NAICS 331 — Primary Metal Manufacturing

A rollup investor primer. NAICS is the North American Industry Classification System, the U.S. government's standard code for industries. This is a subsector (3-digit) inside Sector 33 (Manufacturing). It gathers five child industry groups (4-digit): 3311 Iron and Steel Mills and Ferroalloy Manufacturing, 3312 Steel Product Manufacturing from Purchased Steel, 3313 Alumina and Aluminum Production and Processing, 3314 Nonferrous Metal (except Aluminum) Production and Processing, and 3315 Foundries. This page synthesizes the five child primers and adds the federal figures that exist only at this combined level. It is written for a general investing audience — public-market and private investors alike.

---

1. Overview

NAICS 331 is where the economy makes the metal everything else is built from — and then shapes, recycles, and casts it. It starts with raw production (melting iron ore and scrap into steel, smelting alumina into aluminum, refining copper and other nonferrous metals from ore) and runs through the mid-stream that reshapes purchased metal (pipe, tube, sheet, wire, rod, profiles) and the foundries that pour molten metal into finished parts. If a physical object moves, holds pressure, carries current, or bears weight, some of the metal inside it passed through a plant counted here.

For investors the single most important framing is that 331 is not one market — it is five, and they barely compete with each other. A steel minimill, a copper smelter, an aluminum rolling mill, a wire-drawing shop, and an iron foundry sit in the same subsector but serve different customers, earn money differently, and are owned by different kinds of people. The distinctive value of reading this rollup rather than a single child primer is the contrast across the five — which is biggest, which is growing, who owns each, how concentrated each is, and how their economics differ. That contrast comes first.

Two threads run through all five. First, with the narrow exception of the primary steel mills, there is no clean, sizable U.S.-listed pure play anywhere in this subsector. Public exposure is either concentrated in a short list of steel producers, diluted inside diversified manufacturers and foreign strategics, or simply absent — replaced by a deep base of private, family-owned, and private-equity-held companies. Second, and newly quantified across the revised children, profit in this subsector tracks qualification difficulty and route, not tonnage or size. The reported margins of the high-volume middle sit in the low-to-mid single digits; the qualified, certified corners report margins several times that. Section 5 lays the ladder out.

---

2. What's inside — the five children and how they differ

All five children work metal, but they occupy different rungs of the same chain. Read the contrast table top to bottom and a pattern appears: as you move from making metal toward shaping and casting it, revenue-per-plant falls, concentration falls, pay falls, and ownership shifts from public toward private. The revenue giant of the subsector (iron and steel mills) runs the fewest plants with the fewest workers; the employment-and-plant giant (foundries) earns the least revenue.

Acronyms used below: HHI = Herfindahl-Hirschman Index (an antitrust concentration gauge on a 0–10,000 scale, where U.S. agencies treat readings under 1,500 as "unconcentrated"); CR4 = combined revenue share of the four largest firms; EAF = electric-arc furnace (the scrap-melting "minimill" route); PE = private equity. Shares of the level are our own arithmetic on the child figures; all reconcile to the subsector totals in Section 3. All concentration figures on this page come from the same 2022 Economic Census file, so they are comparable with each other — and, as every child now warns, they are a 2022 snapshot rather than a current reading.[1]

The first divergence: scale versus headcount. Two children anchor the extremes. 3311 (iron and steel mills) is the revenue heavyweight — 40% of the subsector's sales from just 10% of its plants and 22% of its workers, at an extraordinary ~$1.65 million of revenue per employee and ~$116,000 of pay per employee.[2] It is capital-intensive, concentrated (HHI 1,072), and the only child with genuine listed pure-plays. 3315 (foundries) is its mirror image — 30% of the jobs and 37% of the plants but only 10% of the revenue, at ~$305,000 of revenue and ~$67,000 of pay per worker, spread across 1,335 mostly small, mostly private shops (HHI 170).[6] In between sit the three "buy-and-reshape / smelt-and-refine" children (3312, 3313, 3314), each earning a conversion margin on metal it does not itself dig out of the ground, each with limited and diluted public exposure. The level's own averages — ~$910,000 of revenue and ~$84,000 of pay per worker — sit between the two poles and describe none of the five well.

The second divergence: margin does not follow size. The revised children now report enough company evidence to order the subsector by profitability, and the order does not match the order by tonnage. The high-volume middle — iron casting, die-casting, flat-rolled steel processing — reports low-to-mid single-digit margins; the qualified, certified corners — precision aerospace casting, specialty alloys — report margins in the twenties and thirties.[3][5][6] Section 5 lays out the ladder and its caveats. The important structural point for this level is that the widest margin spreads sit inside children, not between them: in 2025 alone, one integrated steel producer reported a negative 5% steelmaking gross margin while a minimill peer earned about a 10.6% operating margin on its steel operations.[2]

The third divergence: measured concentration falls every time you aggregate. This is now a documented pattern at every rung, not a quirk of the subsector. 331's HHI of 234.7 is lower than 3311's 1,072.[1][2] But so is 3312's HHI of 196 lower than both its own children (388.8 and 249.8); 3314's 235.8 lower than all three of its children (881.9, 527, 337.6); and 3315's 170.4 lower than both of its children (337.4 and 177.6).[3][5][6] Bundling non-competing markets mechanically dilutes measured concentration, so the number you should trust is always the one from the narrowest level where firms actually compete — and, as Section 3 explains, at six digits that number frequently does not exist.

