An LNG cargo leaving the US Gulf Coast has a choice. It can sail east to Europe in about eleven days, or it can sail to Japan — and in 2026, with the Strait of Hormuz effectively closed, that means going the long way round the Cape of Good Hope, thirty-three days at sea.
The cargo goes wherever it earns more. That sounds obvious. What is not obvious, and what trips up most people looking at this market for the first time, is that the higher price does not automatically win. Asia is almost always more expensive than Europe. That does not mean cargoes always go to Asia. What matters is whether Asia is more expensive by more than the extra cost of getting there.
That difference — the spread net of the cost of delivery — is the arbitrage. This piece explains how it is calculated, what closes it, and what happened to it in 2026.
Two basins, one fleet
The world LNG market is usually described as two basins. It is a useful simplification, and worth stating precisely because the whole trade rests on it.
The Atlantic basin is the United States, Trinidad, Nigeria, Algeria, Norway and Russia’s Yamal on the supply side, selling into Europe, the UK, Turkey and South America. Its reference price is TTF, the Dutch virtual trading point, which is a pipeline hub price that LNG must compete with rather than an LNG price as such.
The Pacific basin is Australia, Malaysia, Indonesia, Brunei, Papua New Guinea, Russia’s Sakhalin and now Canada’s west coast, selling into Japan, China, Korea, Taiwan and South-East Asia. Its reference price is JKM — the Japan–Korea Marker, an assessment of spot cargoes delivered into North-East Asia.
Between them sits the Middle East — Qatar above all — which physically ships into both basins through the Strait of Hormuz. That is why a single military event in one waterway repriced gas on two continents at once.
And connecting all of it is one global fleet of roughly 700 ships. A vessel does not belong to a basin. It goes where the freight rate is best, which is why freight is not a cost you can treat as a constant.
Picture the world gas market as two large neighbourhoods sharing one fleet of delivery vans. Europe is one, Asia is the other, and Qatar sits between them delivering to both. The vans do not belong to a neighbourhood — they go wherever the delivery fee is best. That is why one blocked road changed prices on two continents at the same time.
The arithmetic
Here is the calculation an LNG desk runs several times a day. The numbers below are illustrative but sit within 2026 market ranges, and every step is shown so you can substitute your own.
Take a standard cargo of about 3.6 trillion British thermal units (roughly a 174,000 cubic metre vessel), lifted free-on-board from the US Gulf.
Cargo-size conventions vary — depending on vessel and heating value you will see anything from 3.4 to 3.8 TBtu quoted for the same class of ship. It matters less than you would think. Charter and port costs per MMBtu scale inversely with cargo size, while boil-off does not (it is a percentage of the cargo either way). Rerun the whole calculation at 3.4 TBtu and the threshold below moves from USD 1.82 to about USD 1.89 — roughly 4% on the number that decides the trade. Use your own vessel’s figure, but do not expect it to flip the answer.
That FOB cost is the same whichever way the ship turns. Everything that follows is the cost of the voyage.
Route A — US Gulf to Northwest Europe
Route B — US Gulf to Japan, via the Cape of Good Hope
Subtract one from the other and you have the number that actually governs the trade:
A US cargo sails to Asia only when JKM exceeds TTF by more than the freight differential — here, about USD 1.82 per MMBtu. Below that, Europe wins even though Asia pays more. The arbitrage is "open" above the threshold and "shut" below it.
Every one of those inputs moves. Double the charter rate and the threshold roughly doubles. Reopen Hormuz and the Asian voyage shortens by a fortnight, and the threshold collapses. That is why LNG traders watch freight and canal transits as closely as they watch gas prices.
At a USD 5 per MMBtu JKM premium — which is where the spread went at the height of the summer disruption — the trade nets USD 3.18 per MMBtu, or about USD 11.4 million on a single cargo. At today’s near-parity between JKM and TTF, the same voyage loses money and the cargo discharges in Europe.
