Natural Gas Breaks $2.80 With Storage Tracking a Record 3,985 Bcf

Front-month September Nymex natural gas futures surged past $2.800/MMBtu early Wednesday, reaching a near four-week high after opening at $2.786. The move came from two sources: hotter overnight weather trends and a pullback in Lower-48 production, with firm crude prices adding support as WTI held $85 and Brent pushed $92.

That break matters technically. The $2.78 to $2.80 zone had been the ceiling on every rally this month, and clearing it opens $2.86 to $2.92 as the next objective.

The path into this level was ugly. The front month dropped roughly 2% to $2.68 on Monday, its lowest since August 7, with a separate reading placing the decline at 3% to around $2.65 near three-month lows. Milder weather models and expectations for larger storage injections drove that selling. Futures then edged higher Tuesday on intense late-summer heat and hotter trends in the European model, before Wednesday's break.

The heat is specific and quantified. AccuWeather projects Houston's average high temperatures from August 20 through 23 at 100°F — five degrees above the typical reading for that stretch. Peak demand on the Texas grid could set a new record as wind generation declines. Prolonged, intense heat is forecast across the southern and western United States.

Five degrees does not sound like much until it runs through an air conditioner for four consecutive days across the largest power market in the country.

What caps every one of these rallies is on the supply side. Average output in the Lower 48 has reached 111.6 billion cubic feet per day in August, exceeding July's monthly record of 110.7 Bcf/d. Gas flows to the nine major U.S. LNG export facilities average 17.2 to 17.3 Bcf/d in August, essentially unchanged from July.

Production at a record. Exports flat. Storage 5% above the five-year average. That is the structure a four-day heat wave is fighting.

The next settlement date for the September contract is August 27.

Production at 111.6 Bcf/d Is the Ceiling on Every Rally

The supply picture is the single most important number in this market and it has moved one direction all year.

Lower-48 dry gas production averaged 111.6 Bcf/d in August, surpassing July's monthly record of 110.7 Bcf/d. That is a fresh record on top of a record, achieved with the front month sitting below $2.80. Producers are not responding to price because they do not need to — associated gas from the Permian and improved well productivity keep volumes climbing regardless of the Henry Hub print.

The EIA raised its 2026 production forecast to match that pace at 111.2 Bcf/d. The agency is not modeling a slowdown.

The rig count tells the same story from the drilling side. Baker Hughes has the count holding at 126, below February's 134-rig high but showing no evidence of a pullback despite prices under $3.00. Producers are not capitulating and they are not accelerating. They are holding output at record levels with fewer rigs, which is the definition of efficiency gains offsetting capital discipline.

Wednesday's rally was partly attributed to a pullback in production. That pullback is measured in fractions of a Bcf/d against a 111.6 Bcf/d base — a rounding error that the market treated as a catalyst because there was nothing else on the supply side to trade.

The demand offset from power generation is real but modest. Natural gas-fired electricity generation increased 2% in the first half of 2026, in contrast to year-over-year declines in 2025 when prices were sharply higher. The EIA expects gas-fired generation to increase again in 2027 as prices remain relatively low while coal generation continues declining.

Renewables are eating into that. Solar generation rose 21% in the first half of 2026, wind 6%, and hydropower 9%, against total power sector generation growth of 37 billion kilowatt-hours, or 1.8%. The new 3.7 gigawatt SunZia wind farm keeps that trend running through 2027.

One constructive wrinkle: hydropower is forecast to fall 3% in the second half on intensifying western drought, which shifts some load back to gas.

Storage Is Heading for a Record 3,985 Bcf

The inventory picture is the structural bear case and the EIA has quantified it precisely.

The agency forecasts natural gas inventories at a record 3,985 billion cubic feet at the end of October 2026 — an increase of 19 Bcf compared with the July outlook and 5% above the five-year average. That would put inventories at their highest level heading into winter since 2016.

Inventories have sat above the five-year average continuously since March. Strong production and relatively mild weather earlier in the year built a cushion that no summer heat wave has been able to erode.