Why the split matters for investing. The five do not rise and fall in lockstep. Steel mills track construction and autos; aluminum leans on packaging and transportation; copper rides the electrical grid and data centers; nonferrous specialty metals ride aerospace and defense; foundries straddle vehicles, machinery, and a high-value aerospace-precision niche. The children now put numbers on the de-correlation: North American aluminum extrusion demand fell 3.1% in 2025 while total aluminum demand rose 0.8% and the aluminum-recycling output index jumped from 93.0 to 108.6;[4] iron-foundry output prices rose about 2.6% year over year while other-nonferrous casting prices rose about 11% in 2025 and a further 9% in five months of 2026.[6] That partial de-correlation is why the subsector as a whole is steadier than any one child — and why "primary metals" is a poor thing to buy as a single idea and a useful map for choosing which metal, and which rung of the chain, you actually want.

---

3. Size (this level's rollup figures)

Our ground-truth federal statistics for NAICS 331 as a whole:[1]

So the whole subsector is roughly a $323-billion-revenue, 355,000-job business run from about 3,570 plants and 2,387 companies. Revenue per worker averages about $910,000 and annual pay about $84,000 a head — both high, reflecting a capital- and skill-intensive, heavily unionized field where materials, not labor, dominate the cost base. (County Business Patterns is the Census Bureau's annual establishment tally; the Economic Census is its every-five-years deep count.)

One operating series exists only at this level. The Federal Reserve publishes capacity utilization for the whole primary-metals category and not for anything inside it: 66.7% in 2025, against 67.0% in 2024 and 73.0% in 2022.[10] That matters because 3314 reports that no utilization series is published anywhere within its group, so this subsector-wide figure is the only read available for copper and specialty-metal processing.[5] A narrower Federal Reserve series for iron and steel products averaged 72.2% in 2025, which 3312 and 3315 use as their proxy.[3][6] Both readings sit below the 2022 level — real slack in a business whose profits are driven by keeping plants full (Section 5).

The children reconcile to the level — cleanly, within one statistical family. The five children's employment sums to exactly 354,914 and their plants to exactly 3,570 — unit-for-unit matches. Receipts (the children sum to ~$323.1B) and annual payroll (~$29.69B) match within rounding. That is a rare and reassuring result. It holds, however, only inside the County Business Patterns and Economic Census series; where independent sources exist they disagree with them (see the measurement caveat below).

Firm counts do not add up — and that is expected, at every rung. The children report 73 + 457 + 257 + 465 + 1,211 = 2,463 firms, but the subsector reports 2,387. The ~76-firm gap is the normal NAICS artifact: a company active in more than one child (say, an integrated producer that both makes steel and casts parts) is counted once at the 331 level but once in each child it operates in. The same artifact repeats inside three of the children — 3312's children report 467 firms against a group total of 457, 3314's report 481 against 465, and 3315's report 1,222 against 1,211.[3][5][6] Multi-industry operators are common throughout.

Measurement quality is excellent here and collapses one level down. This is the sharpest new observation this pass. NAICS 331 and all five of its children publish unsuppressed receipts, firm counts, and concentration ratios. Below them, federal revenue data largely disappears: 3313 reports no unsuppressed six-digit receipts, firm count, or concentration for any of its four industries; 3314 reports none for any of its four six-digit codes; 3315 has exactly one of six grandchildren with a current measured receipts line (steel investment foundries, $3.95 billion across 84 firms); and neither of 3312's rolling-and-drawing grandchildren could establish its own revenue or concentration.[3][4][5][6] Only 3311 is clean all the way down, because it contains a single child at every level.[2] The practical consequence: this rollup and its five children are the finest resolution at which the subsector can be measured, even though competition happens below them.

Undercount and coverage caveats.

