U.S. Alumina Refining and Primary Aluminum Production

1. Overview

North American Industry Classification System (NAICS) code 331313 covers U.S. establishments that refine bauxite ore into alumina and electrolytically reduce alumina into primary aluminum; establishments integrated with smelting may also cast, roll, draw, or extrude their own metal into primary forms.[1] Stand-alone recovery of aluminum from scrap and alloying of purchased metal belong to NAICS 331314, while downstream mills using purchased aluminum fall under other 3313 industries.

Public investors have limited but direct options: two listed companies—Alcoa and Century Aluminum—control the operating U.S. smelter base. Private investors can consider the single domestic alumina refinery or idled smelter assets, but opportunities are scarce and highly concentrated. Any investor should verify that an asset actually produces virgin aluminum rather than recycled metal or downstream fabricated products.

The central investment question is not simply whether aluminum demand rises. Returns depend on the spread between realized metal prices (London Metal Exchange benchmark plus regional and product premiums) and plant-level costs for power, alumina, carbon anodes, and labor—variables that can swing violently even when end-market demand is stable.

2. What it is and industry structure

Production has two chemically and economically distinct stages:

Alumina refining (Bayer process): Crushed bauxite is digested in hot caustic soda, insoluble "red mud" residue is separated, aluminum hydroxide is precipitated, and the hydroxide is calcined into alumina powder. The refinery is exposed to bauxite quality and freight, caustic soda prices, natural gas or steam costs, water availability, and residue-management liabilities.[2]

Primary smelting (Hall–Héroult process): Alumina is dissolved in molten cryolite and aluminum fluoride. Direct current passes between carbon anodes and the cathode lining of large electrolytic cells, or "pots"; molten aluminum is periodically tapped and cast into commodity sow or ingot, billet, foundry alloy, and other premium products. The process consumes both electricity and carbon anodes and must operate continuously: a prolonged power interruption can destabilize or freeze pots and require a costly, months-long restart.[3]

Major product categories include:

• Smelter-grade alumina for internal or merchant smelter use.
• Chemical-grade alumina for abrasives, ceramics, chemicals, and refractories.
• Primary aluminum ingot, sow, and T-bar.
• Billet and foundry alloys (higher-value forms cast at integrated smelters).

The classification is process-based. Important exclusions include secondary smelters recovering aluminum from scrap (NAICS 331314), aluminum sheet and plate mills (NAICS 331315), aluminum extruders (NAICS 331316), and aluminum foundries (NAICS 331524).[1]

Ownership is highly concentrated. Three companies control six U.S. smelters in five states; only two smelters operated at full capacity in 2025, two ran at reduced rates, and two were shut.[4] The single U.S. alumina refinery is privately held.

3. How big it is

The most reliable current measures are commodity statistics rather than Census revenue:

All figures are from USGS Mineral Commodity Summaries 2026.[4][5] The $2.6 billion value reflects published aluminum prices applied to domestic primary output and should not be treated as total NAICS 331313 revenue, which would also include alumina and integrated casting operations.

Census County Business Patterns counted 32 employer establishments under NAICS 331313 in 2023.[6] No unsuppressed Census revenue, value-of-shipments, payroll, or employment total was available for this industry, nor was a concentration ratio or HHI. Establishment count is not equivalent to a smelter count: it also captures alumina refining and integrated forming establishments.

The domestic industry is small relative to both U.S. demand and global supply. USGS estimates 2025 world smelter output at 74 million metric tons, including 45 million tons in China, versus 660,000 tons in the United States.[5]

4. Investable universe

Public companies

Only two listed companies offer meaningful direct exposure to U.S. primary aluminum production:

Neither company is a pure NAICS 331313 investment. Century combines its U.S., Icelandic, and Jamaican operations in one reportable segment. Alcoa reports global Alumina and Aluminum segments; its two U.S. smelters provide domestic upside, but global refinery and smelter costs, currencies, and portfolio decisions are equally important. Their consolidated or global segment margins should not be treated as U.S. industry margins.[7]

Century also holds a proposed 40% interest in an Emirates Global Aluminium joint venture for a greenfield smelter in Inola, Oklahoma, designed for 750,000 tons of annual production. Construction remained subject to engineering, a definitive agreement, and a competitive long-term power contract as of year-end 2025; Century disclosed potential federal support of as much as $500 million. This should be valued as development optionality rather than operating capacity.[7]

More diluted listed exposure is available through international diversified producers such as Glencore, Rio Tinto, Norsk Hydro, and South32, but none is a clean proxy for U.S. primary production. Downstream companies such as Kaiser Aluminum and Constellium can benefit from aluminum demand but are sometimes hurt by high primary-metal premiums; they generally sit outside NAICS 331313.

