Aluminium is living in two calendars at once.
This year’s calendar is tight. Gulf smelters were hit. Visible stocks are thin. London Metal Exchange cash metal was near $3,240 a tonne in late September 2026. That is still far above the $2,600-class averages of 2025.
Next year’s calendar is the one the Street is starting to price. Bank of America Global Research spent 2025 and 2026 mapping that two-speed market. Metals research head Michael Widmer first called a 2026 deficit and prices above $3,000 a tonne. In late April 2026 he pulled a $4,000 target forward from the second quarter of 2027 into the fourth quarter of 2026. The reason was damage to Middle East smelters, not a sudden boom in cans.
He also flagged the swing factor that now sits at the centre of every surplus debate. Indonesia’s new smelters were ramping. Supply growth, BofA said then, could accelerate to about 2.6 percent in 2026. High-frequency output data would matter more than slogans.
That is the investor theme. The 2026 deficit is a war and power story. A 2027 surplus, if it arrives, would be a supply-timing story. Demand does not have to collapse for prices to cool. New metal only has to show up faster than the market can absorb it.
This article is for information only. It is not investment advice. It is not an offer to buy or sell any security or commodity.
What Bank of America actually said
Accuracy matters here. BofA did not publish a neat public table that says “590,000 tonnes of surplus in 2027.” Other desks did. Investors should keep those voices separate.
What BofA did say is still useful.
In December 2025, Widmer told clients the bank expected an aluminium deficit in 2026 and prices pushing above $3,000 a tonne. He tied faster supply growth to Indonesian smelters coming online. He warned that the number of units Indonesia would actually deliver was uncertain. High-frequency data would be the tell.
He also described the structural map. The United States had lost about three-quarters of its smelters, falling from more than 20 plants in 1998 to about five. China had capped primary capacity near 45 million tonnes a year. The biggest new production was coming from Indonesia. In many cases Chinese operators were building those pots. Power cost, not patriotism, would decide where new metal is poured. Data centres, he noted, can pay several times what a smelter wants to pay for electricity.
By May 1, 2026, the war shock had changed the price path. BofA brought the $4,000-a-tonne target forward to the fourth quarter of 2026. Widmer said the bank also raised longer-term price expectations. The demand shock from the conflict, he added, was not yet large enough to rewrite the baseline balance. The supply shock was.
Read that sequence as an investor. BofA first called tightness. Then it called even tighter tightness. Then it kept pointing at Indonesia as the later supply wave. The Street has now put tonnes on that second wave. Goldman Sachs, CRU Group, Shanghai Metals Market and others now sketch a 2027 surplus if Gulf metal returns and Asian pots keep ramping.
The change investors must watch is not a slogan. It is a stack of tonnes.
Why aluminium could move into surplus in 2027
Three supply pipes can refill the tank at the same time.
The first pipe is the Gulf. The Middle East accounts for about 9 percent of world primary aluminium. It punches above that weight in seaborne trade. Strikes on Emirates Global Aluminium’s Al Taweelah complex and Aluminium Bahrain’s Alba plant cut Western-facing supply. Analysts have put lost or delayed Gulf output in a range from about 2 million to 3 million tonnes for 2026, depending on the house and the month of the note.
Goldman Sachs later said Bahrain output may only get back to pre-conflict levels by mid-2027. The United Arab Emirates may take until the end of 2027. That lag is why 2026 stayed tight. It is also why 2027 can loosen if those pots actually restart.
The second pipe is Indonesia. Goldman raised its Indonesian primary forecast to 1.7 million tonnes in 2026 and 2.9 million tonnes in 2027. Earlier it had used 1.6 million and 2.5 million. It cited faster ramps at Adaro, Taijing Morowali and Juwan Weda Bay. Indonesian output was already up about 89 percent year over year in the period covered by that June 2026 note.
The third pipe is China. Beijing’s official cap is about 45 million tonnes of primary capacity. Strong margins have still pulled extra metal. Goldman lifted China production to 45.6 million tonnes in 2026 and 46.3 million tonnes in 2027. That is not a legal rewrite of the cap. It is a forecast that high prices will keep plants running hard and, in places, above the spirit of the ceiling.