• Nonemployers omitted, but that gap is trivial here. County Business Patterns counts only employer establishments — those with payroll — and leaves out self-employed people with no employees.[1] For a furnace-and-mill subsector that needs heavy equipment, power, and working capital, this is not a cottage trade; all five children reach the same conclusion independently.[2][3][4][5][6] (The one soft edge is the smallest tier of family bronze and copper foundries in 3315.)
• Alternative federal and trade sources disagree with the census series inside four children. County Business Patterns puts 3311 employment at 78,532, while the Census Bureau's own six-digit industry profile reports average employment of about 86,000 — an unreconciled ~10% gap.[2] An EPA rulemaking memorandum reproducing Annual Survey of Manufactures data puts 2021 shipments for nonferrous smelting and refining at $15.905 billion, roughly $3.6 billion above the 2022 Economic Census figure for the same activity.[5] The American Wire Producers Association counts 260 wire-drawing facilities and 10,518 employees against County Business Patterns' 198 and 14,260.[3] And the Bureau of Labor Statistics works from a foundry employment base of 106,000 rather than County Business Patterns' 107,273.[6] None of these is a correction of the other — they are different programs, universes, and years. Hold ranges, not points, and never mix series across years.
• The ownership blind spot is real but no longer uniform across the children. A large share of this subsector is private, foreign-owned, or captive, and none of it reports on a NAICS basis. Foreign strategics (Norsk Hydro, Rio Tinto, Hindalco/Novelis, UACJ, Gränges, Emirates Global Aluminium) own much of aluminum;[4] copper fabrication and the specialty tail are dominated by private and family mills;[5] foundries are overwhelmingly family-, employee-, and PE-owned.[6] In those three children, public stock-market value clearly understates the true size of the activity. The revised 3311, however, explicitly withdrew that claim for steel mills: the direction remains plausible, but no public-versus-private share of output could be established from available sources, and neither could the share of revenue attributable to ferroalloys.[2] Treat both as open questions there. Separately, captive plants inside vehicle, machinery, and battery makers are counted under their parent's finished-product code, not here — so the real physical footprint is larger than 3,570 plants.[4][5][6]
• Receipts are a gross, metal-inflated figure. The $323 billion double-counts metal that flows from one stage to the next (a mill sells steel to a pipe maker who sells to a foundry) and rises and falls with metal prices that pass straight through every invoice. Every child says the same thing about its own receipts line.[3][4] Read it as gross revenue, not economic output or value added.
• Vintage mix. Receipts, firm count, and concentration are 2022 Economic Census; employment, plants, and payroll are 2023 County Business Patterns.

---

4. Investable universe — where value concentrates across the children

Two facts organize the whole subsector for an investor.

First, concentration is genuinely low in aggregate — but that is mostly an aggregation artifact. Our 2022 figures put the top four firms at 26.4% of subsector revenue, the top eight at 34.7%, the top twenty at 49.2%, the top fifty at 64.6%, and the HHI at just 234.7[1] — well inside the "unconcentrated" band. But that number pools five non-competing markets. The genuinely concentrated core is iron and steel mills (3311, HHI 1,072, top four 59.5%, top fifty 99.7%), which alone is 40% of revenue; averaging it against four fragmented children pulls the subsector figure down toward the fragmented middle. The revised children add a second warning on top: even a child's own ratio is broader than the market being fought over. Copper fabrication looks "moderately fragmented" across the category while being concentrated within individual product families — wire rod, brass rod, rolled strip, tube — that require different furnaces and customer qualifications.[5] Aluminum sheet looks unconcentrated at 3313's HHI of 500, yet seven producers told the U.S. International Trade Commission they accounted for 91.4% of U.S. common-alloy sheet production in 2023.[4] The precision-casting corner of foundries has a top-four share of about 73.5% inside a child whose HHI is 337.[6] The lesson every child now repeats: read concentration at the level where firms actually compete — and expect the published data to run out before you get there.

Level figures[1]; child figures from the child primers.[2][3][4][5][6] All are 2022 Economic Census readings; no child could establish a current concentration figure.

Second, where value sits for a public investor is uneven — inversely related to physical bulk, and on balance narrowing. The cleanest listed exposure clusters at the two ends: the concentrated steel-mill core, and the narrow high-specification precision-casting niche. The high-tonnage middle (pipe, wire, rolled aluminum, copper mills, iron foundries, die-casting) is mostly private. Three of the five children reported that listed access shrank this pass rather than widened. Where value concentrates, child by child (tickers belong to the investable universe, not the NAICS definition):