Major private owners

• Atlantic Alumina: Owns and operates Gramercy, Louisiana, the country's only operating alumina refinery with capability of approximately 100,000 tons per month. About 70% of domestic alumina output went to aluminum smelters in 2025, with the balance serving abrasives, ceramics, chemicals, and refractories. Atlantic Alumina is associated with Concord Resources and a fund managed by Pinnacle Asset Management; in early 2026 the company announced a $450 million strategic partnership with the U.S. government to secure domestic alumina production and build the first large-scale U.S. primary gallium circuit.[10][11]
• Magnitude 7 Metals: Owns the idled New Madrid, Missouri, smelter. In July 2026 it announced plans to restart one 75,000-ton-per-year potline before year-end, with the possibility of further ramp-up in 2027. This is an announced restart, not yet operating production.[12]

Century sold its idled Hawesville, Kentucky, smelter in February 2026 for $200 million plus a 6.8% non-dilutive interest in the buyer's planned data-center project; that facility is no longer a plausible smelter restart.[7]

No current transaction-derived private valuation for Gramercy or New Madrid was established from available sources. Power contracts, utility infrastructure, anode supply, ports, and logistics can provide second-order exposure without taking direct LME operating risk.

5. How the money works

Primary aluminum is a price-taker business with plant-specific costs. The basic economics are:

Revenue = tons shipped × (LME aluminum price + regional delivery premium + product premium).

The regional delivery premium—principally the Midwest premium in the United States—compensates for transport and regional supply-demand conditions. Product premiums apply to higher-value forms such as billet or foundry alloys. An operator can improve its realized price through product mix but normally cannot pass through an idiosyncratic increase in its own power, alumina, or labor costs.[7]

Cost structure: Century Aluminum identifies alumina, electricity, carbon products, labor, and other controllable costs as more than 84% of its 2025 cost of goods sold. Alumina and power were its two largest individual operating inputs. Some alumina and power contracts are LME-linked, creating a partial natural hedge; spot or market-indexed contracts leave the plant exposed to fuel prices, weather, transmission constraints, and grid conditions.[7]

The power sensitivity is unusually tangible. At full production, Century expects annual electricity use of 3,372,600 megawatt-hours at Sebree and 3,504,000 megawatt-hours at Mt. Holly. A $1-per-megawatt-hour change therefore changes annual cost by approximately $3.4 million and $3.5 million, respectively. Sebree buys market-based MISO power, while Mt. Holly has cost-of-service supply through 2031.[7]

Refinery economics are driven by the spread between alumina prices and bauxite, caustic soda, fuel, freight, and residue costs. Smelter economics are driven by LME price plus Midwest and product premiums, less alumina, power, anodes, labor, maintenance, and freight. Vertical integration into bauxite, alumina, anodes, or power reduces supply risk but does not eliminate commodity exposure.

The earnings cycle can be violent. Century reported average LME aluminum prices of $2,630 per ton in 2025, against $2,419 in 2024, while its average Midwest premium rose to $1,295 from $427. Favorable realized metal prices and premiums added $415.9 million to its year-over-year net-sales movement, but higher raw-material and power realizations offset $115.7 million and $93.0 million, respectively, of the gross-profit benefit.[7]

No representative U.S. industry margin could be established from available sources. High fixed costs and limited short-run supply flexibility create substantial operating leverage to metal prices, regional premiums, and utilization.

Useful indicators are LME aluminum price, Midwest premium, alumina price, power cost per megawatt-hour, utilization rate, tons shipped, product mix, inventory days, and working capital intensity.

