Investment funds cut their LME aluminium short position by 11.2% in the week to 28 August, and that covering drove four-fifths of the increase in net length. It happened without a corresponding price squeeze. Exchange stocks have since fallen to 245,975 tonnes, their lowest since 1990, with the remaining inventory heavily concentrated in Russian-origin metal and roughly two million tonnes of annualised Gulf production removed by the Middle East conflict. Deteriorating deliverability, rather than price, appears to be the binding constraint on holding short exposure.
TL;DR
- Fund net length rose 6,510.24 lots to +148,946.20, the largest position in the LME base-metals complex.
- Only 1,307.03 lots came from new buying. Shorts fell 5,203.21, accounting for 79.9% of the move.
- The remaining fund short base is 41,337.68 lots, down 11.2% in a single week. Longs outnumber shorts 4.6 to one.
- LME aluminium stocks stood at 245,975 tonnes on 3 September, their lowest since 1990.
- Remaining inventory is heavily concentrated in Russian-origin metal, less fungible for many Western users.
- Gulf primary production has fallen by roughly two million tonnes annualised following the February attacks on Iran.
- Three-month aluminium traded at $3,298/t at 09:10 GMT on 3 September, up 0.4%, having touched $3,306.50 — its highest since 13 August.
Market Overview
Aluminium presents an unusual combination: the largest speculative long in the base-metals complex, sitting alongside a price roughly 15% below its cycle high.
Three-month metal traded at $3,298/t at 09:10 GMT on 3 September, up 0.4% on the session, with an intraday high of $3,306.50 — the strongest level since 13 August. That is well below the $3,855 reached on 2 June, when the cash premium over three-month metal blew out beyond $100/t during an acute physical squeeze, and above the January base near $2,986.
The inventory position has not followed the price back toward normality. It has continued deteriorating throughout.
LME aluminium stocks stood at 245,975 tonnes on 3 September, a level Reuters reports as the lowest since 1990.
That gap — between a price that has partially recovered and a warehouse system that has not — is the central feature of the aluminium market entering September.
The Positioning: Bears Leaving, Not Bulls Arriving
The latest LME Commitments of Traders report, covering positions at the close on 28 August, shows investment funds net long 148,946.20 lots. That is the largest fund net position across the LME base metals.
The weekly increase of 6,510.24 lots looks straightforwardly bullish. The composition says something more specific.
Outright long positions rose 1,307.03 lots to 190,283.88. Short positions fell 5,203.21 lots to 41,337.68.
Short covering therefore accounted for 79.9% of the net increase.
That matters because the two sources of buying behave differently. Short covering requires no participant to hold a positive view — only to abandon a negative one. It is mechanically self-limiting: once the bearish positions are closed, the buying they generate stops and does not resume.
New long accumulation reflects capital committed to a directional thesis, and can continue as long as conviction holds.
Aluminium's advance is currently running mostly on the first mechanism.
Fund longs now outnumber shorts 4.6 to one. Funds hold 20.37% of total long open interest against 4.42% of the short side. The position is almost entirely speculative: risk-reducing exposure within the fund category amounts to four lots long and eight short against a total position exceeding 231,000 lots.
Some 307 separate fund entities hold aluminium, the largest participant count of any LME base metal.
The other categories moved the other way. Investment firms and credit institutions deepened their net short by 4,423.91 lots to −36,218.58. Commercial undertakings — the producers, fabricators and merchants — extended their net short by 1,480.08 lots to −136,180.56, the largest short position in the complex and consistent with producers hedging forward output.
What the Covering Does Not Appear to Be
The conventional reading of short covering is that a rising price forced it.
The sequencing does not support that explanation here. The covering occurred during the week to 28 August. Aluminium's recovery to a three-week high came afterwards, on 3 September. Through the reporting period itself, funds closed 5,203.21 lots of short exposure without a corresponding squeeze in the outright price, which remains materially below June's peak.
The more likely pressure sits in the warehouse data.
LME aluminium stocks fell to 245,975 tonnes on 3 September, their lowest since 1990 according to Reuters. At that level, exchange inventory covers less than a day of global aluminium consumption.
The composition of what remains is the more important constraint.
