And that need for scale — for marketing reasons as well as upstream ones — matters most for one company: Range Resources. Range has one of the most attractive remaining-inventory positions in the industry, but as I’ll show, it lacks an attractive path to bulking up, making it a more likely acquisition target than an acquirer.

Why scale matters

Large-cap oil names already trade at a premium to smaller peers, and in the long run, scale will be even more important for US gas E&Ps. Oil E&Ps’ deteriorating rock quality has driven recent (and ongoing) consolidation, as operators both reduce G&A expenses and improve productivity with larger, more contiguous acreage positions. Larger gas E&Ps benefit from the same corporate and upstream synergies plus returns to scale in marketing. In the gas market, transportation and differentials comprise a much larger share of the commodity price, so marketing decisions move EBITDA more.

But realizing that benefit requires substantial fixed-cost investment in trading, scheduling, risk, and fundamentals, which only pays off at scale. This need for scale is even more acute in constrained basins like Appalachia, where intermittent pipeline constraints increase the extrinsic value of pipeline capacity, but only if the capacityholder treats it as an option rather than just flowing it every day.

Expand interim CEO Mike Wichterich targets a $0.20/MMBtu uplift in realized prices from marketing. And while that degree of improvement is unlikely, EQT’s 1Q26 performance shows how these investments can pay off.

Range’s unappetizing acquisition menu

Read more

As a result, future renewable capacity development will likely decelerate. Still, momentum from projects already under construction points to a record ~32 GW of solar capacity coming online this year, reshaping the daily and seasonal call on gas-fired generation. Backward-looking adjustments to the gas supply-demand balance don't capture this shift, which means the industry risks reading seasonal effects as structural — and structural effects as seasonal. The rising solar share of renewable capacity means the gas market will look more bearish than it is in late spring and early summer, and much tighter than it is around the winter solstice.

Wind and solar capacity outlook

The EIA’s 860M electric generator report includes a project-by-project inventory of operational and planned capacity. Typically, this report understates currently operational capacity in recent months, particularly for solar, but the “planned” category includes speculative projects that will never materialize. By analyzing historical versions of this report, I risk-weight each of the EIA’s development lifecycle categories and further discount less-advanced wind projects based on sector-specific headwinds.

Based on the current development pipeline and this risking approach, solar would account for ~53% of combined wind and solar capacity by year-end, up from ~33% when Congress passed the IRA in August 2022.

Figure 1 | Solar and wind capacity outlook

Seasonality of wind and solar generation

If solar and wind utilization rates were consistent day-to-day or at least month-to-month, their impact on gas burn would be straightforward to disaggregate. However, two factors — shown in Figure 2 — complicate this analysis:

1. Load shapes — especially for solar — are highly seasonal

2. Day-to-day volatility — especially for wind — means generation does not reliably follow these seasonal shapes

Read more

But extracting a real-time read on production requires properly accounting for weather impacts, and that normalization is getting harder with rising gas-fired generation.

Building a supply-demand balance

Isolating underlying structural trends requires first accounting for cyclical factors — both direct weather impacts on demand and other seasonal factors, such as LNG maintenance or nuclear refueling outages. For the former, we build a linear model of gas storage injections as a function of heating-degree days, cooling-degree days, and holidays1 over the 2021-25 period. The residual of this model is the gas supply-demand balance, an indicator of underlying market strength or weakness. Because each week’s observation is noisy, analysts typically smooth the time series with an exponential function.

Figure 1 | Gas supply-demand balance

But the resulting indicator still varies throughout the year because non-temperature-related seasonal factors systematically tighten or loosen the gas market at specific times, such as nuclear refueling outages, wind capacity factors, and LNG maintenance. To correct for this seasonality, we net off the five-year2 average supply-demand balance for that week to get a more neutral supply-demand balance.

Figure 2 | Seasonally adjusted gas supply-demand balance

How are changes in the gas market affecting the storage number?

The rise in gas-fired generation means the absolute seasonality of US gas demand is nonetheless increasing, even as the plurality of US gas demand growth has come from LNG exports. Weather-sensitive gas demand — shown in the teal shades in Figure 3 — grew 18 Bcfd from 2010-25.

Figure 3 | US gas demand growth and seasonal range

This growth in weather-sensitive consumption — particularly power generation — means that the impact of degree days on gas burn will increase over time. And if instead of modeling the impact of weather on weekly storage injections over the 2021-25 period, we use shorter, rolling horizons, we can see how the relationship between weather and demand has changed over time.

