Zinc: The Squeeze Unwinds — Bloodstone Research

We discussed on 27 August that zinc's surplus had not disappeared but relocated to China, where London could not reach it. The metal has since started moving, and the London squeeze is unwinding: the cash-to-three-month backwardation has fallen from $231.75 on 27 August to roughly $87 on 21 September, while Chinese visible inventory has dropped about 21%. But the mine-level data point the other way, with ILZSG reporting first-half world mine production down 2.6% and concentrate treatment charges deeply negative. The near-term dislocation is resolving; the raw-material constraint behind it is worsening.

TL;DR

LME three-month zinc closed at $3,901.50 on 23 September, down 0.22%, against a peak of $3,949.50 on 26 August — the highest since June 2022.
The cash-to-three-month backwardation has roughly halved from its peak, running at $52 in the week to 7 August, $132 on 21 August, $231.75 on 27 August, $168.50 on 4 September and approximately $87 on 21 September.
SMM's seven-region Chinese inventory fell to 213,500 tonnes on 17 September from roughly 270,000 in late August, a decline of about 20.9%.
LME official stocks rose from 86,500 tonnes on 17 August to 113,100 by 4 September, but SMM's three-category total was 115,900 tonnes on 17 August and 126,700 on 4 September — a much smaller increase than the headline build suggested.
Registered warrants stood at about 82,900 tonnes on 4 September against 78,000 on 17 August, while the cancellation ratio rose to around 30%.
China exported 40,600 tonnes of refined zinc in August, up 340% month on month, while imports fell to 2,300 tonnes.
ILZSG preliminary data show world zinc mine production fell 2.6% in the first half, reversing 1.1% growth through May, with Chinese refined output up 5.9% while the rest of the world contracted 3.4%.
SMM's imported zinc-concentrate treatment-charge index stood at −$128.85 per dry metric tonne on 18 September, with domestic Zn50 charges at −Rmb2,100 per metal tonne.
Investment funds have added shorts for three consecutive reporting weeks, totalling 11,158.40 lots to 18 September, while holding the largest fund footprint in the LME complex on both sides.

The Squeeze Is Unwinding

The most useful measure in zinc is not the flat price but the spread, and it has moved decisively.

The cash-to-three-month backwardation ran at roughly $52 a tonne in the week to 7 August, $106 to 14 August, $132 on 21 August, $199 on 26 August and $231.75 on 27 August — a near-vertical progression from a $25 cash discount three months earlier. It then eased to $201 by the end of August, collapsed to $113.50 midweek as arrivals landed, and tightened back to $168.50 on 4 September. On 21 September, with cash metal around $4,018 and three-month near $3,931, the backwardation was approximately $87.

That is a fall of roughly 62% from the peak, and it returns the spread close to where it stood on 17 August, before the squeeze reached its extreme. The flat price has softened alongside it, from $3,949.50 on 26 August to $3,931 on 21 September and $3,901.50 on the 23rd.

The honest qualification is that the detailed warehouse data explaining the move are not available for the period in which it happened. The last full stock and cancellation figures used here are for 4 September, when the backwardation was still around $168.50. What produced the subsequent halving — further arrivals, cancelled metal returning to warrant, long holders reducing positions or some combination — cannot be established from the data to hand. What the spread does establish is that the extreme premium for immediately available London metal has diminished substantially.

Why the Early Build Did Not Help

The reason the initial relief took so long to arrive is worth recording, because it explains the shape of the episode.

LME official stocks rose from a low of 86,500 tonnes on 17 August to 113,100 by 4 September, an apparent build of 26,600 tonnes. Yet the backwardation widened across much of that window, from $82.88 to around $168.62 on SMM's measure. The explanation is that the headline inventory number overstated the improvement in immediately available metal. SMM's tracking of registered warrants, cancelled warrants and off-warrant metal shows the three-category total at 115,900 tonnes on 17 August, rising to 126,900 on the 18th as early Chinese export deliveries emerged, and standing at 126,700 on 4 September. The increase in the wider pool was therefore considerably smaller than the official-stock build suggested.

The deliverable pool tells the same story. Registered warrants were 78,000 tonnes on 17 August, increased to 87,900 on the 18th, then drifted back to around 82,900 by 4 September. Meanwhile cancellations went from 4,550 tonnes on 21 August to 22,350 on the 28th — including a single-day jump of 19,125 tonnes, mostly in Singapore, taking the cancellation ratio from 3.6% to 22.82% — and reached 30,250 tonnes by 4 September, with the ratio near 30%.

