China's primary aluminium output is frozen against a hard policy capacity ceiling, the alumina cost shock has fully reversed, and smelter margins sit near ¥7,800/t. The bull case is structural supply; the brake is a 1.36 Mt domestic inventory glut.
Aluminium enters mid-2026 as one of the structurally tightest base metals, and the cleanest supply-side story in the complex. China's primary (electrolytic) aluminium capacity is pressed hard against the policy ceiling imposed under supply-side reform — operating capacity now sits essentially at the policy cap, with operating rates at historic highs and virtually no headroom left in the system. Domestic supply growth has, for practical purposes, peaked. At the same time the violent 2024–25 alumina squeeze has fully reversed: the SMM alumina index has round-tripped from a December-2024 blow-off near ¥5,700/t back to ¥2,691/t, collapsing the smelter cost base and lifting immediate profit to roughly ¥7,800/t from a ¥1,600/t loss eighteen months ago.
As of 9 June 2026, prices have ground higher across every benchmark. SMM's A00 spot is ¥23,900/t (+6.4% YTD), the SHFE main contract ¥24,110/t, and the LME 3-month US$3,608/t (8 June close, +20% YTD), having spiked sharply in Q1 2026 on Middle-East supply disruptions. SMM's base case puts the LME in a US$3,000–3,400/t trading range through the second half of 2026, implying a partial unwind of the Q1 disruption premium (see Chapter 8). There is more than one soft spot — and that is why this is not a runaway bull market. The first is inventory: Chinese ingot social inventory has roughly doubled this year to 1.36 Mt as seasonal builds met persistently weak property demand and a fading solar pull; beyond the structural bright spots of grid, EVs and energy storage — storage the fastest-growing but still small in absolute terms — no downstream segment is strong enough to offset property, capping the upside even as overseas LME stocks fell 36%. The second is macro: market expectations of Federal Reserve rate hikes weigh on metal valuations across the complex.
This report uses SMM's proprietary price, cost, inventory, production and downstream datasets, cross-verified against the exchanges and the observable market consensus, to make five arguments. (1) The market is supply-capped: the capacity policy ceiling means China's primary output has next to no room to grow, so incremental supply must come from offshore (Indonesia, India) or from recycling — a structural support that does not depend on the demand cycle. (2) It is margin-rich: the alumina collapse, not a demand surge, is what fattened smelter margins and drove this profit cycle — the volatility sits in the raw-material leg, and aluminium does not price off its cost. (3) It is demand-electrified: grid investment, EV lightweighting and rapidly scaling energy storage are the structural growth pillars offsetting weak construction — while the solar (PV) pull has come off its peak, with 2026 installation expectations cut to around 210 GW after the mechanism-price reform (Document No. 136). (4) It is inventory-watched: the 1.36 Mt domestic glut versus overseas tightness is the defining tension, and the inventory turning point is the swing variable for H2. (5) It is increasingly carbon-priced: EU CBAM began charging on 1 January 2026, and a green-aluminium premium is emerging that will re-rate low-carbon metal.
For the participants at Singapore International Ferrous Week's Green Energy Metal Forum, the implications are direct. Aluminium is the metal where the energy transition, supply-side policy and carbon pricing intersect most sharply, and where the China-versus-rest-of-world divergence is widest. Singapore sits at the centre of the price discovery (LME, SHFE-offshore), the trade finance, and the green-finance and certification flows that will define which tonnes qualify for a low-carbon premium. The chapters that follow lay out the data, the scenarios and the actionable conclusions.
The evidence base is SMM's proprietary light-metals data complex: daily A00, alumina, bauxite, anode, ADC12 and premium assessments; the SMM aluminium cost-and-profit model; the regional ingot and billet social-inventory survey; monthly output, operating-capacity and operating-rate series; weekly downstream operating-rate panels (processing, extrusion, foil, billet, secondary); and customs trade data — cross-verified with LME/SHFE, the IAI, and the publicly observable market consensus. Every headline figure ties to an SMM indicator ID or a named external source.
The bottom line for the busy reader: aluminium is the rare market where the bear case (a 1.36 Mt domestic glut) and the bull case (a hard supply ceiling, overseas tightness and an electrification demand engine) are both strong, true and simultaneous — which is why the flat price has gone sideways-to-higher rather than breaking in either direction. The cycle's energy has gone into the margin, courtesy of the alumina collapse, and into the structure — the SHFE–LME divergence, the green premium, the regional premium dispersion — rather than into the headline price. A reader who tracks only the LME number will miss most of what is happening; the action is in the cost stack, the inventory line and the carbon axis. Those are the three dials this report asks the forum to watch.
Aluminium moves through five stages — Bauxite → Alumina → Smelting → Semis & Processing → End Use — and is split throughout by route: energy-intensive primary metal (≈14.8 t CO₂/t globally) versus low-carbon secondary (recycled) metal. Carbon intensity, not just cost, is becoming the second axis of value.
Conversion ratios and cost shares are SMM indicative estimates for 2025–2026; carbon intensities are IAI/SMM composite figures.
Prices have ground steadily higher on a capped supply base and collapsing costs — but a record domestic inventory glut keeps the rally honest. This is a margin story as much as a price story.
Aluminium's 2024–2026 price action is a study in how a market can be both structurally tight and tactically capped at the same time. SMM's A00 spot has ground up to ¥23,900/t (+6.4% YTD), the SHFE main contract to ¥24,110/t, and the LME 3-month to US$3,608/t — the latter up roughly 20% year-to-date and having spiked hard in Q1 2026 when Middle-East force-majeure events (Bahrain, Qatar, EGA) cut overseas supply. The A00 premium/discount sits near −¥70/t, a sign of ample deliverable spot in China.
Step back and the 2024–2026 arc is a clean three-act story. Act one (2024) was a cost-push: tight bauxite and alumina dragged the whole stack higher and squeezed smelter margins toward break-even, even into loss. Act two (late-2024 into 2025) was the reversal: alumina collapsed as refining capacity caught up, the cost base deflated, and margins exploded while the metal price held firm. Act three (2026, the present) is the consolidation: a capped-supply, fat-margin market trading sideways-to-higher, with the flat price hemmed between a capped-supply floor and a record-inventory ceiling, punctuated by the Q1 disruption spike. What makes this cycle unusual — and worth a forum's attention — is that the metal price and the smelter margin moved independently: the margin is a cost story, the price is a supply-and-inventory story, and conflating the two is the most common analytical error in aluminium today. This report keeps them deliberately distinct.
The deeper story, however, is in margins, not the flat price. The single most important development of this cycle is not a demand surge but the alumina round-trip: the SMM alumina index ran from roughly ¥3,350/t in early 2024 to a December-2024 blow-off near ¥5,700/t — when lagging alumina-refining capacity, compounded by episodically tight Guinea bauxite, briefly inverted the cost stack — and has since crashed back to ¥2,691/t. Because alumina is 30–40% of the smelter cost base, that round-trip swung SMM's modelled all-in cost from a peak near ¥21,400/t to about ¥16,000/t, and smelter immediate profit from a ¥1,600/t loss to roughly ¥7,800/t. The metal price rose; the margin exploded.