Add India and other ex-China projects and the 2027 maths starts to look different from 2026. Press Metal coverage from Hong Leong Investment Bank sketched 2 million to 3 million tonnes of new supply from Indonesia and India, plus a gradual Gulf return. Shanghai Metals Market put a 2027 surplus in a band from about 690,000 tonnes to 1.42 million tonnes, depending on how fast the Middle East recovers.
Goldman itself cut its 2027 surplus view from 1.3 million tonnes to 590,000 tonnes when it decided Gulf repairs would be slow. That cut is the whole debate in one number. Slow restarts keep 2027 almost balanced. Fast restarts can dump more than a million extra tonnes on the market.
That is what could drive a surplus. Not a collapse in cars or cans. A pile of new metal arriving after a year of missing metal.
What is driving the aluminium supply surplus case
High prices do the recruiting.
When LME aluminium ran toward four-year highs in the spring of 2026, smelter margins widened. Developers brought projects forward. SMM said it raised its estimate of newly added primary capacity outside China from 2026 onward by more than 60 percent between January and May. For 2026 alone it lifted planned commissioned operating capacity outside China to 2.32 million tonnes from 1.38 million. For 2027 it lifted the figure to 2.03 million from 1.63 million.
That is the surplus machine. Price pulls supply. Supply then leans on price.
Power is the brake. Aluminium is frozen electricity. A modern potline needs cheap, steady megawatts. Indonesia’s coal and captive power plants are the reason new metal is landing there. Quebec hydropower is the reason Canada still matters. U.S. smelters lost that fight. Data centres bid more for the same grid.
Bauxite and alumina sit one step upstream. Australia’s Resources and Energy Quarterly said the alumina market itself may stay in surplus into 2027. Middle East smelters need less feed while they are down. Indonesia and India are adding alumina as well as metal. Extra alumina makes extra aluminium easier to pour once pots are ready. It does not pour the metal by itself.
Recycling is the quiet fourth pipe. Secondary metal already covers a large share of Western use. Scrap collection rises when primary prices stay high. That extra flow does not show up in a primary-only balance sheet. It still hits the price.
Put those drivers in one sentence. A 2027 surplus case rests on repaired Gulf pots, Indonesian ramps, Chinese over-run, Indian additions and more scrap. Any one of those can slip. All of them together can overwhelm a market that is still short this year.
Could aluminium prices fall in 2027?
They could. That is the point of a surplus forecast. It is not a sure thing.
Look at the published bands, not at a single print.
Goldman Sachs, after raising its near-term view, still put average 2027 LME aluminium at $2,950 a tonne. That was up from $2,750. It was still below the forward curve at the time of the note. The bank rolled a short into December 2027. It said that contract best expressed a structural surplus view. If Gulf restarts ran fast, Goldman said the surplus could reach about 1.2 million tonnes and prices could sink toward $2,750. If restarts ran slow, 2027 could stay near balance around $3,250.
J.P. Morgan’s August 2026 metals outlook was more generous near term and still softer later. It saw $3,800 in the third quarter of 2026 and $3,700 in the fourth. It then stepped the path down through 2027 toward $2,750 by the fourth quarter. The full-year 2027 average in that table was about $3,075.
CRU Group, as summarized in late-August market notes, expected prices to average about $2,900 in the fourth quarter of 2026, down roughly 16 percent from the second quarter, as the market looked through the deficit toward surplus years.
The World Bank’s spring outlook had $3,200 for 2026 and $3,000 for 2027. Spot prices spent months above that 2026 figure. The bank’s 2027 cooling case still rhymes with the Street.
Australia’s Department of Industry, in its June 2026 quarterly, forecast LME averages of $3,390 in 2026 and $3,310 in 2027, then a real-term slide toward $2,900 later in the decade.