Here, we face a tradeoff between using the most recent data and having enough data points to reliably estimate coefficients. A two-year sample probably does not have sufficient resolution for a reliable model, but even a three-year rolling horizon shows a consistent increase in the impact of cooling degree days, offset by a higher constant, meaning higher injections in a mild week.

Interestingly, HDD coefficients have stabilized in recent years, even as rising power generation has added demand that is sensitive to cold weather. I expect the HDD coefficient is not rising because gas pipeline capacity — contracted by LDCs — limits gas-fired generation on the coldest days.

Figure 4 | Rolling horizon coefficients

The self-correcting seasonal adjustment partially masks this drift, but not fully, because both the model and its seasonal adjustment are backward-looking — we should expect today’s true CDD coefficients to be higher still.

Forward-looking implications

The US gas market grew 70% over the last 15 years, while US underground storage capacity remained essentially unchanged. EQT and others have rightly pointed out that the same storage inventory levels provide much less of a cushion against gas market dislocations from weather, outages, or other factors. Rising gas-fired generation, meanwhile, means that the call on gas storage is nonetheless increasing. The result, then, is that modest dislocations between supply and demand can quickly lead to extreme price outcomes.

Now, a colder-than-normal winter draws down storage inventories much more quickly; similarly, producers’ outrunning demand by a small percentage risks quickly creating a storage glut. So while the EIA’s gas storage number, released each Thursday morning at 9:30 CT, has always been the biggest needle-mover for NYMEX Henry Hub prices, it is now more important than it has been at any time since the pre-shale era, when supply and demand elasticity were much weaker.

The rise in weather-sensitive demand means that models built on lagging data — which is all of them! — will, in summer months, under-predict the impact of each CDD and therefore under-attribute demand to the weather and over-attribute it to structural strength (and conversely in the spring and fall). In other words, a hot July week looks structurally tighter than it really is, and a windy April one looser.

And while we’re mostly looking at the supply-demand balance for a read on production, understanding the level of gas-fired generation in a given week is different than isolating what that means for the future. In the next installment of this series, I’ll show how the rise in renewable generation overstates market looseness in the spring and tightness in the summer.

Against that backdrop, this earnings season confirms that flows out on these four routes are likely to decline further. Both new power plants under construction in Ohio and potential Ohio pipeline projects source gas downstream of the existing constraints. That means the new projects don’t increase market access for the core dry-gas Marcellus E&Ps best-positioned for growth. But the projects do create an opportunity for these producers, and their midstream counterparties, to develop smaller, intra-basin projects to bring core dry-gas volumes across the Ohio River to access what are likely to be much stronger prices.

Williams integrated power projects

In its earnings presentation, Williams expanded and extended its plans to develop and source gas for power plants in Ohio. The company is now building three power facilities, with a combined capacity of more than 1 GW. One of these projects is due online in each of 2026-28: Socrates (North & South), Apollo, and Socrates the Younger.

These power plants are far from Williams’ own Ohio gathering and processing assets near the Ohio River, so they’ll use gas from other systems. The Socrates plants coming online later this year will source gas from a combination of Columbia Gas (TCO) and Eastern Gas, and the Apollo plant will take gas from Nexus. In all three cases, though, the power plants source gas downstream of typical Northeast takeaway constraints on these systems.

Figure 1 | Williams proposed power projects and feedgas pipelines

Columbus Area Project

TC Energy announced that the non-binding open season for its planned 500 MMcfd Columbus Area Project on Columbia Gas (TCO) was oversubscribed by a factor of three. The project proposes moving gas to Ohio customers near Columbus — possibly including Williams for its power plants — from receipt points at Clarington, Majorsville, Waynesburg, TCO Pool, and Leach. The latter two would not source production directly. Instead, they’d pull volumes north that would otherwise flow south on Columbia Gulf. The former two are directly connected to supply but are unlikely to see strong production growth:

Read more

And while mild winter reduces demand just as much as cold weather increases it, the impact on prices is asymmetric. When Henry Hub prices fall below ~$2/MMBtu, prices at supply hubs in the mid-$1/MMBtu range approach short-run operational costs, precipitating producer shut-ins. But no similar feedback mechanism governs price upside.

The result, then, is that the distribution of realistic Henry Hub prices is right-skewed. But that right-skew in prices means that using collars to protect cash flows will increasingly weigh on E&Ps’ earnings, compared with not hedging, purchasing puts, or even traditional swaps.

