Markets · Research note
The Glut That Did Not Arrive: Global Gas After Ras Laffan
Everyone underwrote an LNG surplus for 2027. Two destroyed liquefaction trains and a delayed expansion pushed it to the end of the decade, and contracts written for the old schedule are now mispriced.
For most of the last three years the consensus view on liquefied natural gas was that a large surplus was coming. Projects sanctioned after 2022 were due to complete around 2027, supply would run ahead of demand, and prices would fall toward the cost of US production; buyers signed contracts on that basis, European regulators wrote storage policy on it, and traders positioned for it.
The surplus has been postponed by roughly two years. This piece argues that the postponement is a repricing rather than a delay. The capacity involved was destroyed rather than deferred, and the response of buyers to a second supply scare within five years is unlikely to reverse when the tonnes eventually arrive.
What was actually lost
Figure 1 shows what changed at a single site. Iranian strikes on Qatar’s Ras Laffan Industrial City on 18 March 2026 hit Trains 4 and 6. Both were written off rather than repaired, removing 12.8 million tonnes per annum and about 17% of Qatari export capacity; the South site fell from 36 MTPA to 24 MTPA, and the return-to-service window runs three to five years.
That was not the whole disruption: QatarEnergy halted all LNG production on 2 March and declared force majeure on 4 March, taking roughly 80 MTPA, close to a fifth of global supply, out of the market at once. QatarEnergy put the damage at around $20 billion of lost annual revenue.
The distinction between the two events is the one that matters for the market: a halt reverses, whereas a written-off train does not, and the expansion that was supposed to replace it has slipped. Qatar’s North Field East start has moved to the fourth quarter of 2026, with reports suggesting 2027.
Figure 2 sets the losses against what is still arriving. Capacity from elsewhere adds up to 99 MTPA by 2027, principally from US projects including Plaquemines and Golden Pass. Combined delays to North Field and Ruwais plus damage to existing trains cut 2028 additions by around 28 MTPA.
Note that global supply still grows: the market is not short, it is simply less long than it was underwritten to be, and the difference between those two statements is where the mispricing sits.
The surplus moves to the end of the decade
The widely modelled 2027 to 2028 glut has shifted to 2029 to 2030. Figure 3 shows this schematically rather than numerically, because the underlying quantity is a balance between supply and demand forecasts that no single published series captures.
A two year shift sounds modest, and in most markets it would be. In a market where the standard contract runs ten to twenty years and a liquefaction project takes four to six years to build, two years is most of the window in which a buyer decides whether to sign long or wait for spot.
Figure 4 is included as a record of what the market believed rather than as a forecast. Morgan Stanley expected European and Asian prices below $10 per MMBtu by the fourth quarter of 2026 against an average near $14 last winter, and BNP Paribas saw $8 in 2027. Both were published before the attack and both assumed the surplus arriving on schedule. Whether they were wrong depends entirely on an event neither could have modelled, which is a useful reminder about what commodity forecasts are actually forecasting.
| Assumed before March 2026 | Position after | |
|---|---|---|
| Qatari capacity | Rising through North Field expansion | 12.8 MTPA written off, expansion delayed to Q4 2026 or 2027 |
| Global surplus | Arriving 2027 to 2028 | Arriving 2029 to 2030 |
| 2028 capacity additions | Baseline | Reduced by roughly 28 MTPA |
| European price path | Below $10/MMBtu by Q4 2026 | Forecast predates the disruption |
| Buyer posture | Wait for spot, resist long contracts | Reconsidering contract cover |
Sources: Wood Mackenzie, Morgan Stanley, BNP Paribas, IEA Global LNG Capacity Tracker, QatarEnergy.
Why the behaviour will outlast the outage
The physical shortfall is temporary in the sense that new capacity eventually replaces it; the change in buyer behaviour may not be.
European and Asian importers have now experienced two supply shocks in five years, from opposite directions. The first removed piped Russian gas and forced a rebuild of import infrastructure around LNG; the second demonstrated that the LNG supply chain has a concentrated chokepoint of its own, because a large share of the world’s flexible cargoes originates from a single industrial complex within missile range of a contested waterway.
Hence the lesson a procurement committee takes from 2026 is not about price; it is about the correlation between the two systems they were told to treat as alternatives. A buyer that diversified from pipeline to LNG found that both routes concentrate at a few physical points.
Furthermore, the security-of-supply argument now has a permanent line in the budget rather than a temporary one. Contracting long at a price above the expected spot path is expensive, and it is defensible only if the buyer believes the supply curve is less reliable than the model says. Two events in five years is a small sample; it is also the entire sample most procurement teams have.
Where this reading could be wrong
The strongest counterargument is that 99 MTPA of new capacity by 2027 is a great deal of gas, and that a two year delay to a surplus is still a surplus. If US projects complete on schedule and demand growth in Asia disappoints, prices fall regardless of Qatar, and the repricing described here reverses within a single contracting cycle. Delays to major energy projects are also routine, so a two year shift sits inside the historical distribution of construction slippage rather than outside it.
Furthermore, the behavioural claim is the weakest part of this piece and the hardest to test. It rests on an assertion about how buyers respond to two shocks, and buyers responded to the 2022 shock by signing long and then, within eighteen months, by trying to get out of what they had signed. There is a reasonable case that procurement memory is shorter than the argument here assumes.
Finally, Figure 3 is schematic. It represents the direction and approximate timing of a shift in consensus, and it should not be read as a quantity. Anyone wanting a number should go to the underlying balances instead.
Conclusion
The 2026 attack on Ras Laffan did something unusual for an energy disruption: it destroyed capacity rather than interrupting it. Trains 4 and 6 are not coming back this decade. The expansion meant to sit on top of them has slipped past the year it was needed.
The result is not a shortage: supply still grows, and 99 MTPA arrives from elsewhere by 2027. What has gone is the assumption that made the last three years of contracting decisions rational, which was that a buyer could wait, stay short, and pick up cheap cargoes in 2027.
That assumption has been moved to 2029, and it has been moved by an event that also demonstrated why waiting is risky. Both halves of that sentence will show up in the contracts signed over the next eighteen months, and the contracts are what determine the price long after the trains are rebuilt.