Global aluminum deficit widens as Middle East smelter damage opens a window for US production - if Oklahoma's power contract closes

The geopolitical case for rebuilding US primary aluminum production has sharpened considerably since spring. Strikes on Emirates Global Aluminium's Al Taweelah smelter in Abu Dhabi and Aluminium Bahrain's force majeure declaration have removed an estimated 3-3.5 million tonnes of output from the global market in 2026, according to Wood Mackenzie, pushing the sector into structural deficit at the same moment that domestic tariffs are triggering the first US smelter restarts in years^[1].

The timing matters because US grid buildout is simultaneously creating new domestic demand for the metal. Transformer and cable manufacturers racing to expand grid infrastructure are among the fastest-growing buyers of aluminum, adding an industrial pull to the tariff-driven supply push^[1].

What the Middle East disruption means for supply

The Strait of Hormuz conflict has effectively halted exports from a region that supplies roughly 7% of the world's aluminum, while military strikes have impaired about 3% of global production capacity. Wood Mackenzie estimates some affected smelters may need up to a year or more to resume full operations.

The supply arithmetic is stark:

• Global primary aluminum output reached roughly 73.8 million tonnes in 2025, per International Aluminium Institute data
• GCC smelters contributed around 6.16 million tonnes, or roughly 8% of world supply, with more than 80% destined for export
• Wood Mackenzie's April 2026 estimate put the 2026 market deficit at 4 Mt following the EGA and Alba strikes
• Wood Mackenzie projected aluminum prices rising to around $3,500 per tonne in 2026, with outcomes sensitive to conflict duration

That price environment is what makes the current US restart wave credible. The two smelter restarts already announced - Century Aluminum's Mt. Holly facility in South Carolina and Magnitude 7 Metals' partial restart in Marston, Missouri - together add more than 20% to US primary output^[1]. But they represent restarts of existing, partially idled capacity. The structural question is whether the US can build something genuinely new.

The Oklahoma greenfield and its power dependency

On 26 January 2026, Emirates Global Aluminium and Century Aluminum announced a joint development agreement to build the first new primary aluminum smelter in the United States since 1980, targeting 750,000 tonnes per year at a site in Inola, Oklahoma - more than doubling current US production. EGA holds 60% of the joint venture; Century holds 40%. Bechtel was named engineering lead in February, with a final investment decision and construction start targeted by end-2026.

The project has a single gating constraint that no tariff can resolve: a competitive long-term power supply agreement with Public Service Company of Oklahoma. Century's own SEC filings list completion of that negotiation as a condition precedent to the construction start.

The economics explain why. A single new aluminum smelter consumes roughly 11 TWh of electricity per year - comparable to the annual power use of a city the size of Boston or Nashville, according to the Aluminum Association. To be commercially viable, a smelter requires a power contract of at least 10-20 years at or below $40/MWh. Technology companies are currently committing upward of $115/MWh for power at AI data centers, according to the same source - a price that is structurally incompatible with aluminum smelting economics.

Missouri's large-load tariff adds a second layer of risk

The Magnitude 7 Metals restart in Marston faces a distinct but related obstacle. Missouri adopted a large-load electricity tariff in 2025 requiring major power users to help cover grid infrastructure upgrade costs. The policy was designed with data centers in mind - operators that can absorb the added expense. For aluminum producers operating on thin margins, the same tariff functions as a structural cost penalty^[1].

The Marston facility has a further complication: the smelter's power source has not been publicly disclosed ahead of the restart. New Madrid County recorded the worst air quality in the United States before the smelter idled, according to 2019 EPA data, and community groups have called on Magnitude 7 Metals to disclose whether the restart will rely on the adjacent New Madrid coal plant. Missouri's Department of Natural Resources has noted that sulfur dioxide compliance is assessed on a three-year average, meaning emissions from a restart would be measured against data from the idle period.

What to watch

The Oklahoma power contract negotiation with PSO is the near-term signal to track. A deal would confirm that at least one US utility is willing to price industrial power competitively against data center demand - a precedent with implications well beyond a single smelter. Absent that agreement, the construction start target of end-2026 slips, and the window opened by Middle East supply disruptions narrows before new US capacity can reach it.