warm-winter outcome across the northern tier.

That path returns price to the $2.74 base and, if the December contract starts unwinding its premium, opens the low $2.60s.

The genuine wildcard is the El Niño intensity. A conventional event is bearish. A historic one is unpredictable, and last winter's polar vortex and 2,020 Bcf of record withdrawals are recent enough that nobody wants to be short the December contract on a forecast.

The macro backdrop is neutral to marginally supportive. Brent at $86.93 and WTI at $81.36 have fallen four consecutive sessions on Hormuz diplomacy, which eases the international LNG pull. European gas at multi-year highs pulls the other way.

Neither reaches Henry Hub while liquefaction capacity is the binding constraint.

Forecast And Verdict: $3.00 Is The Ceiling, $2.74 Is The Base

Constructive within a range that has not broken all summer.

The bullish evidence is genuinely fresh. Working gas came in at 3,184 Bcf on a 15 Bcf build — roughly a third below the 23 Bcf projection and below the 17 Bcf injected a year ago. That follows 16 Bcf the prior week against a 19 Bcf consensus and a 29 Bcf five-year average. Injections have collapsed 58% from 36 Bcf three weeks ago. Well-above-average temperatures cover the eastern two-thirds of the country through September 4. Freeport's 2.0 Bcf/d returns from maintenance this month. Appalachian basis sits near curtailment triggers. And the 3,985 Bcf end-October target requires an 80 Bcf weekly average that the market has not produced once this summer.

The bearish evidence is structural and it has won every argument this year. Lower 48 production is at a record 111.4 Bcf/d, up from 110.7 Bcf/d in July, driven by price-insensitive Permian associated gas at a time when US crude output approaches 14 million barrels per day. Storage held a 6.2% surplus to the five-year average as of August 14. LNG feedgas at 17.1 Bcf/d is capacity-constrained regardless of a wide international spread. The 2026 forecast has been cut from $4.31 in February to $3.44 in August, with spot expected below $3.00 until November. Contracts through September remain below $3.00. And a potentially historic El Niño defaults to a warm-winter outcome.

The levels: $3.00 is the ceiling — a round number, a forecast assumption, and a level front-month contracts have not cleared. Breaking it opens $3.20 to $3.30. Below, $2.90 is the current print, $2.80 is Wednesday's level, and $2.74 is the base established on August 20. Losing $2.74 targets the low $2.60s. The December contract above $4.00 is the separate signal to watch.

Call it bullish into the September 4 heat window and capped above it. Two undersized builds and a Freeport restart are real, but 111.4 Bcf/d of record production and a 3,184 Bcf inventory are the wall this market has failed to climb all summer.

That's TradingNEWS