BofA’s own published target of $4,000 was a 2026 event call after the Gulf shock. It was not a 2027 average. Investors who treat a wartime spike target as a floor for the next two years are mixing two different questions.
Cash aluminium near $3,240 in late September 2026 already sits below the spring peaks. The market is no longer only pricing the outage. It is starting to price the refill.
Prices can still stay high if demand surprises or if Gulf repairs slip again. They can fall if the three supply pipes open together. A surplus forecast is a warning about that second path. It is not a promise.
The 2026 deficit is still the base
None of the 2027 surplus talk erases this year’s hole.
Estimates differ because houses count lost Gulf tonnes, Chinese exports and inventory in different ways. The range is wide. ING has spoken of a 1.8 million tonne deficit, later trimmed toward 1.2 million as some Emirates Global Aluminium capacity returned and Chinese product exports jumped. Goldman’s June revision was 720,000 tonnes. CRU has been cited near 1.4 million. Mercuria’s Nick Snowdon called the shock the largest single base-metals supply hit since 2000 and sketched a deficit that could reach 2 million tonnes or more. Wood Mackenzie, in the first weeks after the strikes, floated a deficit as large as 3 million tonnes.
Those numbers cannot all be right. They can all point the same way. 2026 is short.
Visible buffers are thin. LME opening stocks were about 242,125 tonnes on September 23, 2026. Live warrants were about 238,475 tonnes. That is a few days of world primary output, not a warehouse mountain. Earlier in the year stocks had been even lower. A market that lean does not need a huge extra deficit to stay nervous. It only needs one more outage.
Physical premiums tell the regional story. The U.S. Midwest premium blew out when tariffs and missing Gulf tonnes hit at once. Europe felt the same squeeze through energy costs and lost seaborne units. China could export more product. It could not instantly replace every missing Western billet.
Alcoa chief executive William Oplinger said in May 2026 that the company had expected a slight deficit going into the year. Then about 2.5 million tonnes of Middle East capacity went offline. He had not yet seen physical scarcity in Europe or North America. He thought it could appear within six months. That is how a paper deficit becomes a plant problem.
Aluminium production: where the new tonnes live
World primary output sits near 73 million to 74 million tonnes a year in recent tallies. China is about 60 percent of that. The rest of the world is the swing set.
Canada produces about 3.3 million tonnes. That is roughly 5 percent of global supply and enough to rank fourth. About 96 percent of Canadian metal is poured with hydropower. The country runs near 95 percent of capacity. It exports about 90 percent of what it makes. Most of that metal normally goes to the United States.
Rio Tinto, Alcoa and Aluminerie Alouette dominate the Canadian map. Rio Tinto’s AP60 expansion at Arvida in Quebec is adding about 160,000 tonnes of lower-carbon capacity as older pots close. That is not a world-changing number. It is a reminder that Canada grows by upgrading pots, not by opening empty deserts.
The United States poured only about 660,000 tonnes of primary metal in the latest full year cited by industry notes. It needs to import the rest. Canada can cover a large share of that gap. Tariffs of 50 percent on raw metal have distorted the route, not the geology. Alcoa chief financial officer Molly Beerman said in September 2026 that even a tariff cut on Canadian metal would not crush the Midwest premium by itself. The United States still needs about 4 million tonnes of imported supply. Canada can offer about 3 million. The last million still has to come from somewhere.
Indonesia is the growth story. Capacity there is jumping from well under 1 million tonnes toward several million by the end of the decade. Australian official forecasts have Indonesia passing Australia in primary capacity if projects run on time. Power and permits remain the risks. They always are in this industry.
India is the other riser. New pots there add to the 2027 wave. They also compete for alumina and carbon anodes. Lead times on cathode blocks and labour can slip a “2027 surplus” into 2028. That is why Widmer’s old warning still holds. Watch the units that actually start, not the units on a slide.
Bauxite supply and the alumina bridge
Primary aluminium starts as bauxite. Bauxite becomes alumina. Alumina becomes metal.