Supply-capped: China's capacity policy red line keeps domestic primary growth constrained.
Margin-rich: the alumina collapse fattened smelter margins — the volatility sits in the
raw-material leg; aluminium itself does not price off cost.
Demand-electrified: grid, EV and energy storage offset weak construction.
Inventory-watched: a 1.36 Mt domestic glut caps the rally.
Carbon-priced: CBAM charging since January 2026.
The brake on the rally is inventory. Chinese ingot social inventory has roughly doubled this year to 1.36 Mt on seasonal builds and a still-soft property market, and SHFE on-warrant stock has rebuilt sharply — both signs of ample domestic spot. The mirror image sits offshore, where LME stocks have fallen 36%, leaving the overseas market structurally tight. This China-glut / overseas-tight divergence (Chapter 4) is the defining tension of the 2026 aluminium market and the cleanest spread for the trade to express.
The sharp Q1-2026 spike was not a demand event; it was a supply-shock layered onto an already-tight overseas balance. A cluster of Middle-East disruptions — force-majeure and curtailment at Bahrain (Alba), Qatar (Qatalum) and the broader EGA complex — removed meaningful tonnes from the seaborne market just as European smelter capacity, shuttered through the 2022–23 energy crisis, had failed to return. With no China-scale swing capacity outside the country able to backfill, the LME 3-month ran sharply higher before settling near US$3,608/t. The episode is a template for how aluminium now trades: because the world's marginal producer (China) is frozen at a ceiling and cannot respond to price, any ex-China outage transmits straight into the LME with little supply elasticity to absorb it. That asymmetry — abundant capacity that cannot grow versus scarce capacity that will not grow — is the structural reason aluminium's disruption-driven tail risk is skewed to the upside, relative to the base path, even in a year of record Chinese inventory.
A quieter structural support is the copper-to-aluminium price ratio, which has held near multi-year highs as copper pushed toward record levels. When copper trades at several times the aluminium price per tonne — and at a far higher multiple per unit of conductivity-adjusted volume — fabricators and grid operators substitute aluminium for copper in cabling, busbar, transformer windings and PV/EV applications wherever the engineering tolerates it. This is a slow, sticky, one-way demand transfer: once a utility or an automaker qualifies an aluminium conductor or busbar design, it rarely reverts, because re-qualification carries its own engineering cost and the price gap has persisted long enough to look structural rather than cyclical. SMM's read is that the substitution bid is a genuine, if hard-to-quantify, floor under aluminium demand growth, and it compounds the electrification thesis rather than competing with it. The rest of this report unpacks each link in the chain, beginning upstream.
The defining swing of the cycle happened upstream. Alumina blew off and crashed; bauxite from Guinea remains the swing input; anodes and power round out a cost base that has deflated hard.
To understand aluminium margins in 2026 you must understand what happened to alumina. The intermediate that sits between bauxite and metal — roughly two tonnes of alumina per tonne of aluminium — became the most volatile node in the chain. The SMM alumina index ran to a December-2024 peak near ¥5,700/t as refining-capacity additions lagged the smelter pull and episodically tight Guinea bauxite compounded the squeeze; it has since crashed to ¥2,691/t as new refining capacity in Guangxi came online and imports loosened the balance. The SHFE alumina futures main contract trades around ¥2,838/t, and the national weekly operating rate has eased to 72.7% — loose.
SMM's read (1 June 2026) is that alumina fundamentals are "relatively loose," with spot expected to consolidate sideways into June as new capacity ramps and seasonal maintenance ends. For the smelter, this is the single best piece of news in the cycle: a stable, low alumina price locks in the fat margin. The risk, symmetrically, is that any renewed bauxite or alumina squeeze re-inflates the cost stack — which is why the alumina operating rate and Guinea bauxite flows are the two upstream variables worth watching most closely.
Guinea CIF bauxite (Al₂O₃ 47–50%) is US$69.5/t — up roughly 15% from the ~US$60/t lows averaged earlier in the year, but still far below the ~US$105/t peak that accompanied the 2024 alumina crunch. Guinea remains the swing supplier to China — the Winning Consortium's Boké exports exceeded 70 Mt (wet) in 2025 and are guided above 90 Mt in 2026 — and China's monthly alumina imports jumped to 610 kt in April 2026 as cheap overseas tonnes flowed in. Indonesia, for its part, revised its bauxite benchmark down roughly 15% under a new pricing formula (ESDM 144/2026), a reminder that resource-nationalist pricing cuts both ways as producing nations compete for refinery investment.
The other cost inputs round out the picture. Prebaked anodes (East China) are ¥7,362.5/t (+4.9% YTD), with the anode industry's own margin razor-thin near break-even (SMM's instant-profit read sits around ¥15/t) — SMM expects a modest June decline as earlier cost weakness passes through. Power, typically around 40% of the smelter cost, fell month-on-month into May as southern China's wet (flood) season approached, materially cutting electrolytic-aluminium power cost. The net effect of cheap alumina, easing anodes and seasonal hydropower is a cost base near its cyclical lows — the foundation of the margin story in Chapter 3.
The single largest forward-looking tail-risk in the entire chain sits in West Africa. Guinea now supplies the majority of China's imported bauxite, and that dependence is the upstream mirror of China's own supply ceiling: a structural single-point-of-failure. The country's history of political instability — most recently the 2021 coup — and a record of abrupt resource-policy interventions mean a Guinea disruption is the cleanest path to a renewed alumina squeeze. The 2024 episode deserves a precise reading, however: the primary driver of that year's alumina surge was domestic — Chinese refining supply simply could not grow fast enough to keep pace with demand, as capacity additions lagged the smelter pull — and Guinea's export and logistics frictions acted only as a marginal amplifier on an already-tight balance. The alumina index blew off to ¥5,700/t and dragged the smelter cost base to roughly ¥21,400/t in a matter of months. The lesson is therefore prospective rather than historical: with refining capacity now ample, a future repeat would require a genuine bauxite-side interruption, and Guinea — through the Winning Consortium's Boké operations and the expanding 90+ Mt guidance noted earlier — is precisely the node whose interruption could reverse the entire margin story. For a forum focused on secure supply chains, bauxite concentration in Guinea is to aluminium what Australian–Brazilian concentration is to iron ore — except with materially higher political risk and far less spare capacity elsewhere. China's strategic response is twofold: domestic stockpiling, and the steady diversification toward Australian and (prospectively) other West African tonnes — Indonesia, which has banned raw-bauxite exports since 2023, is not an available fallback.