• 3311 — iron & steel mills (the one clean public read). The only child with sizable pure-plays: Nucor (NUE), Steel Dynamics (STLD), and Cleveland-Cliffs (CLF) are listed U.S. steel producers; a steel-industry exchange-traded fund (ETF) offers a diversified basket. The first decision here is route, not ticker — EAF-focused exposure with downstream integration versus integrated ore-to-steel exposure with automotive concentration and higher operating leverage; the two behaved very differently in 2025. One marquee pure-play, U.S. Steel, left the market when Nippon Steel completed its acquisition on June 18, 2025.[2][15]
• 3312 — steel products (niche-and-illiquid or diluted, and the diluted names got thinner). The one genuinely clean listed read is small: Insteel Industries (IIIN) in wire and reinforcement, at $647.7 million of fiscal 2025 sales. Narrower slices via NWPX Infrastructure (NWPX, whose water-transmission segment was 66.7% of 2025 sales) and Friedman Industries (FRD, whose tubular products were 9% of fiscal 2025 sales). Broader-but-diluted exposure via Nucor (NUE), Tenaris (TS), Vallourec (French-listed), Atkore (ATKR), Worthington Steel (WS), and Ryerson (RYZ). Note the direction: Worthington's majority acquisition of Klöckner and Ryerson's merger with Olympic Steel — both closing in 2026 — made those names bigger but thinner reads on actual conversion processing. Much of the tonnage is private/PE.[3]
• 3313 — aluminum (listed specialists, partial). Conversion specialists Kaiser Aluminum (KALU) and Constellium (CSTM) in sheet and plate, Tredegar (TG) in extrusion (its Bonnell business was 86% of 2025 net sales), with Steel Dynamics (STLD) as the large new sheet entrant (aluminum was 2% of its 2025 sales) and Alcoa (AA) and Century Aluminum (CENX) as upstream commodity plays. Century carries a real counterparty concentration — Glencore owned 36.4% of it at year-end 2025 and accounted for roughly 54% of consolidated sales. Recycling has no listed pure-play; foreign strategics (Norsk Hydro, Rio Tinto, Hindalco/Novelis, Emirates Global Aluminium) and private/family recyclers hold much of it.[4]
• 3314 — nonferrous ex-aluminum (least-diluted in copper, and narrowing). Mueller Industries (MLI) is the broadest U.S. copper-fabrication read — its Industrial Metals segment did $1.024 billion of 2025 sales at a 16.0% gross margin — and Materion (MTRN) the specialty one; ATI (ATI) and Carpenter Technology (CRS) for specialty shaping, at 68% and 62% of fiscal 2025 sales from aerospace and defense respectively. Primary smelting reaches investors only inside diversified miners (Freeport-McMoRan, FCX, whose integrated domestic facilities supply approximately 70% of U.S. refined copper production, and Rio Tinto's Kennecott complex); recovery is almost entirely private. Direct public pure-play exposure largely disappeared with the Wieland/Global Brass (2019) and Prysmian/Encore Wire (2024) transactions.[5]
• 3315 — foundries (precision only, all partial — but slightly wider than last pass). Still no pure-play. The cleanest large-cap proxy is Howmet Aerospace (HWM), ~$8.3 billion of 2025 revenue with ~70% from aerospace and an Engine Products segment at a 33.3% adjusted EBITDA margin. Doncasters/DPC Holdings (DPC) listed on the NYSE on June 25, 2026 at $33 a share, raising ~$919.3 million, on $837 million of fiscal-2025 revenue with 77% from aerospace and industrial gas turbines — the clearest casting economics of any listed name. Berkshire Hathaway (BRK.B) owns Precision Castparts. New this pass: Air T (AIRT) holds a 20.1% equity-method interest in a merchant iron foundry — the only direct listed ownership of one, buried inside an aviation holding company. Iron and die-casting — the bulk of the tonnage — remain overwhelmingly private, and die-casting has no clean U.S. listing at all.[6]

Reality check. Outside the steel-mill core, listed exposure to 331 is either narrow-and-illiquid or heavily diluted by other businesses, and the biggest, highest-tonnage assets are private. Corporate segment revenue for any diversified name should never be read as NAICS industry revenue — segments fold in foreign plants, adjacent activities, and other metals, and no public company in this subsector reports on a six-digit NAICS basis.[5][6] The private and PE tail is where much of the volume, and much of the transactable opportunity, actually lives.

---

5. How the money works

Across all five children a single truth governs: reported revenue is a poor guide to profit, because the price of the metal passes straight through it. Rising metal prices inflate sales, inventory, and receivables without necessarily improving earnings. The cleanest illustration anywhere in the subsector is aluminum: Kaiser's 2025 net sales rose to $3.37 billion even as shipments fell 5%, because the hedged cost of alloyed metal rose 29% and passed through — while conversion revenue, the number that measures the mill's actual work, was essentially flat at $1.45 billion.[4] What operators earn splits into two archetypes:

• Commodity-spread businesses — the primary producers. Integrated steel mills, primary aluminum smelters, and primary copper smelters earn the metal price minus the cost of ore, scrap, energy, and processing. Their margin is a spread that swings hard with the commodity cycle, and their economics turn on scale, power cost, and running flat-out. Steel Dynamics reports that metallic inputs are normally 55%–65% of steel-mill manufacturing cost;[2] Century Aluminum expects annual electricity use of roughly 3.4 million megawatt-hours at each of two smelters, so a $1-per-megawatt-hour move changes annual cost by about $3.4 million per site.[4] Merchant smelters instead earn treatment and refining charges — and 2025 showed how brutal that model can be, with the annual copper treatment-charge benchmark at $21.25 per metric ton and some contracts agreed at no processing fee at all.[5]
• Conversion-margin businesses — the reshapers and casters. Pipe, tube, wire, rolled aluminum, copper fabrication, and foundries buy metal at market price, add a shaping step, and sell it at metal cost plus a conversion charge. For aluminum the selling price has a clean shape: London Metal Exchange (LME) price + regional metal premium + conversion premium.[4] Copper fabrication is the same idea: metal cost + conversion charge + product premium + freight.[5] One structural variant exists in only one corner of the subsector — inside 3312's rolled-shape child, plants can run toll processing, where the customer keeps title to the metal and pays only a conversion fee, which is the single genuine dial for turning commodity exposure up or down anywhere in 331.[3]