6. Demand drivers

USGS allocated 2025 U.S. aluminum consumption as follows:[4]

• Transportation (36%): Automotive body panels, castings, heat exchangers, wheels, aerospace structures, and rail equipment. Linked to North American vehicle production and aerospace cycles.
• Packaging (24%): Beverage cans, food containers, and foil. Comparatively steadier demand base with recycling-loop competition.
• Building (13%): Curtain walls, window frames, roofing, siding, and structural components. Construction and commercial-building sensitive.
• Electrical (9%): Transmission and distribution conductors, bus bars, and transformer windings. Benefits from grid expansion and electrification.
• Consumer durables (8%): Appliances, electronics housings, and furniture.
• Machinery (8%): Industrial equipment, HVAC systems, and material handling.
• Other (2%): Defense, sporting goods, and miscellaneous applications.

Electrification is a real but not one-directional demand driver. Aluminum benefits from vehicle lightweighting, battery housings, solar-panel frames, wind equipment, and transmission conductors. The IEA projects annual global aluminum demand from electricity networks rising from 9 million tons in 2020 to 12.8 million tons under stated policies, or 16 million tons in its more aggressive sustainable-development scenario, by 2040. These are global scenario projections, not U.S. forecasts.[13]

An automaker and supplier survey commissioned by the Aluminum Association projects nearly 100 additional pounds of aluminum per North American light vehicle between 2020 and 2030, with electrification more than offsetting aluminum lost from conventional powertrain and transmission components. Because this is trade-association-sponsored research, it is better treated as an industry expectation than an independent forecast.[14]

Recycling is the most important secular countertrend. USGS estimates that 3.6 million tons of aluminum were recovered from purchased scrap in 2025: 56% from manufacturing scrap and 44% from discarded products. Old-scrap recovery equaled about 28% of apparent consumption.[4] Remelting requires only about 5% of the energy needed to make virgin metal, giving recycled aluminum a formidable cost and carbon advantage.[15] Recycling is therefore both a demand enabler for aluminum as a material and a direct substitute for NAICS 331313 output. Primary metal remains necessary where scrap availability, contamination, alloy mixing, or purity requirements prevent closed-loop recycling, but improvements in sorting and scrap utilization should reduce virgin-metal intensity over time.

7. Regulation

Trade policy materially affects pricing. Covered aluminum articles remain subject to a 50% Section 232 duty, although 2026 changes applied lower rates to certain derivative categories.[16] Protection raises the Midwest premium and benefits domestic smelters, but also raises costs for fabricators and consumers, can suppress downstream demand, and could be reduced by a future administration.

Plants face extensive environmental requirements:

• Air emissions: Federal National Emission Standards for Hazardous Air Pollutants (NESHAP) cover potlines, pitch storage, paste production, and anode-bake furnaces; regulated pollutants include hydrogen fluoride, carbonyl sulfide, polycyclic organic matter, mercury, and metal hazardous air pollutants.[17]
• Greenhouse gases: Hall–Héroult cells emit carbon dioxide from anode consumption and highly potent perfluorocarbons during "anode effects."[3]
• Refinery residues: The Bayer process generates high-pH "red mud" containing concentrated naturally occurring radioactive materials and heavy metals; EPA notes that the United States currently approves no secondary use for this waste.[18]

Labor exposure is local and operationally critical. Century reported that 36% of its U.S. workforce was union-represented; its Sebree agreement runs through October 28, 2028, while Mt. Holly is nonunion.[7] Alcoa's 2026 agreement covers approximately 965 workers at Warrick and Massena.[19] A defensible national labor-shortage figure specific to NAICS 331313 could not be established from available sources.

8. Competitive dynamics and consolidation

Concentration is extreme by design. The economics of continuous-process smelting, specialized power infrastructure, and raw-material logistics favor large-scale operations. Only three companies own U.S. smelters, and only one refinery operates domestically.

U.S. primary production has contracted dramatically over decades. Operating capacity has been reduced through permanent closures, mothballing, and curtailments driven by uncompetitive power costs. Century's sale of Hawesville to a data-center developer illustrates that some curtailed capacity will never return to aluminum production.[7] The New Madrid restart announcement in 2026 is an exception driven by elevated premiums, but the facility has undergone repeated shutdown-and-restart cycles historically.[12]

Domestic producers compete with imports—principally from Canada, the United Arab Emirates, Russia (where sanctioned), and other global smelters—and with recycled aluminum. The 60% net import reliance reflects both domestic demand strength and limited domestic primary supply.[4] Section 232 protection supports domestic pricing, but tariff policy is politically and legally changeable.

