LME Aluminium Hits a 36-Year Low, but the Surplus Is Coming

London Metal Exchange (LME) aluminium stocks fell to 244,525 tonnes on 9 September 2026, the lowest level recorded since 1990. That figure is not a market curiosity. It is the measurable endpoint of a chain of events that began six months earlier with an Iranian missile strike on a single Gulf smelter.

The Middle East conflict that escalated in early 2026 did not just rattle sentiment. It physically removed metal from the market. Three producers, Emirates Global Aluminium (EGA), Aluminium Bahrain (Alba), and Qatalum, all cut output or declared force majeure within weeks of one another, taking a meaningful slice of a region that supplies roughly 8-9% of the world’s primary aluminium offline at once.

What follows in this LME aluminium supply shock is a producer-level account of how that geopolitical event translated into hard inventory figures, a multi-year price high, and a 2027 surplus forecast that complicates the picture. The question that matters for positioning is not whether the shortage is real. It is. The question is whether the current price reflects a genuine, durable shortage or a peak the forward curve is already starting to discount.

How three Gulf smelters went from full capacity to controlled shutdowns in six weeks

The first domino fell before the missiles did. On 3 March 2026, Qatalum, the Qatari joint venture between Hydro and QAMCO, began a controlled shutdown of its aluminium lines after QatarEnergy warned of natural gas shortages tied to the regional conflict. The vulnerability, in other words, pre-existed the direct military action.

QatarEnergy later confirmed it would maintain reduced gas supplies, and roughly ten days after the shutdown began, Qatalum resumed at approximately 60% of capacity. Against a nameplate figure of 640,000-648,000 tonnes (sources differ), that implies a curtailment of around 256,000 tonnes.

Then came the shipping disruptions. In early March, Alba declared force majeure on certain contracts after Strait of Hormuz logistics seized up, executing a controlled shutdown of reduction lines 1, 2, and 3. That removed roughly 19% of its annual capacity, about 308,369 tonnes, leaving the smelter running at approximately 81%.

The most severe blow landed on 28 March 2026, when Iranian missile and drone attacks caused physical damage at EGA’s Al Taweelah facility in the UAE. This was not a controlled curtailment. It was a total emergency shutdown and evacuation.

Everything went dark at once:

• Primary smelter
• Casthouse
• Power plant
• Alumina refinery
• Recycling plant

The production numbers confirm the severity. EGA’s cast-metal output in H1 2026 fell to 1,006,000 tonnes from 1,420,000 tonnes in H1 2025, a year-on-year drop of 414,000 tonnes driven almost entirely by the incident.

Industry-wide output losses linked to these Gulf disruptions are estimated at 3-3.5 million tonnes for 2026, roughly 5% of global supply. The distinction that matters for your read on this is between recoverable capacity and damaged infrastructure. Qatalum at 60% and Alba at 81% represent metal that can come back relatively quickly. EGA’s damage-driven rebuild is a multi-quarter problem, and it is the reason the LME drawdown became arithmetically inevitable rather than a market overreaction.

Nine months of unbroken inventory compression: what the LME data actually shows

With finished metal trapped behind Strait of Hormuz logistics and Gulf output curtailed, consumers had one place to turn: LME warehouses. The result was nine consecutive months of drawdown, and the monthly sequence tells the story better than any summary.

Stocks opened 2026 at roughly 511,750 tonnes. From there the decline was continuous, and crucially, it accelerated as the smelter shutdowns took hold in March and April.

The acceleration through March (down 8.47%) and April (down 11.6%) lines up precisely with the timing of the three shutdowns. That correlation hands investors a reliable leading indicator: when Gulf smelter output is removed at scale, LME depletion accelerates within weeks, not quarters.

Across H1 2026, cumulative LME holdings totalled 2.48 million tonnes, a 12.88% year-on-year decrease. A total of 178,356 tonnes exited exchange warehouses over the six months, an average of roughly 35,671 tonnes monthly. That is the pace of missing supply the market has been absorbing, and it cannot continue indefinitely without either price rationing or physical restarts.

For investors wanting granular month-level data on the inflection point, our detailed coverage of July’s price and inventory shift breaks down the specific warehouse movements and cash-to-three-month spread dynamics that marked the transition into the acceleration phase.

36-year low On 9 September 2026, LME aluminium stocks reached 244,525 tonnes, which LME Insight identified as the lowest level since 1990.

H2 2026: when Russian brand concentration became a secondary risk

There is a complication buried inside those thinning stocks. By the end of July, Russian brands accounted for 95% of available LME system aluminium. That matters because sanctions uncertainty makes some market participants reluctant to take delivery of Russian-branded metal in certain jurisdictions. The effective availability, in other words, is even tighter than the headline tonnage suggests, layering a sanctions-adjacent risk on top of an already acute physical shortage.

Why the price is at a multi-year high and what the forward curve is already pricing in

The price response was exactly what you would expect from a real physical shortage. Aluminium climbed 12.9% year-to-date, from $2,968/t at the end of 2025 to over $3,350/t by September 2026, tracking the stock drawdown almost inversely.

The 2026 aluminium price forecasts that preceded the conflict generally assumed a balanced-to-slight-surplus market, making the 12.9% year-to-date climb a significant divergence from consensus that has forced multiple investment banks to revise both their deficit estimates and their forward price assumptions.

The monthly averages show the pace holding through the second half. July averaged $3,155.98/t, August $3,249.63/t (up 2.97% month-on-month), and September an estimated $3,301.78/t (up 1.61%). On 9 September, cash settlement touched $3,352/t.

So far, so logical. The shortage is documented, the drawdown is measurable, and the price has followed. But this is where the story stops being straightforward.