Guinea, Australia and Brazil still set the bauxite tone. Indonesia is building more of the middle step. When Gulf smelters go dark, they stop pulling alumina. Refineries then look long. That is why official Australian work sees alumina surplus into 2027 even while primary metal is short in 2026.
A surplus in alumina is not a surplus in aluminium. It is cheap feed waiting for pots. If those pots light in Indonesia and the Gulf in the same year, alumina surplus and metal surplus can arrive together. If pots stay dark, alumina stays cheap and metal stays tight. Investors who only watch LME aluminium miss that split.
China’s alumina system is huge. Policy can tighten approvals. It has not removed the country’s ability to feed its own pots. Exportable alumina and exportable aluminium products are different valves. In 2026, Chinese product exports rose sharply in some months. That metal showed up in Western warehouses as fabricated goods, not as LME ingot. Balance sheets that ignore those flows overstate the Western hole.
Aluminium consumption has not vanished
A surplus call can sound like a demand call. It is not.
Global aluminium use still tracks power lines, vehicles, packaging, buildings and aircraft. Bank of America’s own packaging work in August 2026 showed cans still taking share from glass and plastic in several U.S. beverage categories. Can volumes were up 2.6 percent year over year in the latest four-week Nielsen slice the bank cited. Carbonated soft drinks in cans rose 3.6 percent. Plastic bottles in that category fell 4.1 percent.
That is not a boom. It is a floor. Packaging does not disappear because a smelter in Bahrain is being rebuilt.
Electrification is the larger pull. Transmission cable, solar frames, wind towers and electric-vehicle castings all use aluminium. Data centres add busbar and building metal. None of those end markets have to grow at wartime rates for 2026 to stay tight. They only have to keep growing while Gulf tonnes are missing.
Construction is the soft spot. China property is no longer the sponge it was a decade ago. That is why China can export more product even when its own cap is binding. Weak housing frees metal. Strong grids and cars soak it back up. The net is slower demand growth, not a collapse.
Aerospace felt the squeeze early. Bank of America cut its Airbus price target in March 2026 on aluminium cost risk after the Gulf strikes. Extra costs in the hundreds of millions of euros were the issue, not a lack of orders. That is demand meeting a supply shock, not demand dying.
Canadian mining stocks and Canadian aluminium stocks
Canada does not mine much bauxite. It smelts. The investment map is therefore a power-and-tariff map.
Rio Tinto and Alcoa are the listed names most investors use to touch Canadian pots. Hydro-Québec power is the hidden asset. Low-carbon branding is the marketing layer. The 50 percent U.S. tariff is the tax.
Jean Simard of the Aluminium Association of Canada said in mid-2026 that high prices were “very nice” and “not sustainable.” That is the producer’s version of the surplus thesis. Canadian plants ran hard while Gulf metal was missing. They still had to reroute cargoes when Washington taxed the old path. Some metal went to Europe and Asia. Not enough to replace the U.S. door.
A 2027 surplus would not close Quebec. It would squeeze margins from the top. Hydro-based metal should still sit on the left of the cost curve. Coal-based new Indonesian metal sits on a different curve. Carbon rules in Europe can keep that gap alive even if the LME print falls.
Junior mining stocks in Canada are a poor direct proxy for primary aluminium. Most Canadian juniors chase gold, copper, nickel or critical minerals. Aluminium exposure for a resource investor is usually a producer, a recycler or a downstream fabricator. Treating every TSX miner as an aluminium bet is a category error.
What Canadian smelters offer in a surplus year is relative defense, not a guarantee. Cheap power. High utilization. A customer base that still needs metal next door. Tariffs can erase that edge overnight. So can a flood of Asian units into Europe.
Aluminium mining companies and aluminium producers
The listed producer set is short. Alcoa. Rio Tinto. Norsk Hydro. Century Aluminum. Emirates Global Aluminium is strategic but not a simple Western ticker. Press Metal is the Southeast Asian growth name. China Hongqiao and other Chinese groups move the global total more than any Western board.