Alumina refining is the chain's pressure-release valve. China's refining capacity — concentrated in Shandong, Shanxi, Henan, Guangxi, Guizhou, Chongqing and Hebei — has expanded fast enough that the national weekly operating rate can sit at 72.7% while still oversupplying the smelters; idle capacity is abundant rather than absent (operating capacity ~86 Mt against far higher built capacity). That spare refining capacity is precisely why the 2024 squeeze reversed so violently: as new Guangxi lines ramped and seaborne tonnes arrived, the balance flipped from deficit to surplus inside two quarters. The import valve is the other half of the mechanism: China's monthly alumina imports surged to 610 kt in April 2026 (+168% on a year-to-date basis). Two forces drove that surge: cheap FOB West Australian and Indonesian alumina opened the import arbitrage, and at the same time geopolitical conflict constrained Middle-East-bound exports, so overseas tonnes were redirected toward China. The practical implication for the smelter is comfort: with both domestic spare capacity and an open import window, a repeat of the 2024 cost shock requires a genuine bauxite-side disruption — not merely strong metal demand. That is why this report treats Guinea, not Chinese refining, as the upstream variable most worth watching.
China's primary aluminium is frozen at a policy ceiling and running at near-ceiling utilisation. With costs collapsed, the result is a rare combination: no supply growth and the fattest margins in years.
The defining feature of Chinese aluminium supply is the capacity policy ceiling, imposed in 2017 under supply-side reform to curb overcapacity and emissions. It is not a soft target: SMM data show operating capacity essentially at the policy cap as of May 2026, with operating rates at historic highs and almost no headroom left in the system. China's primary aluminium output, in other words, has next to no room to grow. New capacity is replacement-only unless Chinese policymakers choose to exempt renewable-powered smelters, and even then the constraint is binding for years.
Monthly output of 3.86 Mt in May 2026 (+2.1% YoY) represents record run-rates against a fixed cap — the supply curve is, for practical purposes, vertical. This is the structural floor under the aluminium price: unlike iron ore or even copper, the dominant producer simply cannot add primary tonnes. Growth must come from offshore or from recycling (Chapters 5 and 7).
It is worth being precise about the mechanism, because it is what makes aluminium structurally different from every other base metal. The policy ceiling is enforced as a hard quota on operating capacity: any new capacity must be matched by the retirement of an equivalent quantum of capacity entitlement. This has created an internal market in "replacement capacity" — entitlements are tradable, so the national total can only be reshuffled (typically from high-cost coal-power regions toward low-cost hydro regions like Yunnan), never expanded. The live policy debate is whether the Chinese government will carve out an exemption for fully renewable-powered smelters as part of its dual-carbon agenda; even if it does, the permitting, construction and ramp timeline means the constraint binds hard through at least the end of the decade. The investment implication is stark: a dollar of capital spent on Chinese primary aluminium buys no incremental tonnes, only a cheaper or greener version of existing ones. That is why the marginal growth dollar — and the marginal new tonne — is migrating offshore, the subject of Chapter 9.
Within China, Yunnan province — roughly 15% of national output — is the recurring wildcard. Its dependence on hydropower creates a seasonal cut risk: in severe dry years (2023) up to 40% of provincial capacity has been curtailed. The current wet season is supportive for both output and cost, but the dry-season cut is a perennial Q4–Q1 risk that can tighten the domestic balance and spike the SHFE price independently of the demand cycle. The migration of capacity toward Yunnan under the replacement-only regime — a deliberate shift from coal-power regions to hydro — is climate-positive but raises the system's exposure to a single province's rainfall, a concentration risk worth pricing as the energy transition pulls ever more smelting onto hydropower.
The cost-and-profit picture is the heart of the 2026 story. SMM's cost model puts the all-in cash cost at roughly ¥16,000/t and immediate profit near ¥7,800/t; the monthly all-in cost fell to ¥15,966/t in May (−1.9% MoM, −2.2% YoY), with 100% of operating capacity profitable. For comparison, the cost model peaked near ¥21,400/t in December 2024 when alumina spiked. The lesson for the forum is important: this is a margin cycle driven by the cost side, not a price-squeeze driven by demand. That makes the margin more durable than a demand-led rally would be — it persists as long as alumina and power stay cheap — but it also means the margin, not the metal price, is where the risk sits. A renewed alumina squeeze would compress profit even if the LME price held.
The fat domestic margin also has a global dimension that is easy to miss. Chinese smelters, especially the hydro-powered Yunnan cohort, now sit in the lower half of the global cost curve, while large parts of the Western and Middle-Eastern fleet — burdened by post-2022 power costs — sit in the upper half. The consequence is that the high LME price is not, at present, calling forth a wave of ex-China restarts: the European capacity shuttered in the energy crisis remains largely cold because the economics of restarting an idled potline at current power prices are marginal, and a restart is a multi-quarter, capital-intensive commitment that managements are reluctant to make against an uncertain forward curve. This is the supply-side reason overseas tightness persists even at a multi-year-high price. For the smelter the read-through is comfort; for the market it is a warning that the usual price-elastic supply response is broken on both sides — capped in China by policy, frozen offshore by cost and caution. A market that cannot easily add supply at the top of the cycle is a market whose disruption-driven tail risk is skewed to the upside relative to the base path, which is the single most important structural fact for a hedger to internalise. It also reframes the inventory glut: China's 1.36 Mt is not spare global capacity waiting to flood the seaborne market — it is metal trapped behind China's 30% export duty on primary aluminium and the semis-level trade frictions described in Chapter 5, which is why it can coexist with a multi-year-high LME price rather than arbitraging it away.
The lesson generalises beyond a single quarter: in a market with a frozen dominant producer and a cautious, high-cost offshore fleet, supply elasticity is asymmetric — abundant capacity that cannot grow versus scarce capacity that will not. That is why aluminium's volatility profile has changed shape, with sharper, faster upside spikes on any disruption and a firmer floor on any dip, and why a hedger should treat the cost stack and the inventory line, rather than the flat LME print, as the primary risk gauges.
A record Chinese domestic inventory glut sits against the tightest overseas market in years. The divergence is the single most important — and most tradable — feature of aluminium in 2026.
If the supply ceiling is the bull case, and cheap alumina merely fattens the smelting margin — not a pricing anchor, for primary aluminium is supply-constrained and far from a surplus commodity, and does not price off its cost curve — then inventory is the brake, and the place where the market's internal contradiction is most visible. Chinese ingot social inventory has roughly doubled year-to-date to 1.36 Mt (from 0.49 Mt at end-2024 and 0.66 Mt at end-2025), peaking at ~1.43 Mt in spring on seasonal builds and weak property demand. Billet social inventory is 0.16 Mt, and SHFE on-warrant stock has rebuilt sharply (+~500% YTD) — all signs of ample, deliverable domestic spot. SMM has been explicit that the inventory turning point is the key domestic variable it is watching.