The weekly data confirms it. For the week ended August 7, energy firms injected 36 Bcf into storage — larger than market expectations of 31 Bcf and above the five-year average build of 33 Bcf. Three Bcf above normal in a week when Texas was running hot is the cleanest illustration available of how much surplus supply exists.

The Thursday storage report is the week's binding event. Expectations have shifted toward a seasonally lean injection given mid-August cooling demand, and that is what lifted futures Tuesday. A build below the five-year average would be the first genuine tightening signal of the season.

The mechanics matter for winter pricing. Storage entering November at 3,985 Bcf means the withdrawal season starts with roughly 190 Bcf of cushion above normal. That cushion has to be consumed before scarcity pricing can develop, and consuming it requires a cold winter rather than a hot summer.

Periods with higher-than-average inventories are generally associated with lower prices; lower storage levels correspond with tighter conditions. The current setup is unambiguously the former.

For historical context, inventories stood at 3,915 Bcf at the end of October 2025, 4% above the five-year average and near a high point for the decade. The 2026 forecast exceeds that on both the absolute level and the percentage surplus.

The market is not pricing scarcity because there is none.

The EIA Just Cut Its Third-Quarter Forecast by 50 Cents

The August Short-Term Energy Outlook delivered one of the larger single-month revisions in recent memory, and every number moved lower.

The EIA now forecasts the Henry Hub spot price averaging $2.87/MMBtu in the third quarter of 2026 — 50 cents below the prior month's estimate of $3.37. Fourth-quarter guidance dropped to $3.14 from $3.57. The 2026 annual average fell to $3.44 from $3.67, and 2027 to $3.31 from $3.49.

The quarterly path for next year runs $3.62 in the first quarter, $2.79 in the second, and $3.20 in the third — a curve that never reaches $4.00.

The agency's language is direct: with high storage heading into winter, 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 previous forecast. Futures prices show a similar pattern, with contracts through September 2026 remaining below $3.00.

That is a government agency and a futures curve agreeing that $3.00 is a ceiling for another two and a half months.

The revision history frames how far sentiment has moved. In January the EIA saw Henry Hub at just under $3.50 in 2026 before rising to just under $4.60 in 2027. In December 2025 it projected winter 2025/26 averaging about $4.30 and the full year 2026 at $4.01. In July it carried $3.67 for 2026 and $3.49 for 2027.

Roughly $1.20 of 2026 forecast has been removed in eight months, and the entire cut traces to production and storage rather than demand destruction.

The reason given for the July-to-August revision: LNG exports in the third quarter average 16.5 Bcf/d, down 0.2 Bcf/d from the prior forecast because of ongoing Freeport maintenance. Reduced feedgas demand pushed more gas into storage, particularly in the South Central region.

The next STEO lands September 9.

Freeport Maintenance Removed 2.0 Bcf/d of Demand

The LNG side is where the demand story got worse before it gets better, and the timing has been unhelpful for anyone long.

Maintenance at Freeport LNG began July 10 and is expected to complete in late August, affecting 2.0 Bcf/d of nominal export capacity in the short term. That is roughly 1.8% of total Lower-48 production sitting idle at exactly the moment producers set a fresh output record.

Feedgas flows to the nine major U.S. export terminals average 17.2 to 17.3 Bcf/d in August against 17.2 Bcf/d in July — flat when the market needed growth. Peak summer pace has run near 19 Bcf/d in stronger periods, and readings as low as 17.5 Bcf/d were attributed directly to Freeport restrictions.

Freeport returning to full capacity in late August is the most identifiable near-term bullish catalyst in this market. Two Bcf/d of incremental demand arriving as injection season winds down would tighten the South Central balance materially.

The international picture argues for stronger pull once capacity returns. LNG vessel traffic through the Strait of Hormuz slowed considerably after strikes on vessels resumed on July 7, and international prices rose in July to levels last reached in early April. Qatari supply disruption pulls Asian and European buyers toward Atlantic Basin cargoes.

That should support U.S. export economics. It has not shown up in feedgas numbers because the constraint is domestic terminal capacity rather than demand.

Pipeline exports provide a smaller but growing offset. Total U.S. natural gas exports by pipeline are estimated at 9.6 Bcf/d in 2026 rising to 10.0 Bcf/d in 2027, up from 9.5 Bcf/d in 2025, driven by demand from the new Energia Costa Azul LNG terminal supplied from the Permian Basin.

