Three-month LME aluminum traded near $3,170 per metric ton on 24 July 2026, about 16% below the four-year high of $3,787.50/t reached in early June 2026, while regional premiums and tariff layers kept the all-in landed cost in the United States and Japan at or above spring peaks [S1].
The decoupling is structural: the LME screens the probability of supply disruption, while regional premiums price the cost of actually landing a tonne in the right alloy, shape, origin and delivery window, and those costs have not fallen at the same speed as the futures curve [S1][S3].
Anatomy of the all-in landed aluminum price
Industrial buyers do not pay the LME quote; they pay an LME-anchored "all-in" price that layers a regional physical premium, conversion costs, freight, tariffs, and a small but rising set of carbon and traceability charges on top of the screen price [S1][S2]. A typical US buyer's cost stack reads: LME cash price, Midwest Premium in cents per pound, conversion (billet to extrusion or rolling), freight to plant, any applicable Section 232 or antidumping duty, plus finishing and inspection [S2][S3].
For reference, the LME itself offers Aluminium Premiums contracts explicitly because the premium portion of the all-in price is large enough to need its own hedging instrument, separate from the LME Aluminium future [S4]. The Midwest Premium is quoted in cents per pound and covers import duty, freight, financing, warehousing, delivery timing, and local demand pressure, not just the duty rate [S3].
What regional premiums actually priced in 2026
The Q3 2026 primary aluminum premium delivered to main Japanese ports settled at $395 per metric ton above the LME cash price on 6 July 2026, up 11.9%–12.9% from the Q2 range of $350–$353/t, and the highest Q3 level in more than a decade, even though initial offers had started at $440–$480/t [S1].
On COMEX, the CME Aluminum futures benchmark (symbol ALI) was last quoted at 3,106.75 on 21 October 2026, up 22.75 (+0.74%) on the session, illustrating that the screen is no longer where the directional risk sits [S5]. The CME companion contract, the physically delivered Aluminum futures, has registered warehouses in the United States, Europe, and Asia, so regional premium differentials still converge through physical flow, not just paper [S5].
Why premiums moved while LME fell
Middle East supply shocks were the dominant 2026 driver: Emirates Global Aluminium's Al Taweelah smelter was restarting, Indonesian primary metal entered the export market, and Chinese semi-finished exports rose, taking the geopolitical "war premium" out of the LME curve by late July 2026 [S1][S5]. At the same time, the CME noted in May 2026 that rising regional premiums had put aluminum risk management back in focus precisely because the all-in price stayed elevated even as the LME eased [S5].
Tariffs are the second amplifier. The Midwest Premium does not move one-for-one with the duty rate; it moves with the difficulty of replacing a tonne of metal when and where it is needed, and tariffs change that difficulty [S3]. Reuters reported on 5 June 2025 that US aluminum premiums had hit record levels after new tariff treatment took effect, setting a higher base from which 2026 premiums were negotiating [S6].
How tariff expectations move the premium ahead of policy
Harbor Aluminum's September 2026 read is blunt: the Midwest Premium prices the probability of a tariff change, not just the tariff itself; when buyers and sellers believe the rules may be revised, bids, offers, and inventory decisions move before a single tonne changes route [S3].
Under a "possible tightening ahead" scenario, buyers accelerate to cover under current terms, holders of tariff-paid metal withhold inventory, and the premium firms before any new duty is collected; under a "possible relief" scenario, importers destock and consumers delay, and the premium softens [S3]. A wide, unresolved range of outcomes turns the premium into a risk price, with wider bid-offer spreads and headline-driven intraday swings even when no rule has been signed [S3].
Cost stack: ranking the landed-cost drivers
In rough order of 2026 magnitude for a US buyer, the cost stack is: (1) LME base, (2) US import duty and any Section 232 layer, (3) regional Midwest Premium in cents/lb, (4) conversion (billet premium, extrusion or rolling, heat treatment), (5) freight from smelter or port to plant, (6) machining and finishing, (7) inspection, packaging, and traceability paperwork, (8) carbon and CBAM-related charges on the EU-bound share of shipments [S1][S2][S3].
For European shipments, the CBAM aluminium cost impact has become its own line item starting 2026, and buyers exposed to both US Section 232 and EU CBAM are effectively paying two parallel regulatory premiums on the same physical tonne; see the 2026 CBAM operating numbers for the quantified hit [CBAM article]. Within that stack, the regional premium plus duty is the single most volatile line, while conversion and freight move more slowly and tend to drift up with energy and labor [S2][S3].
Who this matters for, and who it does not
This premium-versus-LME split matters most to: OEM buyers who quote in USD per pound on long-term contracts, extruders and service centers hedging the premium separately from the LME, and procurement teams writing should-cost models for fabrication shops and aluminum alloy components [S1][S2][S4]. It matters less to: long-position LME traders, financial hedgers using only Aluminum futures on COMEX (symbol ALI), and toll processors who pass the metal cost through to the customer [S4][S5].
For buyers weighing metal substitution in power hardware, the premium layer is also why a switch toward aluminum alloy conductors only works when the design tolerates the lower conductivity per cross-section; the metal-cost saving is real but smaller than the LME-versus-copper spread suggests once Midwest Premium and Section 232 are added.
Total cost of ownership beyond the invoice
Looking past the purchase price, the all-in landed cost includes: carrying cost on tariff-paid inventory (financing, insurance, warehousing), working-capital drag from longer mill lead times when premiums spike, quality-cost risk on off-spec or wrong-origin material that has to be resold at a discount, and 2026-specific carbon-adjustment cost on EU-bound shipments under CBAM [S1][S3].
On a five-year view, the dominant signal is that the premium line item has become a risk price, not a cost price, and procurement contracts that lock only the LME component leave the largest dollar line item exposed. The next trackable signal is the Q4 2026 Japan premium settlement (expected late September 2026): a print below the $395/t Q3 level would confirm the LME fall is finally being passed through; a flat or higher print would confirm the decoupling is structural, not seasonal.