The mirror image sits offshore. LME stocks have fallen roughly 36% year-to-date to ~328 kt, leaving the international market structurally tight — the backdrop to the Q1 supply-shock spike. The result is a genuinely two-speed market: a Chinese domestic glut that caps the SHFE price and the A00 premium, against an overseas shortage that supports the LME price and underpins China's role as a structural net importer of metal (Chapter 5).
China's glut reflects capped-but-full production meeting soft domestic (property) demand; the overseas tightness reflects years of smelter closures in Europe and disruptions in the Middle East with no China-scale capacity behind it. The two do not arbitrage away quickly because China's 30% export duty keeps the primary-metal channel shut, while semis face their own rebate and tariff frictions (Chapter 5). The spread — short SHFE / long LME, or watching the import arb — is the cleanest expression of the 2026 aluminium thesis, and one that clears through Singapore.
Chinese aluminium inventory follows a pronounced annual rhythm, and 2026's level — not its seasonality — is what is unusual. Stocks build through the winter as downstream fabricators shut for the Lunar New Year while smelters, which cannot economically idle, keep running flat-out; they then draw down through the spring and summer peak-construction season. The 2026 build carried the social total from 0.49 Mt at end-2024 and 0.66 Mt at end-2025 to a spring peak of ~1.43 Mt — a higher base than either prior year. The three-region daily estimate of ~1.15 Mt and the sharp SHFE on-warrant rebuild (warrant stock up roughly 500% year-to-date, to ~489 kt) corroborate a market awash in deliverable domestic spot. SMM has been explicit that it is watching for the seasonal draw to assert itself: the pace of the summer draw, more than the absolute level, is the cleanest near-term read on whether electrified demand is finally biting.
The reason the China-glut / overseas-tight divergence does not simply arbitrage away is mechanical. China levies a 30% export duty on primary aluminium, so the channel for unwrought metal to flow offshore is, in practice, welded shut — the principal reason the domestic glut cannot drain into the seaborne market to flatten the spread. At the semis level, the removed export rebate (Chapter 5) and, for US-bound tonnes, a punitive Section 232 tariff layer further friction on top; moving LME metal into China requires the import arbitrage to open, which the domestic glut keeps shut. So the two pools sit side by side at different prices — a structural, not transient, feature. For a trading desk this is the single most reliable relative-value structure in the complex — provided the 30% export duty on primary aluminium and the current semis-rebate regime stay as they are: the SHFE–LME spread and the physical import arbitrage are where the 2026 thesis is expressed, while the flat price stays range-bound between the supply floor and the inventory cap.
The practical signal for the second half lies in how far the destocking pace deviates from its seasonal norm. A peak-season draw running faster than the usual seasonal rhythm would say that aggregate demand — grid, autos and industry together — is finally absorbing the capped-but-full supply, a bullish read; a sluggish peak-season draw, or worse a renewed build, would say that the broad-based softness across property, solar and the other lagging segments is still winning, and the upside stays capped. The direction, slope and persistence of the inventory line carry more information than any single threshold ever could. The inventory line is the scoreboard. What is putting metal into, or pulling it out of, those warehouses is the subject of the next two chapters.
China runs flat-out against the cap, imports metal as a structural net importer, and has seen its semis exports moderated by the 2024 rebate removal. Growth is migrating offshore.
The trade data tells the story of a producer that has hit its limit. China ran a domestic primary-aluminium deficit in 2025 (output ~44.2 Mt against consumption ~46.3 Mt); for 2026, SMM's balance model shows the domestic balance roughly squared — broadly at par rather than in deficit (measured as output plus net imports against consumption; the year-to-date inventory build means the first half has in practice run looser than balanced, so a squared full year rests on the seasonal H2 destock coming through). China nonetheless remains a structural net importer of metal: primary aluminium imports were 265 kt in April 2026 (net ~250 kt), and 2025 unwrought-plus-products imports reached 3.92 Mt. The country that makes nearly 60% of the world's primary aluminium must buy metal from abroad — a direct consequence of the capacity policy ceiling meeting electrified demand.
On the other side of the ledger, the global picture is one of disrupted-but-recovering supply. IAI global primary output was 5,922 kt in April 2026; SMM notes overseas (ex-China) output was down 9.9% year-on-year on Middle-East cuts but up 0.4% month-on-month as US, Spanish, Icelandic, Indonesian and Indian (Balco) capacity resumed or ramped. Alumina imports into China surged to 610 kt in April as cheap overseas tonnes arrived — partly because low FOB West Australian and Indonesian prices opened the import arbitrage, and partly because geopolitical conflict constrained Middle-East-bound exports and redirected overseas cargoes toward China. The alumina-import swing is itself a useful tell: China flips between net importer and net exporter of alumina depending on import profitability — when the import window is open, overseas tonnes flow in; when it closes, flows reverse toward export. The deeper root of this raw-material comfort is alumina overcapacity: built refining capacity far exceeds what the smelters require, and the surplus presses on prices at home and abroad and keeps the arbitrage channels open. The abundance on the raw-material side is, at bottom, an expression of that overcapacity — and the quiet underwriter of the smelter margin. It is also the reason a margin shock, if it comes, is far more likely to originate in a bauxite disruption upstream than in any tightness in the refined-alumina market itself.
The most important policy event for downstream trade was the removal, effective 1 December 2024, of the 13% export tax rebate on roughly two dozen aluminium semis categories (sheet, strip, foil, tube, some bars). The immediate effect was a +37% year-on-year surge in exports in November 2024 (public market data) as buyers front-loaded ahead of the deadline; the lasting effect was a −8% decline in full-year 2025 semis exports. The rebate removal was not, however, the whole story. The United States raised its Section 232 aluminium tariff to 50% in 2025, and the drag this exerted on US-bound shipments — and on the transshipment flows that had previously routed Chinese semis toward the American market through third countries — was comparably heavy. The −8% print is best read as the compound effect of the two policies: a domestic decision that withdrew the export subsidy and a foreign one that walled off the largest premium market. Plate, sheet and strip make up over half of semis exports, so the combined shock moderated rather than collapsed flows — but it accelerated China's structural pivot from export-led to domestically-balanced aluminium demand. (A partial offset arrived in May 2026, when China removed a 15% export duty on certain semi-processed bars and rods.) For the rest of the world, less subsidised Chinese semis is a margin tailwind for local fabricators — and a reason overseas tightness persists.
Layered on top of the rebate shock is a second policy distortion: the United States lifted its Section 232 aluminium tariff to 50% in 2025, walling off the world's largest net-import market and forcing a re-routing of global trade flows. The combined effect of the US tariff wall and China's rebate removal is a more fragmented, regionalised aluminium market — exactly the backdrop a "secure supply chains" forum is convened to address. Tonnes that once flowed China → world and world → US now seek new homes, widening regional premium dispersion (the US Midwest premium, the European duty-paid premium and the Japanese MJP each now tell a more local story). For physical traders this fragmentation is opportunity; for end-users it is supply-security risk; for ASEAN it is an invitation to become the neutral, low-friction hub that re-aggregates these flows — a theme we return to in the Singapore chapter.