Alcoa’s May 2026 message was blunt. The market was not balancing. Nameplate capacity still looked long on paper. Real available metal did not. The company had restarted pots in Spain and elsewhere under viability deals. It still saw a constructive five-year case because China no longer offered an infinite buffer.
That last point is the hinge between deficit and surplus. If China holds near 45 million to 46 million tonnes, the rest of the world must fill growth. Indonesia can do some of that. If Indonesia and the Gulf both deliver in 2027, the hinge swings to surplus. If either slips, the hinge stays tight.
Century Aluminum’s planned Oklahoma project with Emirates Global Aluminium shows the lag. Even with federal support, hot metal is years away. A Pentagon war-game about 2027 aluminium resilience made the same point. New U.S. pots cannot cover a 2027 hole. Imports still rule.
Aluminium market analysis: how to read the next prints
Investors do not need a new religion. They need a checklist.
Watch LME stocks and cancelled warrants. Rising live warrants with falling cancelled tonnes often means metal is staying put. The opposite often means someone wants physical units.
Watch the Midwest premium and the European duty-paid premium. Those spreads show whether the shortage is global or local. A falling LME price with a sticky U.S. premium means tariffs still rule the American plant gate.
Watch Indonesian monthly output. Widmer was right about that in 2025. A country jumping from less than 1 million tonnes toward 3 million can change a world balance by itself.
Watch Gulf restart language. “Cells back on” is not “pre-war run rate.” Goldman’s mid-2027 and end-2027 dates are the base case, not a contract.
Watch Chinese product exports and the official capacity inspections. Extra exports ease the Western deficit without adding LME tonnes. Tighter inspections cut that relief.
Watch power prices in Europe and captive-power news in Indonesia. A dark potline in Germany and a new potline in Morowali can land in the same week. The net is what matters.
Aluminium forecast 2026 versus aluminium forecast 2027
The 2026 forecast across houses is some version of tight. Price averages in official and bank work cluster from the low $3,000s to the mid $3,000s, with spike risk if another smelter goes down.
The 2027 forecast is a fork.
Fork A is a modest surplus and prices drifting toward $2,750 to $3,000. That is Goldman’s faster-restart risk case and the lower end of SMM and J.P. Morgan paths.
Fork B is a small surplus or a thin balance near $3,200 to $3,250. That is Goldman’s slow-restart case and closer to Australia’s official 2027 average.
Fork C is no surplus at all. That happens if Gulf repairs slip into 2028, Indonesian power fails, or Chinese inspections bite just as Western demand holds. Few base cases use Fork C today. It is the tail that keeps $4,000 talk alive.
BofA’s published work sits nearer the tight side of that map than the glut side. It pulled a high price target into 2026. It raised longer-term expectations after the war shock. It still treated Indonesian supply growth as the variable that could change the later years. Other desks turned that variable into a surplus line. Investors can use both without pretending they are the same note.
Aluminium investment and aluminium price outlook
An investment process for this metal is a calendar process.
In a deficit year, the scarce thing is available metal outside China. Canadian hydro metal, Icelandic metal and remaining Gulf units command a premium. Equities with those tonnes can earn more than the LME move. They can also lose that extra if tariffs or power deals change.
In a surplus year, the scarce thing becomes cheap power and low-carbon certificates. High-cost pots get squeezed first. New Indonesian units may still run if their power is captive and their debt is already sunk. Old European pots may not.
Physical buyers should not treat a 2027 surplus forecast as a reason to run inventories to zero in October 2026. Lead times are long. Regional premiums can stay high after the LME cracks. Aerospace and auto plants do not buy average annual balances. They buy the billet that arrives next month.
Speculators who only trade the LME print should remember the curve. A market that is tight now and potentially long later often sits in backwardation first and then flattens. Goldman’s decision to express the surplus view in a later contract was that idea in trade form. It was a view. It was not a law.
Aluminium mining stocks are not the same as aluminium stocks
Bauxite miners, alumina refiners and aluminium smelters do not move as one.
