What the Numbers Reveal About America’s Aluminium Supply Crisis

Washington’s political class talks about rebuilding American heavy industry as though the foundations are still standing. The physical evidence tells a colder story: the country’s ability to make its own aluminium has been hollowed out over two decades, and the numbers behind the US aluminium supply crisis reveal a structural shortage that no single policy can quickly close.

Aluminium prices in the US no longer track a single global number. They track a domestic surcharge that spiked to a record in early 2026 the moment a trade lever moved, exposing just how thin the country’s physical supply cushion has become.

Understanding that gap matters for anyone with exposure to US manufacturing, construction, or the commodity itself. Here is what the data actually reveals about how deep the deficit runs, why physical premiums have decoupled from paper prices, why Canada cannot plug the hole alone, and why the one mega-project meant to fix it remains stuck in litigation and grid limbo.

The physical reality of the domestic capacity deficit

The gap between rhetoric and metal is enormous. Politicians describe a domestic industrial renaissance; the smelters on the ground tell you the country is starting from almost nothing.

Two decades ago, roughly 30 aluminium smelters operated across the United States. Today just four remain active. No new US smelter has been built since 1980, which means the operational knowledge, skilled labour, and supporting infrastructure have eroded alongside the plants themselves.

The production numbers make the erosion concrete. US nameplate primary capacity stands at about 1.31 million tonnes, yet actual output in 2025 came in at only around 660,000 metric tonnes. Roughly half the theoretical capacity sits idle or permanently closed.

The USGS Mineral Commodity Summaries 2026 records US primary aluminium smelter capacity at 1.31 million tonnes against 2025 output of approximately 660,000 tonnes, confirming that actual production runs at roughly half the theoretical nameplate figure.

Here is the contrast in plain figures:

• Historical smelter count: approximately 30 active plants two decades ago
• Current active smelters: 4
• Nameplate primary capacity: ~1.31 million tonnes
• Actual 2025 production: ~660,000 metric tonnes
• Import reliance: ~60% of apparent US consumption

That last figure is the one to sit with. When roughly 60% of the aluminium the country consumes has to arrive from abroad, domestic supply is not a buffer against global shocks; it is a minority contributor.

North American aluminium demand across key end-use segments, including automotive sheet, building products, and electrical conductor wire, has continued to grow even as domestic primary supply has contracted, widening the structural gap that elevated premiums are now pricing in.

The market has felt that thinness in real time. LME aluminium warehouse stocks fell 31.82% in the first half of 2026, dropping from 498,670 tonnes in January to 320,314 tonnes by June, a drawdown linked to Russian sanctions and Gulf-related disruption.

For industrial buyers and investors exposed to raw material availability, the read is uncomfortable but clear. This is not a cyclical dip you wait out. It is a structural deficit built over 40 years, and it means any plan for rapid domestic self-sufficiency is starting from a near standing stop.

How the Midwest premium prices in supply chain fragility

If you only watch the LME headline price, you are missing where American buyers actually feel the pain. The cost of landing physical metal in the US is a separate number, and in January 2026 it told a far more alarming story than the exchange price did.

That number is the US Midwest aluminium premium. It is the surcharge American buyers pay on top of the baseline LME price to have physical metal delivered into North America, covering the freight, handling, tariffs, and scarcity of getting real aluminium to a real factory.

The LME spot price reflects global paper trading, contracts changing hands electronically around the world. As of late September 2026, that spot price sat at roughly $3,250-3,254 per metric tonne. The Midwest premium sits on top of it, and it moves for entirely different reasons.

In January 2026, the premium hit a record 99 cents per pound, equivalent to about $2,183 per tonne. That is not a rounding adjustment. It is a physical delivery surcharge large enough to reprice the economics of American manufacturing.

The trigger was a policy lever. The US doubled its levy on Russian aluminium to 50%, effective from mid-2025, which pulled competitively priced Russian metal out of reach and forced US consumers to chase alternative supply at higher cost.

Reuters reported on 26 January 2026 that the record Midwest premium was directly linked to the US doubling its levy on Russian aluminium to 50%, a move that sharply reduced competitively priced Russian supply and pushed American consumers toward more expensive alternatives.

The mechanics here matter for how you model risk. The LME price can hold steady while the premium spikes, because the premium captures the friction of physical delivery that paper trading simply ignores.

Midwest premium mechanics have grown increasingly detached from LME spot movements as the concentration of US import sources has narrowed, creating a pricing structure where trade policy decisions can transmit directly into manufacturer cost bases within days of announcement.

What this tells you is that your exposure to US heavy industry cannot be tracked through LME spot prices alone. When a trade dispute, sanction, or carbon-border measure hits any concentrated import source, the pain shows up first in the physical premium, and that premium flows straight into the cost base of every US manufacturer that needs real metal.

The mathematical limits of the Canadian import safety net

The comforting assumption is that Canada has America’s back. It is a friendly neighbour, a hydropower-rich producer of low-carbon aluminium, and by far the largest single supplier to the US market. The problem is that the math does not close.