The remaining LME stock is heavily concentrated in Russian-origin metal, which is less fungible for many Western users because of sanctions, tariffs and customer restrictions. Metal produced before 13 April 2024 remains eligible for trading, but acceptance is uneven across the buyer base. Reuters described the remaining inventory in August as predominantly Russian, while analysis of the LME's July country-of-origin data put Russian-origin metal at almost 95% of warranted stock at that point. Indian aluminium, one of the more readily accepted alternatives, has largely disappeared from the exchange system over the past year.
A futures short does not oblige the holder to make physical delivery — positions can be closed or rolled. But scarce deliverable inventory raises the risk and potential cost of carrying short exposure into nearby tightness, particularly where much of the available metal carries an origin the holder may struggle to source or deliver on acceptable commercial terms.
The June squeeze featured a cash premium over three-month metal in excess of $100/t. Whether the curve has retained any of that structure is among the clearer near-term tests of whether deliverable scarcity is being actively priced rather than merely observed.
The Supply Side Behind the Inventory
The stock drawdown has a physical cause.
Primary aluminium production in the Gulf has fallen by around two million tonnes per year since the US-Israel attack on Iran on 28 February 2026. Set against LME registered stocks of 245,975 tonnes, exchange inventory covers roughly one-eighth of the annualised production loss.
Consumers unable to secure metal through normal supply channels have drawn on exchange stocks instead. That is the mechanism converting a production problem into a warehouse problem.
China cannot function as the usual release valve. It produces roughly half of global primary output but operates under a statutory production ceiling near 45 million tonnes a year, which removes the swing-supply response that has historically capped tightness episodes. Chinese domestic inventory has been comparatively well supplied, with SHFE stocks at 455,092 tonnes on 24 July, though declining since mid-June.
The result is the same geographic fragmentation visible in zinc and copper. Metal exists. It is not in the London system in a form much of the market can readily use, and the barriers are structural rather than temporary.
Indonesia represents the principal incremental supply response, supported by low-cost coal-based power and industrial policy. That capacity is real but arrives on a multi-year timeline, not a quarterly one.
Trade Policy as a Fragmenting Force
The 200% US tariff on Russian-origin aluminium articles and derivatives under Proclamation 10522 continues to shape where metal can profitably go, capturing any aluminium smelted or cast in Russia and exposing third-country buyers to secondary risk through content-tracing requirements. The wider Section 232 framework was modified again during 2026, but the 200% treatment of Russian-origin material was preserved through those changes.
The practical effect is to concentrate Russian metal in the venues that will accept it, which increasingly means LME warehouses, while non-Russian units are bid away by markets paying a premium for unrestricted origin.
That dynamic explains why the headline inventory number understates the tightness. The tonnage is recorded, but a substantial share of it cannot readily satisfy a large part of demand.
Bloodstone View
Aluminium's largest-in-complex fund long is not quite what the net figure suggests.
Read as a single number, it implies conviction. Read as components, it shows a market where bears have been leaving considerably faster than bulls have been arriving.
The covering does not appear to be explained by price alone. With LME stocks at a 36-year low and much of the remaining inventory Russian-origin, deteriorating deliverability appears to be increasing the risk of maintaining short exposure even without a renewed price squeeze.
If that reading is right, two things follow.
The covering can continue while the inventory position stays as it is, because the constraint is not primarily about the price level. But it also ends when the short base is exhausted, and that base is now 41,337.68 lots after falling 11.2% in a single week. At that rate, the mechanical support behind aluminium's advance has a measurable end point.
What replaces it is the open question. New long accumulation contributed only 1,307.03 lots last week. For the advance to continue once covering completes, that figure has to grow substantially.
The physical case for it to grow is intact: a 36-year inventory low, a capped Chinese production ceiling, two million tonnes of lost Gulf output, and a tariff regime fragmenting the deliverable pool by origin. Non-Russian primary producers remain the clearest beneficiaries of that fragmentation while the tariff structure holds. The financial evidence that funds are acting on the case, rather than simply retreating from the opposite one, is so far thin.
Aluminium is therefore a market where the fundamentals and the positioning point the same way for different reasons, and where the positioning has less room to run than the headline net figure implies.