There is no "industry margin," and the evidence is now specific enough to say so. The company proxies below come from different children, currencies, periods, segment definitions, and profit measures (gross, operating, EBITDA, adjusted EBITDA). None is an industry average and they cannot be compared decimal for decimal. Read the ordering only:

• Iron foundry conversion work: a UK-listed foundry's segment profit was under 4% of revenue.[6]
• Aluminum die-casting: a Japanese comparable at a 4.1% operating margin; a Mexico-listed one at about 14% EBITDA.[6]
• Heavy steel casting: a modeled ~4.7% pre-tax rate from an OSHA regulatory analysis — a screening estimate, not observed earnings.[6]
• Flat-rolled steel processing: Worthington Steel at a 7.0% adjusted EBITDA margin on $3.093 billion of fiscal 2025 net sales.[3]
• Steel wire and reinforcement: Insteel at a 14.4% gross margin on $647.7 million of fiscal 2025 sales.[3]
• Copper fabrication: Mueller's Industrial Metals segment at a 16.0% gross margin, $105.0 million of operating income on $1.024 billion of 2025 sales.[5]
• Precision casting: Doncasters at a 16.5% adjusted EBITDA margin for 2025; Howmet's Engine Products at 33.3%.[6]
• Specialty alloys: Carpenter's Specialty Alloys Operations at a 23.0% operating margin on $2.564 billion of sales — while shipment volume fell 11%.[5]

The gap from roughly 4% to roughly 33% is far too wide to be a definitional artifact. It is the subsector's central asymmetry: margin rises with qualification difficulty and value density, not with tonnage, headcount, or company size. And the widest spreads are within children, not between them — 3311 held a −5% steelmaking gross margin and a 10.6% steel-operations operating margin in the same year;[2] one tubular segment's gross margin fell from 40% to 14% between 2023 and 2024 on almost unchanged shipments;[3] Doncasters' own gas-turbine work earns more than three times what its transportation turbo-wheel work does.[6]

Three consequences hold across every child:

1. Utilization is the master dial. Furnaces, rolling lines, presses, and molding lines cost the same idle or busy, so keeping them full drives profit; modest volume swings produce large margin swings, up fast and down faster. Nucor's steel-mill utilization rose from 76% in 2024 to 83% in 2025, more tonnage over the same asset base.[2] The subsector-wide figure is much lower: primary-metals capacity utilization was 66.7% in 2025,[10] and aluminum extrusion — the level's softest corner — ran near two-thirds, having fallen from 83.5% in 2021 to 67.7% in early 2024.[4]
2. Inventory timing makes or breaks a quarter. When metal prices fall, firms holding higher-cost stock take holding losses; when prices rise, they book windfalls. Insteel quantifies the exposure: a 10% increase in wire-rod cost would have cut annual pretax earnings by $37.5 million had selling prices not moved.[3] Roughly a quarter of U.S. steel shipments go to distributors,[2] whose restocking and destocking make order books swing harder than end demand.
3. Physical metrics beat headline sales. Watch tons or good parts shipped, conversion margin per ton, capacity utilization, casting/metal yield, by-product credits (in smelting), maintenance capital, and inventory days — not revenue. Maintenance is lumpy and easy to under-model: one copper smelter turnaround generated $73 million of maintenance and idle-facility costs on a three-to-four-year cycle,[5] and foundry maintenance capital is frequently higher than book depreciation implies.[6]

Because pass-through metal drives the profit-and-loss statement, the frameworks used elsewhere in this taxonomy do not apply here: there is no regulated-utility rate base, no real-estate-investment-trust funds-from-operations, and no mining all-in-sustaining-cost. All five children say so independently. The right lens across the subsector is processing spreads, utilization, and physical throughput.

---

6. Demand drivers

The children lean on overlapping but distinct end markets — which is exactly why the subsector is diversified in aggregate even though each child is cyclical alone. The revised children now put shipment-mix numbers on most of them.

