ISSUE #014 · WEDNESDAY, SEPTEMBER 30, 2026
Copper gave back some of last week’s gains, with COMEX down roughly 3.5% from September 22 and LME three-month copper down about 2.1%. The move was largely macro-driven. A stronger dollar, higher bond yields and renewed inflation concerns weighed on the broader metals complex, while weak Chinese industrial-profit data added to the pressure.
Meanwhile, inventories kept following a familiar pattern as COMEX stocks rose another 5,083 tonnes to 702,350 tonnes, while LME stocks fell 3,775 tonnes to 250,475 tonnes. Chinese exchange inventories also remained low, with SHFE stocks falling another 15.9% week-on-week to 47,147 tonnes.
In other words, prices cooled, but the underlying inventory picture did not change much. And that leads us into this week’s feature, because the increasingly bullish copper case rests not just on how much copper exists above ground — but how much of it is actually available to the market.
The Road to $22,000 Copper
01 THE SETUP
This week, Deutsche Bank’s head of metals research, Daniel Ghali, forecast copper at $10/lb, or $22,050/t, by Q2 2027 — about 50% above current levels.
The call is built primarily on stockpiling rather than unusually strong end-demand: Ghali argues that strategic and tariff-driven accumulation is rapidly shrinking the pool of copper actually available to ordinary consumers. Specifically, Ghali points to a rapid decline in unencumbered Above Ground Stock (AGS), which he estimates has fallen to around 29% of a relatively stable ~60 days worth of supply in recent years.
Meanwhile, the other 71% is sitting in either China’s strategic reserve (SRB) (~2.05Mt) or economically encumbered in US Warehouses (~1.3Mt).
Further, Ghali estimates that if stockpiling continues at its current pace, freely available copper runs out by the end of 2028, producing what he’s calling the “most acute copper scarcity on record.”
02 WHY IT MATTERS
If Deutsche is right, today’s copper prices are merely the opening act.
03 THE DECODE
While we love a good copper bull case as much as the next guy, there are several assumptions here that need to go right.
That the recent stockpiling episode has actually depleted free inventory: the ~60 days of global AGS really does contain only ~18 days of economically accessible copper while the rest is meaningfully encumbered.
The major encumbered pools stay encumbered as prices rise: China does not materially release any of its ~2 Mt strategic reserve, US stocks remain economically landlocked, and industry/WIP inventories cannot be reduced enough to provide a buffer.
That US and Chinese accumulation continues at its current pace, or, even if China reaches its reserve target and US warehouse accumulation slows, underlying refined production must eventually fail to keep pace with ordinary end-use demand.
That the underlying refined production balance assumptions are correct.
Apparent concentrate tightness, (derived via negative TCs), actually becomes a refined-copper constraint: negative TCs must ultimately reflect insufficient contained copper feed, not principally excess smelting capacity.
The market cannot mobilise other buffers before ~$22k: that scrap supply, working-inventory reductions, merchant destocking, aluminium substitution and other thrift must respond too slowly or too weakly at $14k-18k to clear the market.
Demand remains sufficiently price-inelastic and end users keep consuming enough copper at progressively higher prices that a much higher price is required to destroy demand. If Chinese and other consumers are already backing away materially at lower prices, the necessary clearing price could be much lower.
For the sake of simplicity, we’ll assume that point #1 and #2 are largely correct — that stockpiling has depleted free inventory and that China’s reserve and US warehouse metal are, indeed, encumbered. Although, we will add one quick caveat — “encumbered” is an economic classification rather than an immutable property: China has drawn on its reserves before, and US metal can presumably move before we hit $22k copper. Particularly in the event of a tariff rejection.
Where we think it’s worth taking a second look is in the remaining assumptions.
Stockpiling continues at its current pace
If freely available copper is to run out by 2028, stockpiling must continue at something close to its current pace.
The two main stockpiles here are the U.S. and China.
China (flat) — The SRB line (Figure 6) is flat near 2.05 Mt. Meanwhile, market reporting has long referred to a ~2 Mt target, although Beijing has never confirmed such a target exists. In any case, Deutsche’s own series shows no fresh accumulation.
U.S. (accumulating) — The US series climbs from near zero at end-2024 to about 1.3 Mt at end-2026, roughly 600 kt a year, which matches the 2028 date. At ICSG’s forecast 2026 usage of 28.66 Mt, 17-20 days of demand equals about 1.3-1.6 Mt of free stock. A US build of ~600 kt a year uses that up around 2028.
However, sustained U.S. accumulation at current rates is not guaranteed. As we noted a little over a month ago, the tariff trade may be testing its limits. And, just this week, SMM is reporting the tariff trade is fading (while noting it isn’t necessarily over yet).
Elsewhere, physical constraints may slow the flow of metal into the U.S. Just last week, the Port of New Orleans, a major COMEX warehouse hub, was reported to be largely full. And while warehouse operators have moved to add storage capacity, we’re yet to see whether this will be enough to keep pace with a sustained ~600kt/year build.
Refined balance deficit
Deutsche’s estimates show refined deficits of about 90 kt in 2027 and 300 kt in 2028.
However, this is far from a consensus opinion. For instance, ICSG’s April forecasts effectively disagree with the sign, showing +96 kt for 2026 and +377 kt for 2027. In either case, neither projects a balance that drains free stock without continued stockbuilding.