Consolidation among the few remaining domestic assets is constrained by antitrust and strategic considerations. The industry structure is better described as oligopoly with import competition than as a fragmented market ripe for roll-up.

9. Risks

• Power cost and availability: A smelter with market-indexed electricity can become uneconomic even when metal demand is sound. Century's Hawesville curtailment followed historically high energy costs and falling LME prices; long-term power availability is increasingly contested by data centers and other large loads.[7][12]
• LME price volatility: Primary aluminum is a globally traded commodity; domestic producers cannot set prices and have limited ability to pass through cost increases.
• Midwest premium swings: The regional premium can move sharply based on logistics, trade policy, and supply disruptions—amplifying or offsetting LME movements.
• Raw-material and logistics risks: Imported bauxite and alumina, caustic soda, petroleum coke, pitch, natural gas, and river or ocean freight expose the industry to weather, port, trade, and counterparty disruptions. The single operating U.S. alumina refinery is a concentration risk.[4]
• Trade-policy reversal: Tariffs currently help domestic smelters but are politically and legally changeable; a future administration could reduce or eliminate Section 232 protection.[16]
• Environmental and carbon liabilities: NESHAP compliance, perfluorocarbon emissions, red-mud management, and potential carbon pricing represent ongoing and potentially growing costs.[3][17][18]
• Labor relations: Work stoppages at unionized facilities can halt production; contract renewals carry cost and operational risk.[7][19]
• Substitution: Steel, magnesium, titanium, and composites compete in transportation and structural uses; steel, plastics, glass, and paper compete in packaging; copper competes in electrical applications.
• Recycling competition: Secondary aluminum's 95% energy advantage makes it a direct substitute for virgin metal in many applications, and improvements in scrap sorting and utilization will increase competitive pressure over time.[15]
• Customer and ownership concentration: Century's dependence on Glencore for 54% of sales and 36.4% ownership creates counterparty and governance concentration risk.[7]
• Restart execution risk: Returning curtailed capacity to service requires months of work and significant capital; frozen or damaged pots may not restart successfully.

10. How to invest and outlook

Public investors have two meaningful options for U.S. primary-aluminum exposure. Century Aluminum is the closer proxy for domestic smelter economics, with high sensitivity to LME prices, the Midwest premium, power contracts, and trade policy—but it also has non-U.S. exposure and significant Glencore concentration. Alcoa offers broader, vertically integrated exposure through global operations; its U.S. smelters provide domestic upside, but the stock is not a pure NAICS 331313 investment.

Valuation should use mid-cycle earnings rather than peak metal prices or premiums. Relevant measures include enterprise value to EBITDA, free-cash-flow yield, balance-sheet leverage, and replacement cost per ton of capacity. The 2024–2025 swing in Century's realized prices and margins illustrates why peak-year multiples mislead.

Private investors face an extremely narrow opportunity set: the single domestic alumina refinery (Atlantic Alumina/Gramercy), the announced New Madrid restart, or potential participation in the EGA–Century greenfield development. Diligence should focus on power contracts, alumina and anode supply agreements, environmental liabilities, union obligations, pot-line condition, restart history, and logistics infrastructure. Power availability and pricing are the single most important determinants of whether a U.S. smelter is viable.

LME or CME aluminum contracts provide direct benchmark-price exposure but do not replicate a producer's economics: futures omit the plant's power position, Midwest premium, alumina spread, product mix, operating reliability, and environmental liabilities.

Forward-looking judgment: The volume outlook is cautiously constructive because electrification, lightweighting, and infrastructure investment support aluminum demand across multiple end markets. However, the outlook for domestic primary production is more constrained: limited operating capacity, high power costs, single-refinery alumina supply, and 60% import reliance mean that demand growth benefits global producers and recyclers more directly than U.S. smelters.[4][13][14] Margin prospects depend heavily on trade policy: current Section 232 protection supports elevated Midwest premiums that make domestic smelting viable, but that support is politically vulnerable.[16] The best-positioned domestic operators are those with long-term, competitive power contracts, product-mix flexibility toward higher-value forms, and the operational discipline to run at high utilization when spreads are favorable and preserve capital when they are not.