• Construction and infrastructure (all five): non-residential building, bridges, water systems, and public works pull steel, aluminum extrusion, copper, structural castings, pipe, mesh, and reinforcement. Construction is 31% of 2025 U.S. net steel shipments,[2] 42% of U.S. copper and copper-alloy use,[5] 13% of aluminum consumption (and 61% of one listed extruder's sales),[4] and the steadiest demand ferrous foundries have, via municipal water pipe, valves, and hydrants.[6] One listed wire producer put 85% of its fiscal 2025 sales in nonresidential construction.[3] Federal spending under the Infrastructure Investment and Jobs Act, reinforced by domestic-content rules, is a multi-year tailwind.[3]
• Distributors and service centers (mainly 3311, 3312): 26% of 2025 U.S. net steel shipments go to intermediaries whose restocking and destocking can temporarily dominate real end-market demand — a genuine amplifier of the subsector's cycle, and one that also sits between specialty mills and their customers (about 40% of one producer's plate volume moved through independent service centers).[2][5]
• Autos and transportation (steel, aluminum, foundries): automotive is 15% of 2025 U.S. net steel shipments;[2] transportation is the largest U.S. aluminum end market at ~36% of 2025 consumption[4] and 18% of copper use.[5] Lightweighting shifts share toward aluminum die-casting and structural "gigacastings" — the Department of Energy credits aluminum with 30–60% mass-reduction potential, and a 10% vehicle-weight cut with a 6–8% fuel-economy gain.[6] Electric-vehicle adoption redistributes rather than lifts casting demand, eroding iron and aluminum engine work while creating structural-casting demand.[6]
• The electrical grid, data centers, and electrification (copper, aluminum, foundries): electrical and electronic products are 23% of U.S. copper use,[5] and the Energy Information Administration forecasts U.S. electricity demand growth of 1.9% in 2026 and 2.5% in 2027 with data centers an important contributor.[5] A newly identified pocket sits in 3312: electrical steel for motors and transformers, where roughly 55% of in-service distribution transformers are more than 33 years old.[3] Data centers cut both ways in this subsector — they pull copper, gas-turbine castings, and grid steel, and they compete directly for the cheap power that primary aluminum smelting needs (Section 8).
• Packaging (aluminum): beverage and food cans are 24% of 2025 aluminum consumption, the most defensive aluminum market and the cleanest source of recyclable scrap.[4]
• Aerospace, defense, and gas turbines (specialty nonferrous + precision foundries): heat-resistant nickel alloys, titanium, and superalloy turbine blades ride record aircraft backlogs, defense budgets, and data-center-driven turbine demand — the strongest single tailwind in the subsector. The listed specialty producers derive 68% and 62% of fiscal 2025 sales from aerospace and defense;[5] one precision caster's defense-aerospace revenue rose 21% and its industrial-gas-turbine revenue 25% in 2025.[6]
• Energy, machinery, agriculture, mining, and rail (across children): oil-country tubular goods and line pipe, cold-finished bar, wear parts, valves, and hydrants. Energy is 3312's most volatile end market and the energy transition cuts both ways — threatening conventional tubulars while creating demand for carbon-dioxide, hydrogen, and geothermal tubulars.[3]

Two cross-cutting themes. The first is supply security and reshoring: a durable national preference for domestic sourcing of strategic metals — copper and lead were both added to the federal critical-minerals list in 2025,[5] and steel, aluminum, and copper are all under Section 232 — supports every child. It is not cosmetic: the United States imports 60% of apparent aluminum consumption and produced 660,000 tonnes of primary aluminum against 1.31 million tonnes of nameplate capacity;[4] net import reliance for refined copper is an estimated 57%;[5] and the country produced no titanium sponge and no primary refined lead in 2025.[5] Steel is the exception that proves the point, with net imports supplying only 13% of what the country used.[2] The second is that near-term demand optimism sits inside a longer contraction in the subsector's largest employer: an industry survey projects foundry casting sales up 4.2% in 2026 — but 52% of that is pricing and only 48% tonnage — while the Bureau of Labor Statistics projects foundry employment falling from 106,000 in 2024 to 90,400 in 2034, a 14.7% decline.[6][11] Because these are late-cycle capital-goods and building end-markets, the subsector rises and falls with interest rates, construction starts, and vehicle and aircraft build rates.

---

7. Regulation

Three forces dominate the whole subsector, and they are common to every child.

Trade policy is the single biggest lever — and the children now show it is a transfer, not a subsidy. Section 232 of the Trade Expansion Act of 1962 lets the President tax imports on national-security grounds. As of mid-2026 the covered rate on core steel, aluminum, and copper articles is generally 50%, with 25% on qualifying derivative articles, subject to metal-specific exceptions (qualifying United Kingdom aluminum at 25%, Russian-origin aluminum at 200%, and tiered treatment for copper derivatives).[12][13][14][15][17] Protection helps domestic primary producers by keeping cheap imports out — but it raises input costs for the reshaping and casting children (3312, the rolling and fabrication layers of 3313 and 3314, and all of 3315), which buy domestic coil, plate, cathode, and alloy at tariff-inflated prices. The U.S. International Trade Commission's study of the earlier Section 232 measures quantifies both sides: affected imports fell 24%, U.S. steel-product prices rose 2.4%, and domestic production rose 1.9% on average over 2018–2021 — while raising downstream prices and reducing downstream production, with U.S. importers bearing nearly the full tariff cost.[9] The behavioral evidence is visible inside the children too: one wire producer's imported share of rod purchases rose from 15% to 27% in fiscal 2025 as the domestic price wall went up.[3]