Indeed, if stockpiling stops, the picture could look quite different. If Deutsche’s forecast deficits draw down inventory, free stock would end 2028 at roughly 0.9-1.2 Mt, or 12-15 days of demand. (Thin by historical standards, but far from zero.) If ICSG is closer to the underlying production-consumption balance, the depletion of free stock becomes much harder to explain without continued off-market accumulation.
Concentrate tightness
Another key assumption in Ghali’s thesis is intensifying copper concentrate tightness. Notably, he points to plummeting Chinese concentrate treatment charges (TCs) as evidence.
While there’s no doubt that negative TCs suggest intense competition for a limited pool of concentrate, we’re not so confident in the leap from negative TCs to a threat to refined availability. A few points worth noting:
TCs measure the wrong gap: TCs measure feed scarcity relative to smelter demand, rather than contained copper relative to final refined demand.
Capacity has outgrown the feed: The broader processing system has expanded rapidly (32.6Mt in 2023, 33.9Mt in 2024, 35.7Mt in 2025, and H1 2026 capacity at 21.3Mt versus an H1 2025 capacity estimate of 20.5Mt), while S&P estimates global smelting capacity rose about 8 Mt over the past decade while concentrate output rose about 3 Mt.
Refined output is still growing: Per ICSG, Jan-Jul, concentrate production fell 2% while SX-EW output rose 3.6%. Primary refined production still grew 1%, and secondary refined production grew 4.3%, although a squeeze may be starting to register at the margin in China.
The number that would settle Deutsche’s claim is the balance between contained copper available in concentrate and the copper needed for primary refining. ICSG’s forecast adjusts for exactly that gap and still produces surpluses.
Demand destruction
For $22,000 copper to become a reality, end-demand must presumably hold up. However, price sensitivity is already visible, with Chinese fabricators adjusting at current prices, well below $22,000.
Just this month:
Sep 4-10, wire and cable: Operating rates fell to 63.71%, down 2.93pp week-on-week and 3.91pp year-on-year, as high prices held back orders.
Sep 4-10, enamelled wire: Operating rates fell 1.87pp to 67.53% and new orders dropped 5.90pp.
Sep 11-17, after prices dipped: Enamelled-wire operating rates rose 5.2pp and new orders jumped 16.67pp. Copper rod operating rates rose 4.32pp to 65.82%.
Sep 18-24, as prices climbed back: New enamelled-wire orders fell 13.11pp. (Although pre-holiday restocking had largely been completed the week before, which may account for part of the drop).
While this doesn’t reveal global demand-destruction at current prices, it does show at least some Chinese fabricators cutting runs, carrying less raw material and releasing orders when prices retreat, which may already be adding some resistance.
tl;dr
Deutsche may well be right that copper is unusually tight. But “tight” and “another +50% leg up while free inventory heads to zero” are very different claims.
The latter requires continued strategic accumulation, a worsening underlying refined balance, concentrate scarcity that actually constrains cathode output, and surprisingly weak supply- and demand-side responses to much higher prices.
04 WHAT TO WATCH
The tariff decision and the COMEX-LME spread. A narrowing spread with slowing inflows means the stockpiling increment is fading. Renewed widening puts Deutsche’s path back on track.
ICSG’s autumn forecast update. April’s numbers had a 377 kt surplus for 2027. Look for a cut toward Deutsche’s -90 kt.
ICSG monthly data. July showed a 51 kt deficit, and refined supply slipped 1% year-on-year against a 3% rise in usage. Consecutive monthly deficits would strengthen Deutsche’s case on the underlying balance.
China in Q4. The ~80 kt of smelter maintenance, the effect of China’s scrap crackdown, and SMM operating rates after the National Day holiday.
H1 2027: Watch whether free stock actually draws down or US inflows plateau. Also watch the recovery pace at Grasberg and Kamoa, and any sign of Chinese reserve activity, whether buying or releasing.
2027-2028: Watch whether increased aluminium substitution shows up before prices approach $22,000, and whether ICSG’s measure of contained copper in concentrate starts to fall short of primary refining requirements.
/ China’s Refined Copper Output Growth Is Slowing Sharply
China’s refined copper output is expected to grow just 3-3.4% in 2026, versus 10.4% last year, as concentrate and scrap availability tighten and smelter economics deteriorate.
/ Peru Wants Another Million Tonnes of Copper
Peru is targeting 1Mt/year of additional copper production within 5-6 years, while cutting around 27 permitting requirements to accelerate projects. Both Chinese and US-backed miners are showing interest.
/ Big Miners Are Still Shopping for Scale
Mining majors continue to favour mergers, partnerships and joint ventures as new copper projects increasingly carry $10-20 billion price tags. The attraction is less about financial engineering than spreading the cost and political risk.
Operations at Escondida have restarted following last week’s accident, but supervisors are now voting on a contract offer their union has recommended rejecting. Rejection moves negotiations toward mediation and, potentially, strike action at a mine capable of producing around 1.28 Mt this year.
Eldorado Gold, one of the relatively rare large copper projects actually progressing toward new production, has connected its Skouries copper-gold mine in Greece to the national power grid, clearing another commissioning milestone.
That's the wire for today. Until next time, follow the tonnes.
Know someone who keeps asking you about copper?
They can subscribe free at copperwire.news.