The arithmetic of China's balance is worth stating plainly because it is the crux of the bull case. Chinese output of roughly 44.2 Mt in 2025 met consumption near 46.3 Mt — a deficit of about 2.1 Mt that imports filled. For 2026, SMM's balance model shows the domestic balance roughly at par — broadly squared rather than in deficit, on the same output-plus-net-imports-versus-consumption basis, and with the same caveat that the year-to-date stock build implies a loose first half and a squared full year hinges on the seasonal second-half destock materialising — but the asymmetry matters: because domestic output is frozen at the ceiling, any acceleration in electrified demand would tip the balance straight back toward tightness, and the gap could not be closed from within China. The only sources of incremental primary metal are the offshore ramp — Indonesia (roughly 4.5 Mt of Chinese-owned smelting capacity planned and under construction, with commissioned lines still ramping), India (Vedanta, Hindalco and the Balco 525 kA expansion taking capacity to roughly 1 Mt, of which about 435 kt is net new) and the gradual restart of idled Western capacity — and recycling. As noted, overseas supply is healing sequentially from the Middle-East outage. The race between that offshore ramp and Chinese demand growth is, in essence, the multi-year supply-demand question for aluminium.
The synthesis of supply and trade is clear: China has next to no room to grow primary output, ran a ~2.1 Mt import-filled deficit in 2025 and — with the 2026 balance roughly squared — remains a structural net importer, while exporting less fabricated metal than it used to. The marginal new primary tonne, therefore, is increasingly an Indonesian tonne — the offshore growth vector we return to in Chapter 9. Demand is what determines whether the capped supply tightens or loosens, and that is the next chapter.
The demand story is a tug-of-war: structural growth from the power grid, EVs and energy storage against a still-weak construction sector — with the solar pull fading after the mechanism-price reform.
Aluminium demand in 2026 is defined by a rotation, not a collapse. The traditional pillar — construction and property — remains weak, mirroring the steel story, and is the reason domestic inventory has built. But the metal's unique exposure to the energy transition is the offset: power-grid investment in aluminium wire, cable and transformers, transport / EV lightweighting plus battery enclosures and battery trays, and rapidly scaling energy storage are the structural growth engines. Solar deserves its own, more sober sentence: the PV pull on aluminium has come off its peak — with the mechanism-price reform (Document No. 136) in force, installation expectations have been revised down, and the pull weakens markedly from 2026 onward. Aluminium remains a cleaner energy-transition play than most base metals, because it sits in the grid, the storage container and the vehicle simultaneously.
The high-frequency read confirms a peak-season recovery. SMM's weekly leading-enterprise survey shows the aluminium processing industry average operating rate at 64.0%, up 2.6 percentage points year on year, with extrusion at 57.6% (+0.6 pct points YoY), foil at 73.3% and billet leaders at 64.5%. The extrusion read, however, needs to be decomposed rather than taken at face value: it blends construction (weak), industrial (relatively resilient) and solar (off its peak) — and the recovery in the rate owes more to grid and industrial extrusion orders than to PV frames and mountings.
It helps to decompose where aluminium actually goes, because the headline demand number conceals a violent internal rotation. Construction and property — historically the largest single end-use at roughly a quarter to a third of Chinese consumption — is the weak leg, dragging on extrusion and window/door demand and mirroring the steel story. Transport (around a quarter) is bifurcated: traditional auto is soft, but EV lightweighting is a structural pull, with battery-electric vehicles using materially more aluminium per car (in body, chassis, battery enclosure and busbar) than their combustion predecessors — and on 2025's high sales base, the extension of the trade-in programme and first-time owners rolling into their replacement cycle should keep releasing replacement demand through 2026. Power and electrical is the heart of the electrification thesis, but the largest incremental contribution within it now comes specifically from grid investment — cable, conductor and transformer — rather than from solar. The PV story has been marked down: following the mechanism-price reform (Document No. 136), China's 2026 new-installation outlook has been cut to around 210 GW, and SMM does not expect annual additions to regain the 2025 level until around 2030. A single gigawatt of solar capacity embodies roughly 11,000–19,000 tonnes of aluminium across frames, mounting structures and cabling (estimates vary with scope), and China's annual additions remain on the order of 200 GW — substantial, but now below the 2025 peak, so the PV increment to aluminium demand steps down rather than compounds. Packaging (can stock and foil) is a steady, defensive ~10%. The net picture is a demand base whose weak leg (property) is shrinking as a share while its durable legs (grid, EV, storage) compound — which is why SMM frames aluminium as the base metal most directly levered to the energy transition.
The segment picking up the baton from solar — a hand-off under way, its substance still to be proven — is energy storage, the fastest-growing aluminium end-use of 2026. Grid-scale ("big storage") projects, procured alongside renewable build-out and grid reinforcement, and commercial-and-industrial systems, whose economics improve as time-of-use power tariffs widen, are both scaling rapidly from a small base. The aluminium intensity is real and spread across the bill of materials: extruded structural frames and racking, battery-pack and container enclosures, busbar and cabling, and battery foil for the cells themselves. Storage draws on much of the same extrusion and foil capacity that PV frames once filled, which is precisely why fabricators describe it as a partial hand-off rather than a separate market: order books that were losing PV frame volume are refilling, in part, with storage enclosures and racking. Storage cannot yet match the sheer tonnage that solar contributed at its peak — the base is still too small — but its growth rate is the highest of any demand segment this year, and its policy tailwind (renewables need firming; the grid needs flexibility) is, if anything, strengthened by the same reform that trimmed PV economics.
The PV downshift is a specific, datable policy event, not a vague souring of sentiment. The mechanism-price reform (Document No. 136) moved new renewable projects from administered tariffs to market-based mechanism pricing, recasting project economics and prompting developers to mark down their build plans: 2026 new installations are now expected around 210 GW, and SMM does not expect annual additions to return to the 2025 level until around 2030. For aluminium, the read-through is a markedly lower PV pull on extrusion (frames, mountings) and cable from this year — a downshift that energy storage offsets only in part. Sizing that residual gap, against the grid and EV pull, is the key demand-side judgement for the back half of 2026.
The demand picture also has a top-down dimension that turned modestly more constructive into mid-2026. China's Q1-2026 GDP grew 5.0%, the PPI registered its first increase after 41 consecutive months of deflation — an important signal that industrial pricing power may be bottoming — and the IMF pencilled global growth at 3.1% for the year. None of this is a boom, but for a metal whose weak leg is property, the relevant question is whether construction has found a floor rather than whether it re-accelerates. Chinese policymakers' current property stance centres on urban renewal, destocking and activating the existing housing stock — optimising supply rather than adding visible incremental stimulus; the aim is to stabilise the floor, not to re-ignite the cycle. If that floor holds, the property drag on aluminium demand stops worsening, and the electrification legs are then sufficient to tip the balance toward an inventory draw. The bull case does not need a property recovery — it needs only the floor to hold and the PV installation downgrade (around 210 GW now expected for 2026 after the mechanism-price reform) to prove a manageable, rather than compounding, drag; the bear case is, at bottom, a bet that the floor gives way.