Canada’s role is genuinely central. Its share of US aluminium imports averaged roughly 56% over the 2021-2024 period, drawing on cheap hydroelectric power and roughly 3.3 million tonnes of annual primary capacity that continues to expand its low-carbon output.

But scale is where the safety net frays. According to Alcoa CFO Molly Beerman, the US needs to import approximately 4 million tonnes of aluminium a year, while Canada can realistically supply around 3 million tonnes.

That residual 1 million tonnes is the crux. Even if every available Canadian tonne flowed south, the US would still be roughly a quarter short and forced to source the balance from third-party suppliers around the world.

Beerman made the policy implication explicit: cutting tariffs on Canadian imports alone would not be enough to materially bring down the elevated Midwest premium, precisely because Canadian capacity cannot cover the full requirement.

Those third-party tonnes are exactly where the trade shocks land. The 50% Russian levy and the January 2026 premium spike were not freak events; they were a direct consequence of the US needing metal from sources that sit squarely in the crossfire of sanctions, tariffs, and geopolitical friction.

For capital allocators, this dismantles the idea that North American trade ties insulate the US from global volatility. Relying on Canada solves about three quarters of the problem, which means you must price ongoing exposure to volatile third-party suppliers into any long-term model for US physical metal.

The Oklahoma mega-smelter and the decade-long reality check

If Canada cannot close the gap, the obvious answer is to build. And on paper, the United States is building something enormous.

Century Aluminum confirmed on 1 December 2025 that it was joining Emirates Global Aluminium’s (EGA) proposed smelter at Inola, Oklahoma, at the Tulsa Port of Inola. The venture is structured as a 60/40 joint development agreement, with EGA holding 60% and Century holding 40%.

The ambition is genuinely large. The facility, designated Oklahoma Primary Aluminum, targets 750,000 tonnes of annual output on a projected investment of roughly $4 billion, which would more than double current US primary production and rank as the largest primary aluminium plant ever built in the country.

The Oklahoma Primary Aluminum project‘s $4 billion capital commitment and 750,000-tonne capacity target represent the most ambitious attempt to reverse US primary aluminium decline in four decades, yet the investment thesis rests entirely on a permitting and grid integration sequence that remains unresolved.

The economic case has been marketed hard. An impact study projects a cumulative $49 billion contribution to Oklahoma’s economy through 2060, an average of about 7,000 jobs a year, and peak employment above 10,600 positions in 2029.

The official timeline is equally confident: construction starting by the end of 2026 and first primary aluminium production by the end of the decade. That framing has been repeated across project materials and mainstream coverage.

Litigation and grid integration roadblocks

Then you look at the ground as of late September 2026, and the optimism runs into hard obstacles. No groundbreaking has happened. The project sits in planning and permitting, blocked by three distinct problems:

1. The local moratorium. The town of Inola extended its moratorium on the smelter through April 2027 and is planning a local referendum on whether the project should go ahead. Construction cannot physically begin before that moratorium lifts, which already makes the end-2026 start date unreachable.
2. The attorney general’s lawsuit. Reuters reported on 3 June 2026 that Oklahoma’s attorney general filed suit to block EGA and Century, citing concerns about the project’s ties to the UAE and potential harm to local health and agriculture. That adds active legal risk to the permitting path.
3. The unresolved power deal. A smelter this size is extraordinarily power-hungry, and per Canary Media the project is still waiting on a power deal. Securing a long-term, competitively priced electricity supply and integrating that load into the regional grid is a slow, complex undertaking that had not been completed as of the reporting period.

The gap between the promotional timeline and the litigation timeline is the whole story. Independent reporting characterises the project as “delayed again,” while official materials still cite the original end-2026 construction start.

For anyone weighting execution risk, the lesson is direct. Capital commitment alone cannot bypass grid constraints and local political opposition, so you should heavily discount headline projections of rapid domestic capacity growth. A 750,000-tonne smelter is a multi-year endeavour even before the courts and the town of Inola weigh in.

Navigating a prolonged structural deficit

The core friction is impossible to escape. The US needs metal now, but new primary capacity, if the Oklahoma project clears its legal and grid hurdles at all, is not realistically online before the end of the decade.

Set the two timelines side by side. New domestic production is a 2030 story at the earliest, while the physical market is already tight enough that LME warehouse stocks fell 31.82% in the first half of 2026 alone.

That mismatch means elevated physical premiums and heavy import reliance are not passing conditions. They are structural features of the US market that are likely to persist through at least 2030.

For industrial buyers and investors, adaptation is the only rational response. Model your exposure around a permanently tighter physical market, track the Midwest premium and trade policy as closely as the LME price, and treat any promise of a fast domestic fix as a long-term gamble rather than an imminent solution.

For readers wanting to understand the industry’s own proposed remedies in more depth, our full explainer on US aluminium supply chain strategy covers how recycling capacity, smelting policy, and trade framework adjustments are being sequenced as a multi-year response to the primary production shortfall.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results. Financial projections are subject to market conditions and various risk factors, and forward-looking statements about project timelines and capacity are speculative and subject to change based on market and company developments.