A 2026 metal deficit can coexist with an alumina surplus. That pairing hurts refiners and helps smelters. A 2027 metal surplus that arrives with still-ample alumina hurts both. Scrap-heavy recyclers can do better than either if primary prices fall and collection stays strong.
Canadian resource investors often hold diversified miners. Rio Tinto’s iron ore and copper can swamp its aluminium line in a given quarter. Using the whole ticker as a pure aluminium price forecast is sloppy. Using the aluminium division’s reported shipments and realized prices is not.
Junior explorers almost never offer clean aluminium torque. If the goal is aluminium market exposure, the honest list is short and mostly large. That is a feature of the industry. New mines and new pots take years and gigawatts.
Global aluminium demand still has a floor
World primary demand in recent consultant tables sits in the mid-70 millions of tonnes. Growth rates near 2 percent a year are common in base cases. That sounds dull. On a 74 million tonne base, 2 percent is about 1.5 million extra tonnes. That is the same order of magnitude as a Gulf restart or an Indonesian ramp.
This is why the surplus debate is so sensitive. A 590,000 tonne surplus is less than 1 percent of the market. A 1.2 million tonne surplus is still small next to a pandemic shock. Either figure can move the price a long way when stocks are only a few hundred thousand tonnes.
Demand risks sit in China property, European industry and a broader slump. Demand supports sit in grids, vehicles, packaging and defense. BofA’s can data and Alcoa’s “everything you touch” line describe the support. They do not cancel the supply wave.
What investors should watch from here
The one theme is timing.
If Indonesian pots and Gulf cells arrive together in 2027, the books can flip from deficit to surplus even if consumption stays healthy. That is the change BofA’s earlier work pointed toward and that Goldman, CRU and SMM have now tried to size. Prices could then fall from 2026 war levels without any crash in the real economy.
If those tonnes slip, the 2026 shortage leaks into 2027. Then BofA’s tighter, higher-for-longer instinct looks better than the surplus slide. $3,000 would be a floor talk again, not a ceiling talk.
Neither path is a buy ticket. Aluminium producers can earn good money in both regimes if their power is cheap. They can lose money in a surplus if they sit on the wrong grid. Canadian hydro metal is better placed than most. It is still taxed at the U.S. border. It is still a cyclical industrial product.
Do not confuse a bank target with a delivery. Do not confuse a 2026 outage with a permanent shortage. Do not confuse a 2027 surplus sketch with a glut you can already touch. Count the pots. Count the stocks. Count the premiums. Then decide whether the refill is real.
People also ask
Why aluminium could move into surplus in 2027
Because three supply sources can rise at once. Damaged Gulf smelters may restart. Indonesian potlines are ramping fast. China is already running near or above its 45 million tonne cap. India is adding metal too. Those extra tonnes can exceed ordinary demand growth if they land in the same year.
What is driving the aluminium supply surplus
High 2026 prices pulled projects forward. Captive power in Indonesia made new capacity viable. China’s strong margins kept plants running hard. Alumina is expected to stay ample, which makes new pots easier to feed. Scrap adds a second stream. The surplus case is a supply response to a shortage, not a demand collapse.
Could aluminium prices fall in 2027
Yes. Several banks already sketch lower 2027 averages than 2026. Goldman’s published band runs from about $2,750 in a fast-restart glut to about $3,250 if Gulf repairs stay slow. A fall from today’s $3,200-class prints is possible. It is not guaranteed. Restarts can slip. Demand can surprise. Tariffs can keep regional prices high even if the LME eases.
Disclaimer
This article is for general information and education. It is not investment advice, tax advice or legal advice. It is not an offer or solicitation to buy or sell any commodity, security or derivative. Bank price targets and supply-demand balances are estimates. They can be wrong. They are not guarantees. Company names appear only to explain how listed producers and regions fit the physical market. Their mention is not an endorsement and not a recommendation. Aluminium prices, premiums, tariffs and mine or smelter output can change quickly. Always check primary filings, exchange data and original research notes before making any decision. Past price moves do not predict future results. You can lose money.