How and how much are gas E&Ps hedging?

Read more

And this earnings season highlighted a maturing industry: consolidating into a single hub, bearing more fixed costs to reduce risk, and closing the door on projects that are no longer viable. It’s also one still prone to provocative — perhaps outlandish — goals and regional price dislocations.

No one wants to hire an oil and gas executive outside Houston

The consolidation of the US oil and gas industry into Houston-based companies dates back decades, but after supermajors ExxonMobil and Chevron moved their headquarters to Houston in 2023 and 2024, respectively, the trend is only accelerating. This earnings season, Devon announced its acquisition of Coterra and concomitant move to Houston, and a week later, Expand announced its own move to Houston.1

In regional-economy terms, Houston is becoming much more like Los Angeles than it is Dallas or Chicago or Atlanta. Which is to say, Houston absolutely attracts a disproportionate share of the top 1% US oil and gas industry talent, just as Los Angeles does in the entertainment industry. For a market leader like Diamondback, with a strong executive team in place, there’s no urgent need to move. But for smaller companies trying to make the leap to big-time, they have a better chance of making the key hires to do that in the place those people disproportionately live. I expect continued relocation — whether organically or inorganically — of public E&Ps headquartered in Denver, Midland, or Oklahoma City.

LNG projects are being canceled instead of remaining on life support

After Energy Transfer officially suspended development of its Lake Charles LNG facility in December, Sempra announced this earnings season that it had terminated the development agreement for Vista Pacifico LNG. Many proposed US LNG projects are going nowhere but have not yet been formally canceled. These two cases are interesting because the developers opted to formally pull the plug.

Perhaps the formal cancellations reflect these developers’ size and backlog depth. But usually, there’s more of an angle than that. For Energy Transfer, perhaps it’s that the economics of Lower-48 pipeline development are so much better than for US Gulf Coast liquefaction; for Sempra, perhaps about the regulatory climate for pipeline development through central Mexico.

Venture Global’s FT pivot

With its first two LNG projects, Venture Global took a minimalist approach to pipeline contracting, focusing on securing sufficient feedgas liquidity. In contrast, Cheniere contracts for pipeline capacity to multiple basins. After this strategy led to Plaquemines’ in-service bidding TGP 500L prices to as much as a $0.50/MMBtu premium to Henry Hub, Venture Global pivoted. For CP2, Venture Global is making two large investments to source Permian gas supplies:

• spending more than $1 billion to develop large-scale nitrogen removal units

• developing CP Express, partnering with WhiteWater to develop Blackfin, and contracting for 2 Bcfd of Permian capacity, presumably on Eiger Express

Contracting further upstream was clearly the right decision for a player of Venture Global’s size. However, Permian associated gas growth is likely to fall well short of takeaway capacity additions, narrowing Waha differentials sharply. Paying a levelized ~$1.25/MMBtu for the FT and NRU may end up much more expensive than committing to ~$1.50-1.75/MMBtu in FT to move gas from much weaker Appalachian hubs.

A $0.20/MMBtu improvement in price realizations?!

In the same press release announcing its move to Houston, Expand’s former CEO Nick Del’Osso “stepped down,” with board chairman Michael Wichterich appointed interim CEO. On Expand’s earnings call the next week, Wichterich announced a target of improving price realizations by $0.20/MMBtu. Analysts pressed for details, and Wichterich reiterated the target but offered few specifics.

Typical third-party marketing margins are a penny or two, so $0.20/MMBtu requires much more than insourcing. In a market that’s been deregulated for more than 30 years, there aren’t any free lunches — improving realizations more than a couple cents entails taking more risk, typically in the form of FT commitments or reduced hedging.

EQT’s approach was to own its wellhead gathering. This lets EQT shut in without the burden of fixed gathering commitments, and therefore hedge less, avoiding the vig paid to market-makers. But with Transco alone likely worth more than Expand, Expand can’t run the EQT play.2

Bigger picture, I don’t buy the construct of improving realizations by $X/MMBtu, regardless of what X is. On EQT’s earnings call, the company detailed its strategic decision to sell all of its February volumes at bidweek rather than take exposure into the month, based on its view that fundamentals did not support such strong futures prices. But EQT rightly did not extend this thinking to a promise of continued outsize realizations. Expand’s goal may be a headwind as it tries to staff up its marketing team, and the company probably needs to walk it back.