So a headline build of 26,600 tonnes produced only around 4,900 tonnes of additional registered warrants, while a substantial share of exchange stock had simultaneously been booked for removal. LME Insight's assessment at the time was that the arrivals were easing rather than resolving the squeeze. The subsequent spread evidence suggests resolution has progressed considerably further — but it took several more weeks.

The Metal Has Moved

The relocation identified in August has since become visible on both sides of the trade. SMM's seven-region Chinese inventory fell to 213,500 tonnes on 17 September from roughly 270,000 in late August, a reduction of about 56,500 tonnes or 20.9% in roughly three weeks.

The export arbitrage provides an important part of the explanation. A falling SHFE-to-LME ratio opened an export window into Southeast Asia during August, with the LME benchmark outperforming SHFE sufficiently to make movement of refined Chinese zinc into the international market commercially attractive, and LME Insight attributed the early-September arrivals partly to increased Chinese refined exports into the LME system. The trade data strengthened that interpretation: China exported 40,600 tonnes of refined zinc in August according to SMM's analysis of customs data, up 340% month on month, while imports fell to just 2,300 tonnes, taking China into a small cumulative net-export position across January to August.

That is the mechanism the August analysis was waiting for. Zinc had not disappeared from the global system; a significant portion of the available refined metal was sitting in a geography from which London could not immediately draw it, and a sufficiently large price differential created the incentive to move some of it. The caveat is what the Chinese draw represents. Inventory falling because metal is leaving the country is different from inventory falling because consumption is accelerating, and SMM has continued to describe downstream order growth as limited, although lower prices and pre-holiday replenishment have improved spot activity in September. This is principally reallocation rather than absorption.

The Mine Data Point the Other Way

While the London dislocation eases, the constraint underneath it has deepened.

ILZSG preliminary figures show world zinc mine production fell 2.6% in the first half of 2026, reversing the 1.1% growth reported through May. Losses have been concentrated at operations including Antamina in Peru, Garpenberg in Sweden and Red Dog in the United States, while the closure of Australia's Lady Loretta at the end of 2025 removed further tonnage. Higher volumes from Kipushi in the Democratic Republic of Congo and the restart of zinc recovery at Aljustrel in Portugal provided partial offsets. Producer disclosures reinforce the aggregate picture, with Glencore's first-half zinc output down 21% year on year to 365,600 tonnes and Boliden's concentrate output down 16.8% quarter on quarter to 74,200 tonnes.

Treatment charges register the consequence. SMM's imported zinc-concentrate treatment-charge index stood at −$128.85 per dry metric tonne on 18 September, while its domestic Zn50 treatment charge remained at −Rmb2,100 per metal tonne. Negative charges of that magnitude show how far bargaining power has shifted towards miners as smelters compete for available feed: instead of receiving the conventional positive processing fee, smelters are effectively accepting deeply negative terms to secure concentrate.

The lag is what matters for the fourth quarter. If smelters cannot replace concentrate at economic terms, sustained feed tightness eventually constrains refined production, so today's concentrate problem can emerge in refined supply later — precisely when Chinese export flows would need to remain available if they are to keep relieving London. The refined balance is also less comfortable than earlier forecasts implied: ILZSG had initially projected a sizeable global surplus for 2026, the first half recorded an actual surplus of around 120,000 tonnes, and subsequent estimates have moved materially towards balance. The direction of revision matters more than the residual. The cushion expected at the start of the year has substantially diminished.

These two conditions can coexist. In the near term, existing refined stocks and Chinese exports can relieve London even while the pipeline producing tomorrow's refined zinc becomes more constrained. The question is how long the first mechanism can outrun the second — and if concentrate shortages eventually constrain refined output, the source of metal currently relieving London becomes harder to sustain.

Funds Have Been Short the Whole Way

Investment-fund positioning has moved with the London spread rather than against it, which is worth stating plainly because the opposite reading was available a fortnight ago.

Funds added 3,067.46 shorts to 4 September, 3,305.69 to 11 September and 4,785.25 to 18 September — an accelerating sequence totalling 11,158.40 lots. The short book reached 70,170.25 lots while net length fell to +53,734.10. The long side tells a different story: funds added 4,667.72 longs in the first of those reporting weeks before removing 216.00 and then 603.91, so this has been predominantly bears arriving rather than existing bulls abandoning the market. Scale gives it weight, with funds holding 32.67% of zinc long open interest and 18.50% of short, the largest Investment Funds presence on both sides anywhere in the six-metal complex, across 302 position holders.