The demand verdict for the forum: the electrification thesis is real and durable, but 2026 demand growth is lumpy and internally divergent. The solar downshift traces to one specific policy — the Document No. 136 mechanism-price reform — rather than to any generic "policy sensitivity"; EVs, sitting on 2025's high sales base, should release additional replacement demand through the year; and energy storage is the new bright spot, the fastest-growing segment of 2026, picking up the aluminium baton from solar — though the substance of that hand-off is still to be proven. Aggregate demand has not yet been strong enough to draw down the domestic inventory glut. That makes aluminium a supply-and-margin story in the near term and a demand story in the long term — the capped supply and cheap costs hold the floor now, while grid, EV and storage demand is the mechanism that eventually clears the inventory and re-tightens the market. The recycling and carbon dimensions, which increasingly shape both supply and demand, are next.
Secondary aluminium is the low-carbon future and the scrap-constrained present. CBAM and a green-aluminium premium are beginning to re-price the metal by its carbon intensity, not just its cost.
Aluminium is the great recycling metal — remelting scrap uses about 5% of the energy of primary smelting — so secondary metal is central to both the decarbonisation story and the supply balance. Yet in 2026 the secondary sector is under pressure. SMM's ADC12 secondary alloy price is ¥23,900/t (+6.5% YTD), tracking primary higher, but the industry-wide secondary-aluminium operating rate has fallen to 46.6% (May 2026, down roughly 4.0 percentage points year-on-year), squeezed by tight, expensive scrap and thin margins.
Tight scrap supply and elevated costs remain the secondary sector's principal constraints through 2026. The clearest price evidence is in the spread: ADC12 has converged to parity with A00 (spread ≈ 0, June 2026) — the traditional secondary discount to primary fully consumed by scrap costs. The sector is strategically essential to decarbonisation yet financially stressed in the near term, and for now secondary supply is a constraint on, not a contributor to, easing the 2026 balance.
EU scrap exports rose roughly 50% from 2019 to 2024 (to ~1.26 Mt), and the EU, US, Japan, UAE and South Africa are all moving to retain scrap as a strategic resource. For Asian ADC12 producers dependent on imported scrap, this is a structural supply risk — the same scrap-nationalism dynamic playing out in steel. Whoever controls scrap controls the low-carbon supply of the future.
The carbon dimension is no longer theoretical. The EU's Carbon Border Adjustment Mechanism entered its definitive charging phase on 1 January 2026, and aluminium is one of six covered sectors. With charging now live, CBAM puts a new marginal cost on China's EU-bound aluminium. The more important effect is structural: CBAM is accelerating a green-aluminium premium and a market bifurcation. European hydro-based primary aluminium carries a carbon footprint significantly below the global average, while coal-based primary runs several times higher (IAI basis, indicative).
The mechanism matters for who pays and how much. Under the definitive regime, EU importers must surrender CBAM certificates for the embedded emissions of covered aluminium — certificate prices are set by reference to the weekly average of EU ETS auction carbon prices — with the charge phased in as the free allocation to EU producers is withdrawn through 2034. Importers either document verified actual emissions or accept conservative default values — and the defaults are deliberately punitive, so the practical effect is to force carbon accounting onto the supply chain. For a coal-powered smelter the embedded figure can imply a levy on the order of several hundred euros per tonne at plausible ETS prices (the precise number turning on the carbon price and the verified emissions), large enough to swing the delivered economics of a tonne of metal. Yet the modelled near-term aggregate cost to China is modest — a small fraction of the value of China's EU-bound exports — because China sends relatively little primary metal directly to Europe. The strategic significance is therefore not the first-year bill but the price signal: CBAM establishes, for the first time, that a tonne of low-carbon aluminium is worth measurably more than a chemically identical high-carbon tonne. That signal is what creates the green premium and the bifurcated market.
The corporate response is already visible: hydro- and solar-powered "green" brands (Rio Tinto, EGA CelestiAL, the Alcoa–Rio Tinto ELYSIS inert-anode venture), a Rotterdam green premium, and China scaling its own low-carbon aluminium certification, led by the China Nonferrous Metals Industry Association (CNIA) — though any relief under CBAM ultimately depends on EU recognition, and China's domestic low-carbon certificates are not currently recognised by the EU. The strategic read for the forum: over the decade, aluminium's value will be set on two axes — cost and carbon — and the producers who can credibly document certified low-carbon metal will earn a structural premium. That is a Singapore green-finance and certification opportunity as much as a metallurgical one.
How large can the green premium become, and who captures it? Today the differential for certified low-carbon primary aluminium is modest and inconsistently quoted — a Rotterdam premium of a few tens of dollars per tonne over the LME — but its trajectory, not its current level, is what matters. The premium is anchored at the bottom by the CBAM certificate cost a high-carbon competitor must pay (which rises as EU free allocation withdraws through 2034) and at the top by the cost of actually decarbonising a smelter (hydro relocation, solar PPAs, inert-anode retrofit). As the floor rises mechanically with the carbon price, the premium widens almost regardless of demand. The crucial unresolved question is the unit of account: a premium expressed per tonne of metal rewards primary producers who switch to clean power, while one expressed per tonne of avoided CO₂ would reward recyclers far more — so how the premium is defined will itself decide who captures it. For SMM, the commercial opportunity is to build the assessment, the index and the verified-emissions dataset that let this premium be priced, hedged and financed transparently — turning a qualitative ESG aspiration into a tradable, auditable differential. Whoever owns that benchmark owns the carbon axis of aluminium value, and the contest to define it is already under way.
SMM frames the 2026 outlook through supply-demand balance, the inventory trajectory, raw-material disruption risk and the macro (Fed) path — a scenario framework, not a point forecast.
SMM's price view for 2026 rests on four pillars, and deliberately not on the cost curve. First, the balance: the domestic supply-demand account is roughly squared this year — output frozen at the policy ceiling, demand growing but lumpy — and the global market is close to balance on a roughly 75 Mt base. A balanced market with no supply elasticity on either side (China capped by policy, the Western fleet frozen by economics) is a market that trades on inventory and disruption headlines, which is exactly the price behaviour of 2026. Second, the inventory trajectory: the 1.36 Mt domestic glut is the live brake, and the direction, slope and persistence of the draw — not any single level — will decide whether the structural bull case is released or stays contained. Third, raw-material disruption risk: with refining capacity ample, the genuine tail sits in bauxite — above all Guinea — and an upstream interruption would squeeze the smelter margin first; overseas smelter outages, by contrast, pass through to the LME almost undamped. Fourth, the macro path: expectations of Federal Reserve tightening weigh on metal valuations across the complex, and the timing of any pivot is the principal macro swing factor for dollar-denominated pricing. For completeness: the distribution of market expectations for 2026 — roughly US$2,900–3,700/t across unattributed public forecasts — has converged toward near-balance readings of the same fundamentals.