The coming Ohio gas shortfall

I previously wrote about the specific connectivity deficiencies that constrain various western Appalachia takeaway utilization, leaving ~900 MMcfd of spare capacity in aggregate. But across all four pipelines, the overarching issue is that the Utica in Ohio, and to a lesser extent the rich-gas Marcellus, are short inventory relative to the dry-gas Marcellus further east.

Against that backdrop, this earnings season suggests the situation is likely to get worse before it gets better: TC Energy announced that its Columbus Area Project and Crossroads expansion could each add 1.5 Bcfd, while Williams upsized its planned Ohio power projects. I’ll dive into the details in a future post, but the pre-2019 TCO Pool dislocation from other Appalachian hubs looks set to return.

On February 28, US and Israeli forces launched surprise airstrikes on several sites in Iran, killing Supreme Leader Ali Khamenei among other Iranian officials. This attack, along with subsequent Iranian counterstrikes, destabilized the region’s production and especially shipping routes, with attendant impacts on global energy and agricultural markets.

The degree of disruption in global oil and LNG markets, and by extension North American gas markets, depends on when and how the crisis is resolved. But regardless of the ultimate resolution, I’m reevaluating my view that it’s a bad time to be contracting for US LNG. Being an analyst means not just building a forecast but also thinking about how solid its foundations are. I underestimated the likelihood of war in the Middle East; for both LNG-market and oil-market reasons, instability there increases the option value of US LNG.

What are the war’s impacts on LNG markets?

Read more

Figure 1 | Spare western Appalachia takeaway capacity

Getting more Northeast gas into one or more of these pipelines is likely the cheapest and easiest way to grow Appalachian production in the near term. Range looks set to grow ~250 MMcfd into new Rover capacity, and new power plants will likely pull almost as much gas south on TGP. EOG could grow into ~150 MMcfd of unutilized Nexus capacity, but the largest tranche of unutilized takeaway — ~300 MMcfd on Columbia Gulf — does not yet offer a compelling risk-reward for producers.

Columbia Gulf: Spare capacity likely not held by producers

Read more

After no new Permian gas takeaway capacity was sanctioned between mid-2022 and mid-2024, an unprecedented 9.3 Bcfd of Permian gas pipelines took FID between July 2024 and August 2025. Remarkably, operators then upsized these commitments by a further 2 Bcfd in late 2025.

The first ~4.5 Bcfd of this capacity is due online later this year: ~550 MMcfd from a Gulf Coast Express expansion in the second quarter, along with 2.5 and 1.5 Bcfd from Blackcomb and Hugh Brinson pipelines in the fourth quarter. For E&Ps, it can’t come soon enough: gas takeaway capacity in the basin has been consistently constrained since March 2024, with only a brief period of basis relief following Matterhorn Express Pipeline’s in-service in winter 2024-25.

But this buildout is much larger than what would be needed to accommodate today’s Permian activity, or even a realistic high-drilling case. Either E&Ps will target more gas-weighted acreage or intervals, or Permian takeaway capacity will be consistently underutilized.

How does gas production compare to takeaway capacity currently?

Read more

Read more

But that assumption is startin…

Read more

Free subscribers will still receive occasional pieces, about once per month.

If my analysis has been useful, I hope you’ll subscribe. (If you prefer annual billing, email me at amber@measureddepth.com, and I’ll send an invoice.)

For teams of seven or more, I offer an enterprise license — please get in touch for details.

Finally, a couple of firms have reached out about a higher-touch relationship, including quarterly in-person presentations. If that’s of interest, please reach out to discuss further.

Building this independent analysis over the last four months has been some of the most fun and rewarding work I’ve done, and I hope it has resonated with you, too.

Amber

Read more

Read more

A former colleague texted me last week, amid the disruptions from winter storm Fern, to ask what I thought about freeze-offs. He had heard estimates he thought were too high but was skeptical. My answer to him, and to all of you here now, is that we really don’t know, and we won’t have any idea until we see the storage number for the week ending January 29.

What do we actually know

Let’s start with what we actually can observe in real-time.

Typical Permian gas flows

El Paso, TransWestern, NNG, and NGPL typically move1 ~6.5 TBtu/d, from a combination of direct connections to gathering systems and processing plants (shown in teal) and receipts from intrastate pipelines within the Permian (in pink). Since the final Matterhorn Express compressor station came online, it (in purple) and Whistler (in light blue) usually move ~2.5 TBtu/d to south Texas. Interstate receipts from Enterprise typically total an additional ~0.3 TBtu/d.