Protection is also not uniform, and the children correct a common assumption here. Section 232 covers the metal; targeted antidumping and countervailing orders do not follow the same map. Orders on common-alloy aluminum sheet and foil remain in force, reinforcing that industry's cover — but the broad 2024 aluminum-extrusion case covering 14 trading partners ended without new orders after the Commission found no material injury, even though Commerce made affirmative dumping and subsidy findings.[4] A live oil-country-tubular-goods investigation initiated in April 2026 alleges dumping margins from 43.64% to 126.08% depending on origin — petition allegations, not final duties.[3] And a July 2026 proclamation offers approved investors in new or expanded U.S. primary aluminum capacity a half-rate import allowance, an explicit push to reshore smelting that does nothing for imported sheet or extrusions.[16] Do not assume "metals tariffs" mean the same thing for every child, or that the regime is permanent — the copper framework alone was rewritten twice within three months of 2026.[5] Domestic-content rules (Build America, Buy America, which generally require U.S. iron and steel through all manufacturing processes) reinforce the tailwind for federally funded infrastructure.[3]

Environmental rules grow heavier the hotter and dirtier the process. The Environmental Protection Agency sets hazardous-air-pollutant standards (National Emission Standards for Hazardous Air Pollutants, NESHAP) that bite hardest on integrated steel mills, primary aluminum and copper smelting, and secondary-lead recovery, and lightest on clean rolling and drawing.[2][4][5] Three separate effluent-guideline regimes apply across the subsector and are easy to confuse — nonferrous metals manufacturing, copper forming, and nonferrous forming — plus a fourth for foundries.[5][6] Compliance costs are real but not uniformly ruinous: one large steel producer puts environmental-compliance capital expenditure at existing facilities below $100 million a year in both 2026 and 2027,[2] while the EPA calculated that requiring wet electrostatic precipitators at eight secondary-lead facilities would cost $621 million upfront and $73 million annually — and did not propose it.[5] Decarbonization is a genuine multi-year capital theme for the primary producers; in steel it favors lower-emitting electric-arc production, which already makes 70% of domestic steel, over blast furnaces.[7]

Worker safety and labor are now regulatory-grade constraints, not just cost lines. The Occupational Safety and Health Administration limits respirable silica in foundries to 50 micrograms per cubic meter, and exposure to lead, beryllium, and copper fume in smelting.[5][6] The outcome shows in the injury data: every child that publishes a recordable injury rate runs above the 2.3-per-100 private-industry benchmark — iron foundries 5.5, steel foundries except investment 5.6, die-casting 4.6, other nonferrous castings 4.5, steel investment foundries 4.0, nonferrous smelting and refining 4.1 — and 3312's rolled-shape child also reports a rate above the private-industry average.[3][5][6] On the labor side, more than 90% of one integrated steelmaker's hourly workforce is union-represented, with major site agreements up for negotiation in 2026, alongside shortages of electricians and other experienced industrial workers; work stoppages, wage escalation, and retiree obligations bear hardest on integrated mills.[2] Skilled-worker scarcity — metallurgists, furnace operators, molders, patternmakers, inspectors — is cited independently by 3312, 3314, and 3315 as well.[3][5][6] Defense sourcing rules (the DFARS specialty-metals clause) favor qualified domestic mills throughout.[5][6] A strong permit is a double-edged asset everywhere: a high barrier to entry that protects incumbents, and a standing liability if emissions or legacy contamination force shutdowns.

---

8. Consolidation

The aggregate picture is deceptively fragmented (HHI 234.7), but consolidation is real and takes a different form in each child.