Sources

1. U.S. Census Bureau, 2022 NAICS Manual, 2022, https://www.census.gov/naics/?details=33&input=33&year=2022
2. U.S. Environmental Protection Agency, Bauxite and Alumina Production Wastes, 2025, https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100RH6N.TXT
3. U.S. Environmental Protection Agency, Aluminum Industry Partnership, 2025, https://www.epa.gov/eps-partnership/aluminum-industry
4. U.S. Geological Survey, Mineral Commodity Summaries 2026: Aluminum, 2026, https://pubs.usgs.gov/periodicals/mcs2026/mcs2026-aluminum.pdf
5. U.S. Geological Survey, Mineral Commodity Summaries 2026, 2026, https://pubs.usgs.gov/periodicals/mcs2026/mcs2026.pdf
6. U.S. Census Bureau, NAICS 331313 Industry Profile, 2023, https://data.census.gov/profile/331313_-_Alumina_Refining_and_Primary_Aluminum_Production?codeset=naics~331313
7. Century Aluminum Company, 2025 Form 10-K, 2026, https://www.sec.gov/Archives/edgar/data/949157/000162828026013788/cenx-20251231.htm
8. Alcoa Corporation, Massena Long-Term Energy Contract Announcement, 2025, https://investors.alcoa.com/press-releases/press-release-details/2025/Alcoa-Corporation-Secures-Long-Term-Energy-Contract-and-Announces-60M-Capital-Investment-in-U-S--Smelter/default.aspx
9. Alcoa Corporation, Warrick Operations Fact Sheet, 2025, https://www.alcoa.com/global/en/pdf/Alcoa-Warrick-Fact-Sheet.pdf
10. Atlantic Alumina, Company Overview, 2026, https://www.atalco.com/
11. Business Wire, Atlantic Alumina Announces $450 Million Strategic Partnership with United States Government, 2026, https://www.businesswire.com/news/home/20260111484503/en/Atlantic-Alumina-Announces-%24450-Million-Strategic-Partnership-with-United-States-Government-to-Secure-U.S.-Alumina-Production-and-Build-Americas-First-Large-Scale-Primary-Gallium-Production-Circuit
12. Mining Weekly (Reuters), Aluminium Supply Shock Revives Long-Idled Western Smelters, 2026, https://www.miningweekly.com/article/aluminium-supply-shock-revives-long-idled-western-smelters-reuters-says-2026-07-10
13. International Energy Agency, The Role of Critical Minerals in Clean Energy Transitions, 2021, https://www.iea.org/reports/the-role-of-critical-minerals-in-clean-energy-transitions/mineral-requirements-for-clean-energy-transitions
14. The Aluminum Association, New Survey of Automakers Confirms Aluminum Use Expected to Grow in New Electric Vehicles, 2022, https://www.aluminum.org/news/new-survey-automakers-confirms-aluminum-use-expected-grow-new-electric-vehicles
15. U.S. Department of Energy, Alcoa Aluminum Recycling Facility, 2024, https://betterbuildingssolutioncenter.energy.gov/showcase-projects/alcoa-aluminum-recycling-facility
16. White House, Further Adjusting the Tariff Regimes for Imports of Aluminum, Steel, and Copper, 2026, https://www.whitehouse.gov/presidential-actions/2026/06/further-adjusting-the-tariff-regimes-for-imports-of-aluminum-steel-and-copper-into-the-united-states/
17. U.S. Environmental Protection Agency, Primary Aluminum Reduction Industry National Emission Standards, 2025, https://www.epa.gov/stationary-sources-air-pollution/primary-aluminum-reduction-industry-national-emission-standards
18. U.S. Environmental Protection Agency, TENORM: Bauxite and Alumina Production Wastes, 2025, https://www.epa.gov/radiation/tenorm-bauxite-and-alumina-production-wastes
19. Alcoa Corporation, Ratification of Labor Agreement at U.S. Smelters, 2026, https://news.alcoa.com/press-releases/press-release-details/2026/Alcoa-Announces-Ratification-of-Labor-Agreement-at-U-S--Smelters/default.aspx