Figure 1 | Observable Permian scheduled gas volumes

Reduction in gas flows last week

During winter storm Fern, these typical ~9.4 TBtu/d of flows fell to a low of ~4.9 TBtu/d. As of Monday, February 2, ~22 TBtu of production had been lost in aggregate. I show these flows in Ttu/d rather than converting to MMcfd because, as I'll discuss below, the conversion factor itself is uncertain.

Figure 2 | Recent observable Permian scheduled gas volumes

Historical Permian wellhead production

While the EIA only reports state-level production, operators must report lease- and/or well-level production to state regulatory agencies, for tax purposes. Aggregating this well-level data, then, gives insight into Permian-specific production.

Figure 3 | Reported wellhead Permian gas production

What we don’t know

However, these two data sets are not directly comparable for three reasons: volume versus energy, state reporting lags, and intrastate flows.

How much ethane was extracted

The first challenge in grossing up scrapes to total Permian dry gas production is that the latter is never actually reported. Analysts typically model it using a fixed (or slowly moving) dry-to-wellhead gas ratio, but in reality, the extent of ethane extraction varies with market conditions. In both Texas and New Mexico,2 the EIA’s estimates of dry-to-marketed gas ratios have been trending down over time, but the frequent step-changes between the end of one year and the beginning of the next point to uncertainty even in EIA’s state-level dry gas production estimates.

Figure 4 | Texas and New Mexico dry gas share of marketed gas production

On January 26, gas quality at a range of Permian interstate locations ran about 2% hotter than earlier in the month, as operators extracted less ethane in response to much higher gas prices. For a basin producing ~30 Bcfd at the wellhead, this ethane rejection could offset more than 1 Bcfd of freeze-offs.

How complete state-reported production data is

I’ve built many a well-level production model, and while 95% reporting is more than sufficient to construct an accurate type curve for a given operator, county, or play, it’s nowhere near sufficient for real-time production estimation. In a ~30 Bcfd play, the difference between 94% reporting and 96% reporting is 500 MMcfd — and you have no way of knowing which one you’re looking at. In both Texas and New Mexico, some well- or lease-level reporting trickles in over the course of a year or more after first production.

Figure 5 | EIA and state-reported New Mexico gas production

How observed routes behaved relative to unobservable ones

Two increments of “missing” dry gas production — lease/plant fuel and local consumption — are small and predictable. But most of the gap between wellhead production and interstate nominations is gas that moves on intrastates, which is neither.

All of the post-2018 Permian intrastate pipelines — Gulf Coast Express, Permian Highway, Whistler, and Matterhorn Express — move gas to south Texas. We can reasonably assume that these pipelines’ 9.6 Bcfd of capacity is fully utilized in normal market conditions, given how weak Waha prices have been. The typical ~2.5 TBtu/d of Whistler and Matterhorn deliveries into south Texas therefore accounts for just ~25% of this capacity.

But the freeze-offs in west Texas (and elsewhere) last weekend changed the underlying economics of Permian gas flows. Waha prices traded over Houston Ship Channel, meaning that if anything, the reduction in Permian→south Texas flows on these new intrastates was modest. Given this price signal, it is possible that GCX and PHP shippers made even bigger changes, and that the actual observed deliveries on Matterhorn Express and Whistler constituted more than the typical share of flows — or even all the flows — along this route.

Implications for freeze-off estimates

Given these uncertainties, to estimate freeze-offs, we need to compare the two series we actually know: scheduled interstate pipeline volumes and wellhead gas production. Even putting these two series on the same chart requires an assumption about calorific value, which, in reality, varies over time depending on market conditions.

During winter storm Uri in February 2021, the observable share of Permian wellhead production increased by 1.7 percentage points relative to the surrounding six months, as shown in the gray line in Figure 6. Given that these disruptions lasted only 12 days and were severe for just five, the observable share of wellhead production likely rose 6-8 percentage points during the shut-ins. Conversely, freeze-offs in January 2024 were more evenly distributed, with the observable share of total wellhead production broadly consistent with the surrounding six months.

At the scale of today’s Permian production, whether this storm is more like Uri or like January 2024 cold weather could swing Permian freeze-off estimates by 3-4 Bcfd at peak, and 15-20 Bcf in aggregate. Uncertainty, due to state reporting lags and varying levels of ethane extraction, about what the baseline Permian dry gas production was before the storm widens the possible range by an additional 5-8 Bcf.