It is worth spelling out why SMM does not anchor its aluminium view to the cost curve, useful as the cost model is for reading margins. Cost-floor logic governs prices in structurally oversupplied markets, where the marginal loss-making tonne eventually exits and the curve sets the clearing level. Primary aluminium in 2026 is the opposite case: supply is constrained by policy, the dominant producer cannot respond to price, and the metal is far from a surplus commodity — so the gap between price and cost is set by the supply-demand balance and the inventory picture, not by any cost-side gravity. The alumina collapse fattened the smelter margin; it tells you nothing, by itself, about where the metal will trade. With smelting economics comfortable across effectively the whole of Chinese operating capacity, the analytical question is not whether smelters are viable but how much premium above cost a balanced, disruption-prone, inventory-heavy market will pay — and that question is answered by the four pillars above, which is why the outlook below is built as scenarios around the balance rather than as a band around the cost line.
A sustained, established destocking trend takes hold in China as electrified demand bites; LME tightness persists or worsens on Middle-East/logistics disruption; Yunnan dry-season cuts tighten the domestic balance. Capped supply + a confirmed inventory downtrend = breakout.
The two-speed market persists: capped supply and fat margins hold the floor, while the 1.36 Mt domestic glut caps the upside. Near-balance globally. A00 ¥23,000–25,000/t; margins stay rich on cheap alumina. The base range (US$3,000–3,400) is a 2H26 trading-range call: as Middle-East supply normalises, part of the Q1 disruption premium unwinds; "firm" means a solid floor and shallow drawdowns, not further upside from spot. Treat US$3,400–3,700 as the transition band between base and bull: a price lingering there says the disruption premium has not yet unwound, but no fresh evidence of shortage has emerged either.
The solar downshift cuts deeper than storage and grid demand can offset; property stays weak; peak-season destocking fails to materialise and stocks keep building; Indonesian/Indian capacity loosens the global balance faster than expected; geopolitical de-escalation unwinds the disruption premium. A renewed alumina squeeze would hit margins even if the price held.
The base case breaks up if a sustained Chinese destocking trend establishes itself while the overseas market stays tight — i.e. demand finally clears the glut into a capped market. It breaks down if the peak season passes without a meaningful draw — stocks building against the seasonal pattern with warrants rebuilding alongside. The margin leg needs separate monitoring: an alumina re-spike above ¥3,500/t would compress the margin that is the whole story — but that hits profits, not the price direction. The distinctive feature of aluminium is that the margin and the price can diverge: a forecaster must track the cost stack (alumina, power) as closely as the metal price, because in 2026 the profit — not the LME print — is where the cycle lives.
Distilled to a single dashboard, four indicators carry most of the forward signal through end-2026. (1) Chinese social inventory — the scoreboard: read it as a trend, not a level — a confirmed, persistent destocking trend through the peak season is the bull signal; a failed peak-season draw (no destock, or a renewed build) with warrants rebuilding alongside is the bear confirmation, off-season builds counting only as corroboration. (2) The SMM alumina index — the margin's pulse, a variable that moves profits rather than the price direction: stability in the ¥2,500–3,000/t band keeps the profit cycle intact, ¥3,000–3,500/t is the neutral observation band, and a re-spike above ¥3,500/t is the single fastest way to gut smelter margins regardless of the metal price. (3) LME stocks and the SHFE–LME arbitrage — the two-speed gauge: continued overseas drawdowns confirm persistent ex-China tightness, and the arb tells you whether the two pools are about to converge. (4) The Rotterdam green premium and CBAM certificate price — the carbon axis: the speed at which a low-carbon differential establishes itself will determine how quickly the market bifurcates. A reader who tracks just these four — inventory, alumina, the LME-stock/arb pair, and the green premium — will understand the aluminium cycle better than one watching the flat price alone. That, ultimately, is the analytical message of this report: in 2026, aluminium is a market to be read in its structure, not its headline.
Aluminium is where supply policy, electrification and carbon pricing converge — and where the growth is migrating to ASEAN. Singapore prices, finances and certifies the transition.
The structural forces in this report converge on Singapore and its region. With China's primary capacity frozen at its policy ceiling and the country a net importer, the marginal new primary tonne is increasingly an Indonesian one: Chinese-owned Indonesian smelting capacity planned and under construction totals roughly 4.5 Mt — commissioned lines are still ramping, and the build-out is far from complete — with company plans targeting 13 Mt by 2035 (compiled from public disclosures, indicative). The aluminium growth story, in other words, is an ASEAN story — and the capital, the offtake and the trade clear through Singapore.
This migration is the aluminium counterpart to the iron-ore Simandou story: capped or constrained supply at home pushing investment offshore to lower-cost, often coal- or gas-powered geographies — which collides directly with the carbon-pricing trend. The central tension of the next decade is that the cheapest new aluminium (ASEAN, coal-powered) and the most valuable new aluminium (low-carbon, certified) are pulling in opposite directions, and Singapore is where that tension is priced and financed.
The offshore build is not speculative; it is under construction. Indonesia is the centre of gravity: integrated bauxite-alumina-aluminium projects, several Chinese-owned or Chinese-financed, are clustering around the country's industrial parks, drawn by domestic bauxite, captive (largely coal) power and a government determined to move up the value chain after its ore-export ban. India is the second vector, where Vedanta and Hindalco are expanding and the Balco 525 kA expansion takes capacity to roughly a million tonnes (about 435 kt of it net new) — growth powered by domestic coal and bauxite. The Middle East (EGA, Alba, Ma'aden) rounds out the gas- and increasingly solar-powered tier. The strategic problem this creates is a carbon paradox: the lowest-cost incremental tonnes are also, on current power mixes, among the highest-carbon — which is precisely the metal that CBAM and the green premium will penalise. Resolving that paradox — financing the build while greening its power source — is the defining commercial challenge of the aluminium transition, and it is a challenge that runs through ASEAN's capital and trading hub.
Three Singapore franchises matter. First, price discovery and hedging: the LME complex and its Asian liquidity, the SHFE-offshore relationship, and the growing need to price a green-aluminium differential that the flat LME contract cannot express. Second, green and trade finance: financing both the ASEAN capacity build and the low-carbon transition (inert-anode, solar-powered, recycling) — and the certification that determines which tonnes earn the CBAM-relief premium. Third, the scrap and recycling supply chain, as scrap becomes a contested strategic resource across the region. For SMM, whose A00, alumina and ADC12 assessments increasingly underpin physical and financial contracts, Singapore is where the data becomes the infrastructure of the trade.