Even with Freeport fully operational, exports remain limited by slow growth in additional export capacity. The next wave of Gulf Coast expansions is a 2027 and beyond story.

The structural bull case for gas depends entirely on that export ramp arriving faster than production grows. Currently it is not.

Technical Structure: $2.92 Is the Real Test

The chart has been range-bound all summer and the boundaries are well established.

The $2.78 to $2.80 zone functioned as the critical resistance area through the first half of August, capping every heat-driven rally. Wednesday's move past $2.800 clears that band, and the breakout exposes $2.86 to $2.92 as the next objective.

Above $2.92, the structure opens toward the $2.991 area that marked the upper boundary of the two-week range earlier in the summer, then $3.00 — the psychological level and the exact figure the EIA says futures do not clear until November.

Downside markers are equally defined. The $2.68 low from Monday is the first support, followed by $2.65 as the three-month low. Beneath that, the $2.556 to $2.543 zone appears in near-term modeling as the lower bound.

The intermediate reference is $2.823, the floor of the earlier two-week consolidation range that ran $2.823 to $2.991. Losing $2.80 after clearing it would confirm a failed breakout and put $2.68 back in play quickly.

Momentum readings support the move without confirming a trend change. Based on technical indicators and moving averages, the daily buy/sell signal reads Buy — but the same signal has flipped repeatedly through a summer where every rally met sellers within days.

Near-term modeling puts Wednesday at $2.691 with a maximum of $2.826, Thursday at $2.677 with a maximum of $2.811, and Friday at $2.612 with a maximum of $2.743. Those projections were built before the breakout and now look conservative on the upside.

The monthly framework has August averaging $2.724 with a maximum of $2.916 and a minimum of $2.481, ending the month at $2.726 for a 1.7% decline. September is modeled at $2.687 average, October at $2.632 with a 6.2% monthly decline.

Every quantitative framework and the government forecast agree: rallies are capped below $3.00 until the heating season arrives.

The Pattern That Has Repeated All Summer

The trading behavior in this market has been mechanical, and recognizing the pattern is worth more than any single forecast.

A hot forecast lifts the bid. Cash prices strengthen in the physical market, ERCOT sets load records, and futures rally two to five percent. Then production data prints at a record, the storage report shows a build at or above the five-year average, and sellers reload into the rally.

That sequence has run at least four times since June. Each cycle ends with the front month back near the bottom of its range and each rally high sits slightly lower than the last.

The recent evidence is direct. Texas ran hot, ERCOT hit load records, cash prices firmed across the West — and the EIA printed a larger-than-expected storage build while LNG feedgas demand ran below its summer peak and production did not slow. The market had reasons to rally and could not hold any of them.

The critical judgment for anyone trading this breakout: what would break the pattern rather than repeat it.

Three things would. Freeport returning to full capacity in late August, adding 2.0 Bcf/d of demand as injection season ends. A storage print materially below the five-year average, confirming the surplus is finally eroding. Or a genuine production decline rather than the fractional pullback that supported Wednesday's move.

Absent those, the current rally is the same trade with a different date attached, and $2.92 becomes the level where sellers appear.

Working against a repeat is the calendar. Injection season is running out. Cooling demand tapers into the fall, but so does the window for building storage, and the market has to decide whether 3,985 Bcf is enough cushion for a winter it cannot forecast.

Late-summer heat matters more when there are fewer weeks left to inject.

The Oil Link Is Working in Gas's Favor

The cross-commodity dynamic has become an unusual tailwind, and it operates through two separate channels.

Firm oil prices were explicitly cited as support for Wednesday's gas rally. WTI reached $85 and Brent pushed toward $92, extending gains for a fourth consecutive session as the 60-day U.S.-Iran memorandum expired with no replacement and Trump confirmed the naval blockade remains in full force.

The first channel is substitution. Cheaper natural gas relative to crude positions gas as the viable alternative in industrial and power applications, and a widening spread pulls incremental demand toward gas. At $85 crude and $2.80 gas, the oil-to-gas ratio sits above 30 to 1 — deeply in gas's favor on an energy-equivalent basis.