In a June 2026 note, Goldman Sachs laid out a two-stage trajectory that should unsettle anyone reading the spot price as a durable signal:

1. 2026 deficit of 720,000 tonnes (revised up from 570,000), driving a Q3 2026 price target of $3,300/t. This is the shortage the market is currently pricing, and Goldman agrees it is real.
2. 2027 surplus of 590,000 tonnes (revised down from 1.3 million), as new smelting capacity from Indonesia and China arrives and Gulf production is progressively restored.

That second stage is the complicating variable. Because building a primary aluminium smelter typically takes five to six years, no meaningful greenfield capacity can arrive before 2027 except from projects already in progress. But those projects, plus EGA’s targeted full restoration by Q1 2027, are enough to flip the balance.

Goldman consequently lowered its 2027 average price forecast to $2,950/t, well below where the forward curve sits.

The forward curve warning Goldman Sachs cited the 2027 forward curve near $3,250/t as “likely overstating the duration of current tightness,” flagging clear downside risk.

Here is the read you should take from that gap. The spot price rationale is sound; the physical shortage is not in dispute. The open question is whether the forward curve, pricing tightness well into 2027, is charging you for a shortage that may already be past its worst point. Positioning on current spot strength without accounting for the surplus timeline carries specific downside risk by the middle of next year.

The recovery mechanics: EGA’s restart, Qatalum’s June force majeure, and what restoring Gulf capacity actually requires

Recovery in primary aluminium is not a switch. It is measured in potline restarts and refinery ramp schedules, which is why “recovery” here means quarters, not weeks. EGA’s restart is the clearest illustration, and the numbers are granular.

By mid-to-late August 2026, EGA had re-energised all three potlines and restarted between 227 and 315 of its total 1,262 reduction cells. That works out to roughly 18-25% of capacity, and the company reported it was tracking ahead of internal schedules.

The sequence tells you how staged this is:

EGA’s recovery timeline involves infrastructure interdependencies that go beyond the potline cell count, including power plant sequencing, alumina refinery pressure testing, and casthouse recertification steps that each carry their own failure modes and schedule risks.

1. All three potlines re-energised
2. 227-315 of 1,262 cells restarted (~18-25% of capacity) by August
3. Alumina refinery restarted on 10 July at 50% capacity, full technical capability targeted by end-2026
4. Recycling plant full production expected late Q4 2026
5. Smelter at full pre-incident output targeted for Q1 2027

Two details matter for your timeline. EGA explicitly stated the smelter ramp does not depend on the alumina refinery returning to full capacity, because alumina feedstock can be imported. And the company committed roughly AED 1.5 billion (about $400 million) across 2026 and 2027 to the rebuild, a state-backed signal that the UAE intends to hold its position in primary aluminium despite the conflict risk.

Then there is Qatalum, which complicates the recovery narrative. In mid-June 2026, after its partial resumption, Qatalum declared a second force majeure due to an agreement termination. A repeat force majeure after a partial restart signals ongoing operational fragility, not stabilisation, and its full restart timeline remains unknown.

Alba and Qatalum: slower recovery, less transparency

Alba’s situation differs from EGA’s in a useful way. Its shutdown was controlled rather than damage-driven, so its theoretical path back to full capacity is faster. The problem is that no public timeline has been confirmed. Neither Alba nor Qatalum has signalled any government-level strategic response, no strategic stockpiles, no region-wide insurance or force-majeure frameworks, in reaction to the year’s events. That leaves both exposed to the same logistical chokepoint risk if regional tensions re-escalate, a contrast with EGA’s committed rebuild.

Put the pieces together and the supply-side picture is clear. With EGA at 18-25% in August and Qatalum’s June force majeure unresolved, the Gulf base will not be functionally restored before Q1-Q2 2027. That means the inventory drawdown has further to run before any meaningful LME rebuild can begin.

What the Goldman surplus forecast means before investors act on the current price

Three threads now sit in tension. Gulf production is structurally impaired through at least Q1 2027. LME stocks are at a 36-year low of 244,525 tonnes, with Russian brand concentration adding a secondary availability risk. Yet Goldman’s revised 2027 surplus forecast of 590,000 tonnes and its $2,950/t price target suggest the forward curve is already discounting a longer shortage than the restart timelines support.

That frames the decision squarely. Is the current price a peak, bought at the top of a supply shock, or a floor established ahead of a prolonged deficit? The Goldman data argues for the former.

The most actionable number in all of this is the gap between Goldman’s $2,950/t 2027 forecast and the forward curve near $3,250/t. That is an implied downside of roughly 9% from forward pricing, and it quantifies the risk of holding a long position past the EGA restart milestone without a view on whether the 2027 surplus arrives on schedule.

Three variables will determine which scenario plays out:

The Iran conflict’s commodity market effects have not been uniform across metals: aluminium absorbed the sharpest supply-side hit because of Gulf smelter concentration, while other base metals with more geographically distributed production saw price responses driven more by demand-side uncertainty than physical supply removal.

• EGA’s restart progress, tracked against the Q1 2027 full-restoration target
• Qatalum’s force majeure resolution, still unresolved since mid-June
• Whether new Indonesian and Chinese smelting capacity arrives on Goldman’s projected timeline

The number that frames the trade Goldman’s 2027 forecast of $2,950/t against a forward curve near $3,250/t is the implied downside for anyone currently long on the shortage thesis.

The takeaway is not that the shortage is fake. It is real and documented. The point is that the trade of buying the shortage has largely been made, and the highest-risk moment for longs may be approaching as Gulf restarts compound with new global capacity. The forward curve has not yet fully reflected that transition, and that is the gap worth watching.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results. Financial projections are subject to market conditions and various risk factors, and forward-looking statements are speculative and subject to change based on market developments.