When the arbitrage opens
JKM minus TTF, against the cost of the longer voyage
Indicative spread levels through 2026 against a USD 1.82 threshold calculated at an USD 80,000/day charter rate. In the crisis weeks charter rates themselves spiked toward USD 300,000/day, which pushed the threshold up even as the spread widened — the two moved together, which is exactly why the arb is harder to capture than it looks.
A cargo does not sail to whoever pays the most. It sails to whoever pays the most after you subtract the cost of getting there. Asia nearly always pays more than Europe — but Asia is also far further away, so unless Asia's extra payment beats the extra shipping bill, the ship turns towards Europe. Here that break-even is about USD 1.82 per unit of gas.
What 2026 did to the price signal
Three benchmarks, three completely different years.
Three benchmarks, three different years
USD per MMBtu at three points in 2026. TTF converted at €1 = USD 1.16 and 1 MWh = 3.412 MMBtu
Asia and Europe converged as both competed for the same shrinking pool of flexible cargoes. Henry Hub barely moved. That gap — roughly USD 19 per MMBtu between US wellhead and Asian delivered — is the entire economic case for the US export build-out. Sources: IEA Gas Market Report Q3-2026, EIA Short-Term Energy Outlook, market assessments.
Note what did not happen. Henry Hub sat between USD 2.87 and USD 3.44 all year, because the United States is a closed pipeline system with abundant domestic supply and a liquefaction bottleneck. American gas is only loosely connected to the world price, and stays that way until export capacity outgrows the resource base. This is why a US producer hedging in Henry Hub is not hedged against a global gas event at all.
American gas stayed cheap all year while European and Asian gas roughly doubled. That is not an error. America has plenty of gas and only a limited number of plants that can chill it into liquid for export, so its price sits behind a fence. The height of that fence — about USD 19 — is precisely why so many export terminals are being built.
The physical picture behind the prices
Prices are the summary. Underneath them is a straightforward supply-and-demand story, and 2026 was a year in which supply broke and demand rationed itself.
| Basin | Market | 2026 position |
|---|---|---|
| Atlantic supply | United States | 17.4 Bcf/d of LNG exports forecast for 2026, rising to 18.6 Bcf/d in 2027. Corpus Christi Stage 3 and the first two Golden Pass trains ramping this year. |
| Trinidad, Nigeria, Algeria | Legacy suppliers, broadly flat, feedstock-constrained. | |
| Norway | Pipeline plus Hammerfest LNG; Europe's most reliable non-LNG lever. | |
| Russia (Yamal) | Still flowing to Europe and Asia under sanctions carve-outs; politically fragile. | |
| Atlantic demand | European Union | Gas demand down more than 2% on high prices and renewables growth — but storage only 63% full against a 79% five-year average, the lowest in fifteen years outside 2021. |
| Germany / Netherlands | 51% and 44.3% storage respectively. The weakest points in the European system going into winter. | |
| South America | Brazil and Argentina buy on hydrology; a dry year in Brazil competes directly with Europe. | |
| The swing | Qatar and UAE | Combined LNG exports forecast down 45% year on year — some 54 bcm — following the Hormuz disruption and damage at Ras Laffan. Supplies both basins, so the loss is felt on both sides. |
| Global fleet | Roughly 700 vessels. Longer voyages absorb ships, tightening freight independently of cargo availability. | |
| Pacific demand | Asia Pacific total | 257 Mt forecast for 2026, down 4.1% from 268 Mt and well below the 2024 peak of 278 Mt. A second consecutive annual decline. |
| China | 62.4 Mt, down from 66.4 Mt. Regasification terminals running at 29% of nameplate. Holds the region's largest inventory buffer and the most diversified portfolio — which makes it the marginal buyer that can simply walk away. | |
| Japan | More than 90% of 2026 demand covered by term contracts. Structurally the least exposed to spot. | |
| South Korea | Spot exposure above 20%, with two 2 mtpa contracts tied to the damaged Ras Laffan Train 6. The most exposed of the large Asian buyers. | |
| India | The largest South Asian exposure by volume, with potential curtailments of up to 1.5 Mt a month forcing industrial fuel-switching and fertiliser production cuts. India is the classic price-elastic buyer: above roughly USD 12–14 per MMBtu, demand destruction starts. |
The pattern is worth naming. Europe’s demand is inelastic and its storage obligation is legal; Asia’s demand is elastic and its buyers can switch to coal. So when the market tightens, Asia rations by price and Europe pays. That is why the JKM–TTF spread compressed toward parity in August despite Asia being nominally short — Asian buyers stepped back, and Europe, needing to fill storage before winter, kept bidding.