The second is associated production. Higher crude prices normally incentivize more drilling in oil-directed basins, which produces more associated gas as a byproduct and adds supply. That channel is bearish for gas over a multi-quarter horizon and is part of why Lower-48 output keeps setting records with the gas rig count at 126.

The Hormuz disruption cuts differently for gas than for oil. LNG vessel traffic through the strait slowed considerably after strikes resumed July 7, which removed Qatari supply from the global pool and lifted international prices to levels last reached in early April. That is structurally supportive of U.S. export economics.

It has not translated into higher feedgas demand because Freeport maintenance capped throughput. Once that constraint lifts, the international pull becomes visible in domestic balances.

Coal is the other substitution story running in reverse. U.S. coal exports surged in April and May, lifting the 2026 forecast to 102 million short tons, with steam coal exports rising in the second quarter as natural gas-to-coal switching in Europe and Asia favored American coal. Domestically, coal generation continues declining on the shift toward lower-cost gas.

That split — coal gaining share abroad while losing it at home — reflects the same fundamental: U.S. gas is cheap relative to everything else, and cheap gas displaces coal domestically while high international gas prices make coal competitive overseas.

Where This Market Has Been

The historical arc is worth stating because it frames how compressed current prices are relative to what this contract can do.

Henry Hub fell to a multi-decade low of $1.63/MMBtu in June 2020 as the pandemic crushed industrial demand. It recovered through 2021, averaging $3.91 — a 91% increase — then spiked to a 14-year high of $9.85/MMBtu on August 29, 2022 during the global energy crisis. It crashed back below $2 in early 2023, recovered through the 2024 LNG export ramp, and then executed another round trip: from below $2 in early 2024 to $7.72/MMBtu in January 2026, and back below $3 by spring.

That $7.72 January 2026 monthly average was set during a polar vortex event and stands as the highest monthly average on record. Seven months later the front month trades at $2.80 — a 64% decline.

The volatility is the point. This is a contract that moved from $1.63 to $9.85 to below $2 to $7.72 to $2.80 inside six years, and each move was driven by weather or a supply shock rather than by any durable change in the underlying balance.

That history argues against treating $2.80 as a settled price. It also argues against assuming a break of $2.92 means anything durable, because every one of those prior moves reversed.

The longer-dated forecasts sit meaningfully higher. The EIA's Annual Energy Outlook projects Henry Hub reaching $3.80/MMBtu by 2030 as LNG exports scale past 20 Bcf/d and data center demand adds a persistent new load floor. Deloitte's 2026 commodity work places it at $5.40/MMBtu on sustained global LNG demand and data center power growth. The consensus 2030 range runs $3.80 to $5.40 depending on export growth, AI demand and weather.

Those numbers require the export capacity that does not exist yet and the data center load that is still being built.

The near-term curve says something entirely different: below $3.00 through October, $3.03 average across the final five months, and a 2027 annual average of $3.31.

What the Data Center Demand Story Actually Delivers

The most heavily promoted structural bull case for natural gas is AI-driven power demand, and it deserves scrutiny against what has actually shown up.

Total U.S. electric power sector generation rose 37 billion kilowatt-hours in the first half of 2026, or 1.8%, compared with the same period in 2025. Natural gas-fired generation increased 2% over that stretch.

Two percent. That is the measured contribution of every data center, every AI training cluster and every electrification trend combined, running through the gas-fired generation stack in the first six months of 2026.

Against that, solar generation grew 21% and wind grew 6% over the same period, with continued capacity additions expected to sustain those rates through 2027. Renewables are absorbing a disproportionate share of load growth, and the 3.7 gigawatt SunZia wind farm adds to that.

The gas-fired share is growing because prices are low and coal is retiring, not because demand is outrunning supply. The EIA expects gas-fired generation to rise in 2027 specifically because natural gas prices remain relatively low.

That is a price-driven substitution story, not a demand shock. It supports volumes without supporting price.

The offsetting near-term factor is drought. Hydropower generation was up 9% in the first half but is forecast to fall 3% in the second half on intensifying western drought conditions, which shifts load toward thermal generation including gas.