Outlook
Base case: Aluminium holds a broad range while deliverable exchange stock remains scarce. The advance loses momentum as the remaining short base is exhausted, with the price supported by physical tightness rather than by continued speculative accumulation.
Bull case: Gulf supply disruption persists or deepens, exchange stocks fall further, and funds convert from covering shorts to adding outright length. A return to meaningful backwardation would signal deliverable scarcity being priced again rather than merely observed.
Bear case: Gulf production restarts, non-Russian metal returns to warrant, and exchange inventory rebuilds. Short covering completes with no new buying behind it, leaving a large and unsupported net long exposed to reduction.
Key Risks
- The short base is finite. At 41,337.68 lots and falling 11.2% weekly, the covering that has driven four-fifths of recent buying has a measurable end point. What replaces it is unresolved.
- New long accumulation is weak. Only 1,307.03 lots of fresh buying last week. If covering completes without conviction buying behind it, the net position becomes vulnerable to reduction.
- Gulf supply could restart. Two million tonnes of annualised production loss is the physical foundation of the inventory drawdown. Restoration would relieve it directly.
- Russian share of available stock could fall. Non-Russian units returning to warrant would signal the deliverability constraint is easing, which is the mechanical opposite of the current dynamic.
- Exchange stocks could rebuild. A sustained inventory recovery would undermine the scarcity argument even with the production loss unresolved.
- China's production ceiling could be relaxed. The 45-million-tonne cap removes the market's traditional pressure valve. Any policy signal on it would be the single most consequential input to the medium-term balance.
Intelligence Monitoring Points
- Fund short positions: currently 41,337.68 lots. The rate of decline determines how much covering support remains.
- Fund long accumulation: whether outright longs grow beyond last week's 1,307.03 lots, indicating conviction replacing covering.
- LME total and on-warrant stocks: the gap between them measures how much reported inventory is actually deliverable.
- LME monthly country-of-origin report: published in arrears, and the clearest measure of how much of the warranted pool is Russian.
- Cash-to-three-month spread: a return to backwardation would indicate deliverable scarcity being priced rather than observed.
- Gulf production restarts: the physical driver behind the inventory drawdown.
- SHFE inventory and Chinese export volumes: whether Chinese metal is moving towards the international market.
- China's 45-million-tonne cap: any policy statement easing the ceiling.
FAQ
Q: Why is short covering more significant than the net position? A: Because it is self-limiting. Covering requires only that bears leave, not that bulls arrive. Once the short base is closed the buying stops. Aluminium's short base fell 11.2% in a week to 41,337.68 lots, so the remaining support is measurable.
Q: If the price didn't squeeze, why did shorts cover? A: The evidence points to deliverability rather than price. Through the week to 28 August, funds closed shorts without a corresponding price move, while exchange stocks sat at a 36-year low with much of the remainder Russian-origin. That raises the risk of carrying short exposure into nearby tightness even absent a price squeeze. It does not establish the motivation of any individual participant.
Q: How low are LME stocks? A: 245,975 tonnes on 3 September, reported by Reuters as the lowest since 1990. At that level, exchange stock covers less than a day of global consumption.
Q: Why does the Russian share matter? A: Because reported tonnage overstates readily deliverable tonnage. Metal produced before April 2024 remains tradeable, but sanctions, tariffs and customer restrictions make it less fungible for many Western users, so a substantial part of the headline inventory cannot easily satisfy a large part of demand.
Q: What would most change the view? A: Non-Russian metal returning to warrant in volume, or a restart of Gulf production. Either would ease the deliverability constraint that appears to be shaping the covering.
Data: London Metal Exchange Commitments of Traders report, positions as of 28 August 2026, published 2 September 2026; positions reported in lots. Total long and short open interest may not balance owing to trades suspended at clearing and differing modelling methodologies for delta hedge equivalent option positions. LME inventory and price data as of 3 September 2026 where stated.
This document is published by Bloodstone Research for informational and institutional research purposes only. It does not constitute investment advice, an investment recommendation, an offer or solicitation to buy or sell any financial instrument, commodity or security, or a forecast of future performance. Market conditions and data may change without notice. Readers should conduct their own analysis and, where appropriate, seek independent professional advice before making investment decisions. For institutional enquiries contact research@bloodstonecapital.co.uk.