Europe and Asia reacted to expensive gas in opposite ways. Asian buyers, who can switch fuels and sit on stock, simply bought less. Europe could not — its storage caverns have to be filled before winter, and that is a legal obligation, not a preference. When one buyer can walk away and the other cannot, the one who cannot ends up paying.
The ceiling nobody quotes
There is a limit on how far JKM can run, and it is not set by the gas market at all. It is set by what an Asian buyer could burn instead.
Gas competes with oil products and coal on energy content, not on volume. To compare them you convert everything to dollars per million British thermal units: divide a crude price by roughly 5.8, a fuel oil price by about 6.3, and a coal price by about 23.8 per tonne.
| Fuel | Price | Energy content | $/MMBtu | Against JKM at $22.61 |
|---|---|---|---|---|
| LNG, JKM | $22.61 | — | 22.61 | the reference |
| Gasoil / diesel | $175/bbl | 5.83 MMBtu/bbl | 30.02 | 33% dearer — no switching |
| Brent crude | $92/bbl | 5.8 MMBtu/bbl | 15.86 | 30% cheaper |
| High-sulphur fuel oil | $75/bbl | 6.29 MMBtu/bbl | 11.92 | 47% cheaper |
| Thermal coal, API2 | $130/t | 23.8 MMBtu/t | 5.46 | 76% cheaper |
Read the bottom three rows and the whole of Asian demand behaviour in 2026 becomes obvious. At USD 22.61, LNG costs nearly twice what fuel oil costs and more than four times what coal costs for the same heat. Any buyer with a dual-fired boiler, an idle coal unit or a flexible industrial process stopped buying gas.
That is what the demand numbers are describing. Asia Pacific LNG demand fell 4.1%. Chinese regasification terminals ran at 29% of nameplate. Indian buyers faced curtailments of up to 1.5 million tonnes a month and cut fertiliser production. None of that is a forecast about the energy transition — it is arithmetic, executed by plant managers with a calculator.
The switching level is a soft ceiling on JKM. Push Asian gas far above the price of the fuel next door and the buyers simply leave, which is exactly what compressed the JKM–TTF spread toward parity in August even though Asia was nominally short. Europe has no equivalent escape route — it has coal-plant closures behind it and a legal storage-filling obligation in front of it — so Europe becomes the price setter by default.
Note the one row that does not switch. Gasoil at USD 30 per MMBtu is dearer than LNG, so nobody moves from gas to diesel. Switching runs downward into the cheap, dirty end of the barrel, never upward.
And there is a second-order effect that connects this article to the next one. Every tonne of gas demand that switches into fuel oil or coal becomes additional demand for those fuels. In a year when refining capacity was already scarce, that switched demand landed on top of an overloaded product market — one of the underappreciated reasons Atlantic Basin refining margins reached all-time highs. See Crack Spreads Across the Barrel for what happened at the other end of that chain.
Gas has an invisible ceiling. If it becomes dearer than the oil or coal a factory could burn instead, the factory switches and stops buying gas. Asian gas currently costs about four times what coal costs for the same heat — so the buyers left. That, not sentiment, is why Asia stopped competing for cargoes.
What closes the arbitrage
Five things, in roughly the order a desk worries about them.
Freight
The fastest-moving input. Atlantic spot charter rates opened 2026 near USD 26,250 a day and spiked toward USD 300,000 a day during the Hormuz scare — a 600% move in a week. Every dollar on the charter rate raises the threshold the spread must clear.