Texas is where the demand story is most visible in real time. ERCOT load records with Houston at 100°F for four consecutive days produce genuine gas burn, and cash prices strengthening across the West confirm the heat is reaching the physical market rather than sitting on a weather model.

That is real demand. It is also seasonal, and it disappears in six weeks.

The persistent load floor that would justify $3.80 to $5.40 gas is a 2028 to 2030 proposition on current construction timelines.

The Equity Complex Is Pricing the Same Thing

The producer and infrastructure equities offer a useful cross-check on where the market thinks this goes.

The pure-play gas producers — EQT, Chesapeake and Coterra — trade on the strip rather than on spot, which means they are discounting the same sub-$3.00 curve the EIA published. Their capital discipline is visible in the rig count holding at 126 against February's 134, and in output that keeps rising anyway on efficiency rather than on new drilling.

Cheniere is the leveraged play on the export thesis, and its economics improve when international spreads widen — which they have, given Hormuz disruption pushing Asian and European prices up while Henry Hub sits at $2.80. That arbitrage is the widest it has been since early April.

The leveraged ETF complex is where the volatility gets expressed. UNG tracks front-month futures with the roll cost that a contango curve imposes; BOIL and KOLD provide the double-long and double-short exposure that gets destroyed in a range-bound market like this one.

A summer of failed breakouts and failed breakdowns is the worst possible environment for leveraged products on either side, and the compression from $2.65 to $2.92 across a full month reflects exactly that.

The rig count is the number to watch for a genuine turn. Producers holding at 126 with prices under $3.00 means the marginal barrel is still economic. A drop toward 110 would signal capital discipline arriving, and that is historically what precedes a price recovery — supply response rather than demand growth.

Nothing in the current data suggests that response is close. The EIA raised its production forecast rather than cutting it.

Storage at a record 3,985 Bcf entering winter means producers have no incentive to slow before the withdrawal season proves whether the cushion is needed.

The Forecast: Levels, Triggers and the Verdict

The base case is that the September contract holds above $2.78 into Thursday's storage report and tests $2.86 to $2.92 on the Texas heat.

The bull sequence is short. Hold above $2.800 on a settlement basis, take $2.86, then run $2.92. Clearing $2.92 opens the $2.991 upper boundary of the prior consolidation range and then $3.00 — a level the EIA forecasts futures will not sustainably clear until November. That entire ladder covers 4.3% from spot.

The triggers: a Thursday storage injection materially below the five-year average, Freeport LNG returning to full capacity in late August and adding 2.0 Bcf/d of feedgas demand, Houston delivering the forecast 100°F run from August 20 through 23 with ERCOT setting a new peak load record, or a genuine production decline from the 111.6 Bcf/d August average.

The bear sequence starts with a failed breakout. Losing $2.80 after clearing it puts $2.68 in play immediately, then $2.65 as the three-month low. Below $2.65, the $2.556 to $2.543 zone becomes the reference, with $2.481 as the modeled August floor.

The triggers on that side: another build at or above the five-year average, weather models moderating after the August 23 heat peak, production holding above 111 Bcf/d, or feedgas flows staying flat at 17.2 Bcf/d as Freeport maintenance extends.

The verdict: this is a weather rally inside a structurally oversupplied market, and the supply data is unambiguous. Lower-48 production at 111.6 Bcf/d in August already exceeds July's record of 110.7. The rig count holds at 126 with the EIA raising rather than cutting its output forecast. Storage is tracking toward a record 3,985 Bcf at end-October — 5% above the five-year average and the highest heading into winter since 2016. The EIA cut its third-quarter Henry Hub forecast by 50 cents to $2.87 and expects the spot price below $3.00 until November.

Four days of 100°F in Houston is worth a few cents. It is not worth a regime change against a 36 Bcf injection that beat both consensus and the five-year average during a hot week.

Base case targets $2.92 on the heat and a lean Thursday print. Failure at $2.80 targets $2.68 first and $2.65 as the floor. The genuine turn requires Freeport back at full rate, the rig count falling toward 110, and a winter cold enough to consume a 190 Bcf surplus. None of that is priced, and none of it arrives before November.

That's TradingNEWS