Figure 6 | Permian wellhead gas production and pipeline scrapes

In a basin like Appalachia, where more than 95% of gas flows on interstate pipelines, and where higher transportation costs mean ethane extraction doesn’t swing in and out of the money as frequently, pipeline scrapes provide a reliable real-time production estimate, even during market disruptions. In the Permian, though, the combination of slow production reporting, mostly intrastate flows, and significant month-to-month variation in ethane extraction means that scrape-based production models often provide more noise than signal.

This cold front sent gas demand forecasts and Henry Hub prices skyrocketing. As recently as January 16, the February NYMEX futures contract traded at just $3.16/MMBtu. Monday, it closed at $6.52/MMBtu. The dislocations were even larger in the January market, with Henry Hub cash printing ~$28/MMBtu for the January 24-26 weekend package and balance-of-month trading ~$20/MMBtu.

Gas liquefied this weekend would be sold no sooner than February in destination markets. TTF trades at ~€40/MWh for February delivery, equivalent to ~$14/MMBtu. At ~$28/MMBtu, then, US LNG was ~$14/MMBtu underwater on a variable-cost basis.

On Saturday — before the storm disrupted Haynesville production and likely precipitated additional involuntary cuts — feedgas reductions totaled ~3.8 Bcfd, with tollers cutting ~36%, Cheniere ~25%, and Venture Global hardly at all. The financial stakes are significant: I estimate Cheniere could have made as much as $80 million by unwinding its feedgas purchases over the weekend.

When Henry Hub price spikes push US LNG this far out of the money, liquefaction is no longer baseload demand but rather a financial option. But who holds that option — and can exercise it for tens of millions — varies by project.

Who holds the options?

In the first wave of US liquefaction, most projects had a commercial structure similar to an interstate gas pipeline, on which a shipper pays a monthly reservation charge in exchange for access to an increment of capacity. The shipper, though, is responsible for buying (or producing) the gas, scheduling the volumes, and then selling (or consuming) the gas. Freeport, Cameron, Elba, and Cove Point liquefaction contracts are structured similarly: the LNG offtaker pays a monthly reservation charge for access to the liquefaction capacity but is responsible for procuring/shipping the feedgas. These LNG offtakers maintain title to the gas through the value chain.

Cheniere’s projects were structured differently, with Cheniere responsible for feedgas sourcing, pipeline and storage contracting, and pipeline scheduling. Cheniere’s counterparties still pay a reservation charge, but at these projects, that is in exchange for the right to receive FOB LNG. Cheniere, instead, has title to the gas until after it is liquefied. Cheniere’s offtakers also pay 115% of the Henry Hub price for volumes they do lift, to cover losses in pipeline transportation and especially liquefaction.

In that first wave, feedgas procurement and transportation were straightforward for companies like BP (at Freeport) or Shell (at Elba), as these firms are among the largest US gas marketers. Other companies, most significantly Toshiba (at Freeport), underestimated the complexity of US gas marketing and ultimately exited US LNG. Not coincidentally, all second- and third-wave US LNG projects use the FOB model pioneered by Cheniere.1

At Cheniere’s facilities, LNG offtakers must exercise their option by the 20th of the month two prior to when they would receive the cargoes.2 That means offtakers made decisions about January cargoes by November 20, when January TTF traded ~$6/MMBtu over Henry Hub.

So while LNG tollers at Elba, Cove Point, Freeport, and Cameron could sell back gas purchases this weekend in response to changing market conditions, the window for Cheniere’s and Venture Global’s customers has long since passed. In the case of Plaquemines, although its final trains came online in December, the project is still pre-commercial, meaning that Venture Global remains the project’s only offtaker.3

But while Calcasieu Pass, Sabine Pass, and Corpus Christi offtakers no longer hold a January (or February) US LNG option, Cheniere and Venture Global themselves still do. Because their customers only take title to the LNG FOB, Cheniere or Venture Global can serve their customers from LNG storage rather than continuing to liquefy gas. They could then cover this short by fulfilling their own DES sales with ~$14/MMBtu LNG from elsewhere or, in a best-case scenario, by refilling LNG storage with ~$6/MMBtu February gas.

Who’s exercising their options?

By analyzing LNG transportation portfolios and pipeline nomination data, I estimate which US LNG option-holders unwound their gas purchases this weekend. This exercise isn’t completely straightforward — some Freeport and Cameron tollers have overlapping feedgas portfolios; some intrastate pipelines serve Corpus Christi and Freeport; Cove Point Pipeline serves other load besides the liquefaction facility — but nonetheless, the results are directionally informative.

