The buffer is thin in two places at once: European storage at 68.5% while the Asian spot climbs to 30 dollars

The storage gap: a volume problem, a price symptom

As Europe closes the injection season it finds its underground storage 68.5% full; according to TradingPedia's report of 17 September 2026 that level is roughly 16 points below the seasonal average. On the same day two sources diverge on where the TTF front-month contract traded: TradingEconomics reports 78.60 euros/MWh and a daily rise of 0.73%, while TradingPedia reports a fall of 1.23% to about 76.98 euros/MWh and a one-week low. The 1.62 euro difference between them may stem from the intraday trading hour; it could not be independently verified and has been recorded here as a contradiction.

This one-day fall in the price is not a structural easing. What produced it was the end of a 24-hour strike at Dunkirk, France's largest LNG import terminal, which had cut daily capacity from at least 9.4 GWh to 4 GWh: the one-day return of a single terminal. The source of the gap, meanwhile, remains in place; with the Strait of Hormuz largely closed to commercial shipping, Qatari LNG exports are constrained, and planned maintenance at Norwegian facilities is reducing pipeline supply. Fullness is a measure of stock, while TTF is the cost of replenishing that stock: a stock 16 points below the seasonal average means fewer days of buffer going into winter, wherever the price stands.

The same cargo, two buyers: the measure of competition is not price

The only quick way to close Europe's storage gap is spot LNG. But there is a queue of demand for the same cargoes in Asia, and that queue is culled not by price but by ability to pay. The Asian spot LNG price rose this week from a pre-conflict floor of about 10 dollars/mmBtu to roughly 30 dollars/mmBtu; it was the year's second price jump. On Shell's estimate, some 36 million tonnes of LNG supply from the Middle East was lost in 2026. That lost volume forces Europe and South Asia to draw from the same pool.

The outcome has to be read not in the price curve but in industrial output. Bangladesh, which generates more than 40% of its electricity from imported LNG and used to source 95% of those imports from Qatar, has in effect been out of the market at a spot level of 30 dollars. According to BKMEA's survey of the knitwear industry, 78% of factories have partly halted production and 55% reported that buyers had cancelled or cut orders since the end of August. At one factory an order of 50,000 pieces was reduced to 40,000, and meeting the delivery date produced an air freight cost of 50,000 dollars on a single shipment. Bangladesh's energy ministry confirmed that industrial output was slowing.

On the Pakistani side the picture is tighter still. Against a power sector gas requirement running to 400 million cubic feet a day through the winter, only two LNG cargoes were confirmed for September; a standard cargo is about 140,000 cubic metres, that is 3 billion cubic feet, and corresponds to roughly a week of supply. When petrol reached 391 rupees (1.41 dollars) and diesel 421 rupees (1.52 dollars), the government turned to a subsidy of 100 rupees a litre and to closing bazaars at 21.00. These are not a price response but instruments of rationing.

This picture may look like good news for Europe: as a rival withdraws, cargoes become easier to find. But that is not how a buffer is measured. The withdrawal of Bangladesh and Pakistan shows not that supply has increased but that price has begun to ration supply. Once rationing has started, the cost of closing Europe's gap no longer rests on the most elastic part of the demand curve but on the most rigid part: its own storage and its own winter.

Simultaneous squeezes on the edge lines

Within the same week two more supply nodes locked up, and both teach the same lesson: dependence on a single line turns into fragility even in a planned operation. Gazprom halted gas flows to Armenia from 15 September to 25 September because of planned maintenance on the North Caucasus-Transcaucasus pipeline. In 2025 Russia supplied Armenia with 2.7 billion cubic metres of gas; that is 82% of the country's gas imports, at a price of 177.5 dollars per 1,000 cubic metres. The gap is to be covered from domestic reserves and from Iranian imports whose volume has not been disclosed. The sources diverge on the length of the cut: the Moscow Times takes the 15-25 September window while OC Media reports a ten-day interruption. The current agreement ends at the close of 2026 and the renewal terms have not been announced.

In Libya the interruption is on the oil rather than the gas side, but the mechanism is the same. On 15 September the Petroleum Facilities Guard closed a valve on the Hamada-Zawiya line, halting the Hamada and Tahara fields and the NC5 station; it threatened to shut seven fields, including Sharara and El Feel, if its demands were not met within a week. The NOC deemed the closure unlawful and warned that it might declare force majeure, but had not done so as of 15 September, and published no estimate of the output lost. In a country where oil makes up about 90% of the economy, the fact that a single group can close a valve shows that the supply risk is institutional rather than technical. Gas and oil are separate markets, but they are written into the same winter budget.

Where Türkiye stands in this equation

Türkiye is tied to both channels at once: an importer that relies on long-term pipeline gas contracts and on spot LNG alike. TTF is one of the main references for Türkiye's spot purchase cost; the price staying in the 77-79 euro band keeps the import bill and the energy pressure on the current account high. But unlike Bangladesh and Pakistan, Türkiye is not a buyer priced out at a spot level of 30 dollars; it stands on the expensive side of the competition rather than the losing side. The distinction matters: the loss here is measured not as a halted factory but as a bill. Because the source event records contain no direct volume or bill data for Türkiye, this link is drawn at the level of mechanism, without figures.

The second-round effects are on the trade and corridor side. With production partly halted at 78% of factories in Bangladesh's knitwear industry and an air freight cost of 50,000 dollars arising on a single shipment, delivery time risk moves ahead of price risk for buyers; suppliers with the advantage of proximity gain in relative terms under such conditions. Armenia seeing the cost of taking 82% of its imports through a single line, meanwhile, turns the debate about alternative corridors in the region into a political heading. The cutting of oil revenue in Libya additionally pushes back the collection timetable for contracting receivables tied to public payments.

What to watch

Three indicators will determine how the winter goes. First, the difference between EU storage fullness and the seasonal average: whether the gap, currently about 16 points, narrows or widens during the withdrawal season is a truer measure of the buffer than price. Second, whether the Asian spot LNG price stays around 30 dollars and whether the number of cargoes Pakistan confirms for October to December rises above September's two. Third, whether Qatari LNG flows through Hormuz recover: unless flows normalise, the loss Shell estimates at 36 million tonnes is not replaced.

The structural question is this: a storage gap can close within a few weeks if supply normalises. But in a market where price is culling buyers, normalisation can come not from supply rising but from demand disappearing for good. That orders have already been cancelled or cut at 55% of Bangladeshi factories is the first sign of it. On the day a fall in the gas price is measured, it is possible that the loss has already been written into another ledger. That is the real cost of the buffer being thin in two places at once: the price does not say which side has been repaired.