That does not establish foresight. Positioning records what funds did, not what they concluded, and the same sequence is consistent with a view on the export arbitrage, hedging, relative-value positioning or simple mean reversion after an extreme spread. The narrower point is the useful one: funds have not been proved correct about the zinc supply chain, where mine production has deteriorated and treatment charges remain deeply negative. What has normalised is the immediate scarcity of London metal.

Outlook

Base case: The backwardation continues normalising as Chinese export flows persist, while the flat price remains elevated by mine-supply losses and deeply negative treatment charges. Chinese visible inventories continue drawing without evidence of a comparable acceleration in domestic downstream demand.

Upside risk for price: Concentrate scarcity feeds through into weaker refined production during the fourth quarter, reducing the availability of Chinese metal for export. Mine disruption persists, Chinese inventories continue falling and the London backwardation begins widening again.

Downside risk: The export arbitrage remains open for a full quarter, arriving metal becomes available rather than being rapidly cancelled, and the London deliverable pool rebuilds materially. Continued spread normalisation combines with the growing fund short book to remove some of the scarcity premium from the flat price.

What would change the view: A sustained recovery in concentrate treatment charges would weaken the upstream-tightness thesis. Conversely, a renewed widening of the London backwardation alongside falling registered inventories would indicate that the relief from Chinese exports was temporary.

Key Risks

Warehouse detail stops on 4 September. The spread halved after that date, and the registered-warrant and cancellation data explaining the move are not available in the evidence used here.
The $87 spread is derived. It comes from cash zinc near $4,018 against three-month metal near $3,931 on 21 September and may not represent a same-timestamp quotation.
Total stock is not deliverable stock. The 26,600-tonne official-stock build between 17 August and 4 September produced only around 4,900 tonnes of additional registered warrants.
The Chinese inventory draw is not simply demand. Exports account for part of the reduction, while downstream order growth remains limited.
Concentrate and refined zinc are different markets. Deeply negative treatment charges establish feed scarcity but do not automatically imply an immediate refined-metal shortage, and they affect refined supply with a lag.
Balance forecasts remain unstable. The expected 2026 refined-market cushion has narrowed materially, but preliminary estimates remain subject to revision.
Positioning is not a forecast. Three consecutive weeks of short building record fund positions, not the reasoning behind them.

Intelligence Monitoring Points

Cash-to-three-month backwardation, last approximately $87 against a $231.75 peak, as the clearest measure of whether London relief is holding.
LME registered and cancelled warrants, last detailed at roughly 82,900 and 30,250 tonnes on 4 September.
SMM seven-region Chinese inventory, at 213,500 tonnes, and whether the draw continues towards 200,000 tonnes.
Chinese refined-zinc exports, after the August surge to 40,600 tonnes.
Imported concentrate treatment charges, at −$128.85/dmt, and domestic Zn50 at −Rmb2,100 per metal tonne.
ILZSG mine-production revisions, after the first-half decline of 2.6%.
Fund positioning, with the short book at 70,170.25 lots after three consecutive weekly increases.

FAQ

Is the zinc squeeze over? It is unwinding rather than over. The cash-to-three-month backwardation has fallen from $231.75 on 27 August to roughly $87 on 21 September, but cash metal still commands a substantial premium over three-month zinc.

Why did the early LME stock build not solve the squeeze? Because the headline increase overstated the improvement in immediately deliverable metal. Registered warrants rose much less than official stocks, while cancelled warrants rose sharply.

Why can the squeeze ease while mine supply worsens? Because they operate on different horizons. Existing refined stocks and Chinese exports can relieve an immediate London shortage even while insufficient concentrate threatens future refined production.

Were the funds right? The London spread has so far moved in the direction implied by the short build, but the underlying supply problem has not disappeared. Mine output and concentrate treatment charges have moved the other way.

Data and source note: The approximately $87 backwardation for 21 September is derived from reported cash metal near $4,018 against three-month near $3,931 and may not represent a same-timestamp quotation. LME warehouse and cancellation figures cover 17 August to 4 September 2026; more recent detailed data were not available in the evidence used here, and the spread halved after that date. The attribution of LME arrivals to Chinese exports is LME Insight's and SMM's reading of the flows, not a disclosed shipment record. Investment-fund positioning is from LME MiFID II Weekly COTR reports, Investment Funds, non-risk-reducing, for positions as at 4, 11 and 18 September 2026; zinc lots are 25 tonnes.

Sources

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.

Zinc analysis — 2026-09-24