In gas markets, the weekend trades on Friday as a Saturday-Sunday-Monday package. Companies can and do make single-day deals, but typically these reflect operational disruptions, such as cuts due to Haynesville production freeze-offs. On Saturday, before Haynesville production declined sharply, LNG feedgas reductions totaled ~3.8 Bcfd, with tollers (in aggregate) and Cheniere cutting much more aggressively than Venture Global. The latter’s reduction came almost entirely at Plaquemines, where it does not yet have customer obligations.

Figure 1 | January 24 feedgas reductions

As long as upstream cuts on Sunday and Monday were not significant relative to the total feedgas reduction, the financial impact of Cheniere’s operational changes is likely to be material to the company. In 3Q25, the company reported $1.6 billion in adjusted EBITDA, whereas I estimate that Cheniere could have made as much as $80 million by unwinding its feedgas purchases over the weekend and backfilling these sales with $14 LNG. If the company has sufficient storage flexibility to replace these sold volumes with February (or March) Henry Hub gas at $6 (or $4), the impact would rise to ~$140 (or ~$150) million.

Venture Global, in contrast, made much smaller changes, perhaps due to broader corporate considerations. If the company were to change Plaquemines operations in response to market conditions, it may risk jeopardizing its claim that the facility is pre-commercial. In other words, the value of Henry Hub gas resold this weekend was very high, but nowhere near as high as the aggregate value of all pre-commercial Plaquemines cargoes. Shell and Venture Global have been locked in a lengthy legal dispute about when Calcasieu Pass was commercial, and Shell is also a Plaquemines offtaker.

Among tollers, Shell made the biggest change, reducing Elba feedgas to zero. Despite Elba’s small 2.5 MMtpa size, the change could have been worth as much as ~$65 million to Shell, because Transco Zone 5 traded at more than $60/MMBtu this weekend. At Cameron, feedgas reductions varied by pipeline, with the largest cuts on routes where Total holds the most capacity — though differences in upstream freeze-offs may also explain the pattern.

Figure 2 | Estimated LNG option impact on earnings

How long will this last?

At the four tolling facilities, reduced feedgas is likely to last as long as Henry Hub prices stay higher than those for LNG — at least through the end of the month, based on the latest forecasts. If anything, tollers’ feedgas nominations are likely to continue to decline if prices remain high, as operational factors likely limit how quickly and efficiently a liquefaction plant can ramp down. Inasmuch as US LNG tollers have already on-sold their US cargoes, they could instead procure LNG from elsewhere and pocket the difference between Henry Hub and LNG prices. (Of course, reduced US LNG output is likely to bid up February global gas and LNG prices.)

Cheniere’s ability to maintain FOB LNG deliveries to its customers with lower feedgas intake will dwindle quickly through the week. Most of Cheniere’s customers take LNG FOB, so by week’s end, Cheniere likely will need to renegotiate with them — and surely share the value of its now-deeply-in-the-money option. But with ~$6/MMBtu of value at stake and some of the largest LNG marketers as counterparties, unwinding this option is likely to continue accruing profits for Cheniere.

Applying that fra…

Read more

Read more

Read more

About 10 years ago, I gave a presentation to a large utility’s board, and afterward, one of the directors asked me if I had been doing this for a long time. Thinking about his question crystallized something for me — that I’ve always been an analyst. In kindergarten, I checked the Rangers and Mavericks box scores and standings in the Dallas Morning News every morning.

I’m writing this because methodology matters, especially when markets are noisy and data are incomplete.

Conviction and uncertainty

Although I’ve been an analyst for a long time, a book I read this summer about religion, of all things, helped me articulate how I approach my work as an analyst. In The Road to Wisdom, Francis Collins writes about a web of beliefs, wherein the believer is most sure of (and attached to) the views at the center, but less so to ones further out.

And this resonated with me because I think the core job of being an analyst is deciding when to change your mind: thinking not just about models but about heuristics and how well-suited your tools are to capture changing market dynamics, about where models might not be accurate even though they’re precise, and updating your priors accordingly. Do it too infrequently, and the market passes you by; do it too often, and no one will take you seriously. It’s essential to be transparent not just about your analysis but about how strongly you hold it, which requires both methodological rigor and epistemic humility.