The timing is not incidental. Singapore International Ferrous Week convenes these conversations — across the Green Energy Metal Forum, the green-finance track and the trade-and-supply-chain sessions — at the precise moment the aluminium market is fragmenting along policy and carbon lines that no single exchange contract yet prices. A green-aluminium differential, a CBAM-compliant certification standard, a financing structure for low-carbon ASEAN capacity, and a transparent benchmark for the SHFE–LME arbitrage are all, in effect, products waiting to be built; each is a natural fit for a neutral, well-regulated, deeply-capitalised hub sitting between the Chinese, ASEAN and Western poles of the market. The argument of this report is that whoever builds the data, the benchmarks and the finance for the two-axis (cost-and-carbon) aluminium market will capture a disproportionate share of its value — and that the most credible candidate to host that infrastructure is Singapore.
The cycle is a margin cycle: protect it by locking low-cost alumina and power, and prepare for a world where carbon intensity is priced. Low-carbon certification (hydro, solar, inert-anode, recycled content) is the route to the structural green premium. The offshore growth option is ASEAN — but with a carbon cost.
The alpha is in the China-glut / overseas-tight divergence: the SHFE–LME spread, the import arb, the A00 premium, and the emerging green differential. The flat price is range-bound; the structure is where the opportunity sits — and Singapore is where it clears.
Electrification demand (grid, EV, energy storage) is the durable pull; the export-rebate removal reshapes where semis are made. Securing certified low-carbon metal and recycled content is now a commercial as well as an ESG decision, especially for any product bound for CBAM-covered Europe.
The ASEAN capacity build, the recycling supply chain and the low-carbon transition all need capital — a Singapore green- and trade-finance franchise. The policy agenda is robust carbon accounting, scrap-resource security, and a credible certification regime that rewards genuine decarbonisation.
Aluminium in 2026 is supply-capped, margin-rich, demand-electrified, inventory-watched and carbon-priced. China's capacity policy ceiling and fat, alumina-driven margins hold a structural floor; a 1.36 Mt domestic glut against overseas tightness caps and bifurcates the market; electrified demand is the long-term clearing mechanism; and CBAM is beginning to price the metal by its carbon as much as its cost. The growth migrates to ASEAN, the value migrates to low-carbon, and both clear through Singapore. For the Green Energy Metal Forum, aluminium is the metal where the energy transition's promise and its contradictions are most visible — and most investable.
| Indicator | Latest | Unit | Date | SMM / source ID |
|---|---|---|---|---|
| SMM A00 spot | 23,900 | ¥/t | 2026-06-09 | s20003505 |
| LME aluminium 3M | 3,607.5 | $/t | 2026-06-08 | a10003557 |
| SHFE main contract | 24,110 | ¥/t | 2026-06-09 | a10022370 |
| A00 premium/discount | -70 | ¥/t | 2026-06-09 | s20003511 |
| SMM alumina index | 2,691 | ¥/t | 2026-06-09 | a10171865 |
| Alumina futures (SHFE) | 2,838 | ¥/t | 2026-06-09 | a10168364 |
| Guinea bauxite CIF | 69.5 | $/t | 2026-06-09 | s20098953 |
| Prebaked anode (E. China) | 7,362.5 | ¥/t | 2026-06-09 | s20003683 |
| Smelting total cost | 16,236 | ¥/t | 2026-06-09 | a10004081 |
| Smelting immediate profit | 7,764 | ¥/t | 2026-06-09 | a10004082 |
| Monthly all-in cost (national) | 15,966 | ¥/t | 2026-05 | a12818129 |
| China output (monthly) | 3.86 | Mt | 2026-05 | a10004370 |
| Yunnan output | 0.587 | Mt | 2026-05 | a10148350 |
| Ingot social inventory | 1.36 | Mt | 2026-06-08 | a10004205 |
| SHFE warrant | 488,605 | t | 2026-06-09 | a10022641 |
| LME stocks | 327,750 | t | 2026-06-08 | a10004095 |
| Processing operating rate | 64.0 | % | 2026-06-04 | a10031804 |
| Secondary-Al industry op rate | 46.61 | % | 2026-05 | a10004257 |
| ADC12 (SMM) | 23,900 | ¥/t | 2026-06-09 | s20003650 |
| Net primary Al imports | 249,836 | t | 2026-04 | a10004512 |
| Alumina imports | 610,081 | t | 2026-04 | a10167790 |
| SMM scenario | 2026 LME range | Conditions |
|---|---|---|
| Bull (~30%) | US$3,700–4,000/t | Sustained peak-season destocking trend confirmed, faster than the seasonal norm; overseas (LME) tightness persists or worsens |
| Base (~50%) | US$3,000–3,400/t | Two-speed market persists; the seasonal destock comes through at a slope broadly in line with, or slightly faster than, past peak seasons, with social inventory easing back toward the year-earlier level by year-end; alumina holding ¥2,500–3,000/t |
| Bear (~20%) | US$2,600–2,900/t | Peak season fails to destock — stocks build instead, with warrants rebuilding alongside; overseas supply heals faster than expected on geopolitical de-escalation, with the import window staying open; the solar downshift outweighs the storage-and-grid offset |
Volume series are converted from SMM's native 10,000-tonne units to million tonnes (Mt) or thousand tonnes (kt) for readability. Daily price, cost and inventory series were downsampled to one observation per month plus the latest print for file size; the underlying SMM series are full-resolution. The capacity-migration, carbon-intensity and price-band charts are illustrative figures anchored to SMM/IAI compilations and company guidance; year-by-year interpolations are indicative.
| A00 | Market shorthand for Al99.7 primary aluminium (the Chinese national-standard grade), aluminium content ≥99.70%; SMM's A00 spot is the domestic benchmark (¥/t). |
| Alumina | Aluminium oxide (Al₂O₃), the intermediate refined from bauxite; ~2 t per t of metal; 30–40% of smelter cost. |
| Bauxite | The ore; Guinea is the swing supplier to China. |
| Prebaked anode | Carbon anode consumed in electrolysis; a key cost input. |
| Primary vs secondary | Primary = smelted from alumina (energy-intensive); secondary = recycled from scrap (~5% of the energy). |
| ADC12 | The benchmark secondary (recycled) aluminium alloy, used in die-casting. |
| SSSR / capacity cap | Supply-side structural reform; the 2017 policy ceiling on Chinese primary aluminium capacity. |
| LME / SHFE | London Metal Exchange (US$/t, global) and Shanghai Futures Exchange (¥/t, domestic). |
| CBAM | EU Carbon Border Adjustment Mechanism; aluminium is a covered sector; charging from Jan 2026. |
| Semis | Semi-fabricated products (extrusion, sheet, foil, wire); the bulk of aluminium's downstream trade. |
| Social inventory | Metal held in monitored warehouses outside producers; the key Chinese spot-balance signal. |
| Green premium | The emerging price differential for certified low-carbon (hydro/solar/recycled) aluminium. |