Route availability
Suez plus Hormuz is the short way east. With Hormuz disrupted, cargoes go via the Cape, adding roughly three weeks round trip. The Panama Canal is an alternative for US Gulf cargoes but LNG carriers compete for slots with container ships and lose on willingness to pay.
Boil-off
An LNG carrier loses roughly 0.10% of cargo a day to evaporation. On an eleven-day run that is a rounding error. On thirty-three days it is 3.3% of the cargo — and at USD 22 per MMBtu, 3.3% is a real number.
Destination flexibility
Only cargoes with free destination rights can be diverted at all. Most Qatari long-term contracts historically carried destination restrictions; most US contracts do not. The US export model exists precisely to sell flexibility.
Regasification access
A netback assumes you can discharge. Terminal slots, send-out capacity and pipeline entry charges differ by country, and in a tight winter the binding constraint can be a berth rather than a price.
The cancellation option
A US offtaker paying a fixed tolling fee can decline to lift. When delivered prices fall below feed gas plus shipping, cancelling is rational — the tolling fee is sunk either way. This puts a soft floor under global prices at roughly the US variable cost of export.
Six things can kill this trade, and five have nothing to do with the gas price: what the ship costs, how far it must sail, how much cargo evaporates on the way, whether the contract even permits changing destination, and whether there is a free berth at the far end.
How this is traded and hedged
The spread itself is tradeable. JKM and TTF both have listed futures, so a desk can take a view on the basin differential without owning a molecule — long JKM, short TTF, in matched volumes. That is a pure spread position and it is how most financial participants express a basin view.
A physical trader’s problem is harder, because the position has three legs that do not settle together: a gas price, a freight rate, and time. A cargo loaded today and delivered to Asia in five weeks is exposed to whatever JKM does over those five weeks, whatever the charter market does if the voyage extends, and whatever the currency does if the sale is priced in a different unit. Hedging the price leg while leaving freight unhedged is a common and expensive mistake — in 2026 the freight leg moved further in percentage terms than the commodity.
One. The arbitrage is a spread net of delivery cost, not a price comparison. Asia is nearly always dearer; that alone tells you nothing.
Two. The threshold and the spread move together in a crisis. Freight spikes when cargoes are being diverted, which is precisely when the spread widens — so the visible opportunity is usually smaller than the headline spread implies.
Three. Henry Hub is not a global price. A US producer hedged in Henry Hub carries the entire international basis unhedged.
If you sell gas in America and protect yourself using the American gas price, you were not protected at all this year — that price barely moved while everything else doubled. Insure against the thing you actually sell, not the thing that is easiest to insure.
Looking into 2027
The IEA expects markets to stay tighter than previously assumed for the next two years, with cumulative supply losses of around 140 bcm over 2026–2030. But the new-build wave is still coming: Port Arthur Phase 1 and Rio Grande Trains 1 and 2 add roughly 3 Bcf/d of US capacity in 2027, and Wood Mackenzie expects Asian demand to recover to 279 Mt in 2027 and 297 Mt by 2028.
That sets up the interesting question. If Gulf supply returns on schedule and the 2027 capacity lands and Asian demand only recovers to its 2024 level, the market flips from scarcity to surplus quickly. In a surplus, the JKM–TTF spread narrows toward the freight differential and stays there, because the marginal cargo is always indifferent between basins. The wide, volatile spread of 2026 is a feature of a broken market, not a normal one.
The wild price gaps between Europe and Asia in 2026 are a symptom of a broken market, not a normal one. When the missing supply returns and the new export plants start up, those gaps should shrink back to roughly the cost of the extra shipping and stay there. The volatility is the anomaly, not the baseline.
Sources and further reading
- Gas Market Report, Q3-2026 — International Energy Agency
- U.S. natural gas exports to grow nearly 30% by 2027 — U.S. Energy Information Administration
- Short-Term Energy Outlook, August 2026 — U.S. Energy Information Administration
- Asian LNG demand to decline for a second consecutive year — Wood Mackenzie
- The EU will need higher LNG imports to refill gas storage ahead of winter — ACER