Figure 1 | Views about the gas market

Heuristics versus models

Some analysts — typically older and oil-focused — eschew models, arguing that the job is about more than pressing run and that understanding how companies (and sometimes countries) make decisions matters more. It probably goes without saying that I do not belong to this school. I’m an unabashed model lover, not because they are magic, but because they are tools for rationalizing how disparate pieces fit together, in ways that aren’t obvious when looking at problems in isolation. But I nonetheless think the OPEC-watcher’s critique has some merit.

The most important thing about models is to understand what’s not in them. In the 2010s, our Henry Hub forecasts were wrong because they didn’t account for future breakeven improvements. More broadly, you need to consistently ask yourself not just whether the model’s input data are robust but whether the model’s algorithm needs to change to accommodate new features of the market.

For example, Permian gas pipeline development is accelerating and is likely to outpace the growth in associated gas production. That, in turn, will narrow basis differentials. But what would those narrower basis differentials, in conjunction with higher Henry Hub prices, mean for upstream activity and therefore the pace of production growth? Before you can consider using a model, you need to have a heuristic.

At Wood Mackenzie, I spent 10 years using one linear programming model to forecast North American gas flows and basis differentials; another to forecast global LNG development, flows, and prices; and others to optimize power systems. Next, I went to RS Energy Group, where we had a best-in-class understanding of onshore North America upstream but no long-term activity model, to say nothing of a gas pipeline flow model. While I prefer having those tools, a strong mental model is more important. Because at the end of the day, having a model isn’t about updating the numbers and pressing run — it’s about understanding what the model does and what it doesn’t, and how to rework it as the market changes.

In practice, I can tell you where specific compressor stations are on Transco and TGP, but I can’t rattle off all the regas terminals in Japan. So when I write about sustainable long-term global gas prices, or clearing the medium-term LNG oversupply, I’m using a heuristic rather than detailed dispatch or flow models.

Accuracy versus precision

Sometimes, though, the problem isn’t the heuristic but the availability of data. My econometrics professor told the joke about the econometrician looking for a lost ring under the lamppost, not because that’s where he lost it, but because that’s where the light was. And I think of this often with respect to models of gas production scrapes, whose problems are twofold:

• First, the independent variable is itself unknown: operators report well-level two-stream production to state regulatory agencies, but they do not report three-stream volumes. The EIA reports dry gas production only at the state level, and even then with a three-month lag. In the Permian, wellhead NGL content and ethane extraction both vary widely, meaning no one really knows what gas production was, even historically.

• Second, the relationship between observed and unobserved gas production changes frequently. With scheduled quantity data reported for interstate gas pipelines, you can observe how much production moves onto interstates in real-time. But modeling total gas production from the volume moving onto interstates requires that relationship to be stable, which it is not.

Recently, the in-service of Louisiana Energy Gateway re-routed flows from less-visible intrastates.1 Applying the same gross-up factor resulted in Haynesville production being substantially overstated until analysts recalibrated their models. But even that recalibration doesn’t solve the problem, because you don’t have any way of knowing how much gas was redirected versus newly turned in line. You can make an assumption, but it’s that view that’s driving the model results, not the observable data.

If growth were mostly coming from Appalachia or the Rockies, production scrapes would probably have more signal value than noise. But growth is overwhelmingly coming from the Haynesville and the Permian, where intrastates move a large and growing share of production, and NGLs make historical dry-gas production estimates uncertain. With infrastructure as large as LEG, these model deficiencies are obvious. But gas flows change not only with new pipelines like LEG but also with new gathering interconnects, processing outages, or shifts in destination prices.

So when someone asks me to explain why gas production has unexpectedly risen or fallen, my answer is often to question whether it has indeed.

How I reevaluate when evidence changes

I make a good conference panelist because I’m opinionated2 and not afraid to express views outside the consensus. With this framework, I tend to grow less certain before eventually changing my mind. This year, though, I doubted CP2 FID until the moment it happened:

• First, I weighed soft factors — like the ongoing lawsuits and arbitration between Venture Global and its Calcasieu Pass offtakers — too heavily relative to cold hard dollars: Venture Global offered the lowest capacity charges.

• Second, I held too much conviction relative to how connected I was with LNG developers and offtakers. SPAs are public, but tracking those agreements isn’t enough to call which projects will proceed. Pointing out that some of CP2’s SPAs might have lapsed was appropriate; extending that to FID risk was not.

For me, good analysis is less about never being wrong than about knowing when to become less certain. With that in mind, don’t hesitate to comment or e-mail when your analysis suggests I should be changing mine. And in the meantime, merry Christmas!

Subscribe now