• 3311 — actively concentrating at the top, against a rising external threat. Already the most concentrated child, iron and steel mills continue to consolidate: Nippon Steel completed its acquisition of U.S. Steel on June 18, 2025, taking a marquee American pure-play off the public market, while integrated capacity has clustered under a few large producers and private and foreign owners hold regional, specialty, and ferroalloy assets.[2][15] Further consolidation may be limited by antitrust review, particularly for deals combining integrated capacity serving automotive customers. The larger competitive pressure is external: the OECD estimates global excess steelmaking capacity reached 640 million tonnes in 2025 and projects 745 million tonnes by 2028 — displaced foreign steel depresses world prices and reroutes through third countries even where tariffs block direct imports.[8]
• 3312 — consolidating within a still-fragmented base, and the centre of gravity has moved. This is the clearest correction of this pass. The group's two largest transactions are now on the rolled-shape side, not in wire: Worthington Steel took approximately 62% of Klöckner & Co (June 3, 2026) and Ryerson closed its merger with Olympic Steel (February 13, 2026), with former Olympic holders taking roughly 37% of the combined company. Wire drawing is still shedding plants — from 252 establishments in 2017 to 198 in 2023, a 21% decline, while rolled shapes rose 20% from 204 to 244 — but its deals are small bolt-ons, on the order of $67 million and $5 million. Rising plant counts and shrinking owner counts are not contradictory: that is what an industry looks like when capacity grows while the buyers get bigger. Pipe-and-tube incumbents grow by acquisition, though antitrust bites — a Department of Justice inquiry ended a proposed seamless-pipe combination.[3]
• 3313 — bought by strategics and PE, not public markets, and the scarce input is now power. Rio Tinto took 50% of the Matalco recycling network for $700 million and Emirates Global Aluminium 80% of Spectro Alloys; Apollo-managed funds acquired Arconic at an announced enterprise value of about $5.2 billion and American Industrial Partners took the mill now called Commonwealth Rolled Products; PE platforms are rolling up regional extruders. Upstream, consolidation is effectively blocked by antitrust and the tiny asset count, so the action is exits and restarts — and the most telling deal of the pass was a departure: Century's idled Hawesville smelter sold in February 2026 for $200 million plus a 6.8% interest in the buyer's planned data-center project, and will not return as a smelter. Meanwhile roughly 1.25 million tonnes of new sheet capacity (Steel Dynamics' 650,000-tonne Mississippi mill, Novelis' 600,000-tonne Bay Minette project) must find customers and pass qualifications.[4]
• 3314 — closed at the top, consolidating below, thinning by attrition at the bottom. Primary smelting is effectively shut to new entry (multibillion-dollar replacement cost, long permitting, power, acid markets, secure concentrate feed), so capacity comes from restarts and transfers — Korea Zinc bought the sole U.S. primary zinc smelter in April 2026. Copper fabrication consolidates through deals that each removed a U.S. listing (Wieland/Global Brass 2019, Prysmian/Encore Wire 2024 at ~$4.2 billion, International Wire/Hussey Copper 2025) yet stays only moderately concentrated, because the industry is concentrated within product families and fragmented across them. The recovery tail consolidates substantially by attrition: the operating secondary-lead smelter count has fallen to 11 sites under six owners, five fewer than at the EPA's 2012 rulemaking, as legacy environmental liabilities cap what buyers will pay.[5]
• 3315 — rolled up by PE, still fragmented, and shrinking. A decade of private-equity roll-ups (Waupaca to Monomoy, Grede to Gamut, Signicast, Pace, Gibbs) has not concentrated foundries in aggregate. The most consolidated corner is precision aerospace casting, and the change there is now measurable: the steel-investment grandchild went from 130 companies with a top-four share of 59.2% in 2002 to 84 firms with a top-four share of 73.5% today. The antitrust ceiling is real — in a 2021 municipal-castings deal the Department of Justice found the parties were two of only three significant suppliers across 11 states and forced divestiture of rights to more than 500 patterns. Elsewhere the child is shrinking by attrition: steel-except-investment fell from 248 establishments and 16,831 employees in 2002 to 141 and 9,928.[6]

The consistent pattern: buyers pay for qualified, permitted, well-located capacity with secure feed and customer approvals — not merely cheap furnaces. National scale alone is rarely the moat; patterns, permits, qualifications, competitive long-term power, and freight-advantaged location are.

---

9. Risks

The risk set is shared across the subsector and compounds at the aggregate level:

• Deep cyclicality — construction, autos, machinery, energy, and aerospace can turn down, sometimes together, and high fixed costs punish underutilization. Primary-metals capacity utilization was already 66.7% in 2025, below its 2022 level.[10]
• Metal-price and inventory-timing swings — inputs move before customer prices reset, whipsawing both revenue and, via inventory holding gains and losses, margins and working capital. Distributor destocking amplifies the swing, particularly in steel, where a quarter of shipments pass through intermediaries.[2][3]
• Input and energy cost volatility — scrap, ore, coal, cathode, nickel, cobalt, and electricity can move faster than customer pricing; high U.S. power costs keep primary aluminum and copper smelting structurally uncompetitive, and power is now actively contested by data centers willing to pay more for the same grid connection.[4]
• Trade-policy whipsaw — tariffs protect output but inflate inputs for the reshaping and casting children, importers have historically absorbed most of the duty,[9] protection is not granted uniformly across metals or products,[4] and the regime is changeable by executive action, exemption, or litigation — the copper framework was rewritten twice within three months of 2026.[5]
• Global overcapacity — the OECD puts global excess steelmaking capacity at 640 million tonnes in 2025, heading toward 745 million by 2028; the aluminum equivalent is roughly 1.25 million tonnes of new domestic sheet supply that must find customers during ramp-up.[8][4]
• Decarbonization and melt-modernization capital expenditure — a looming multi-year demand on the primary producers, and on foundries replacing cupolas with induction furnaces (one project runs $285 million through 2029).[2][6]
• Structural contraction in the employment-heaviest child — foundry employment is projected to fall 14.7% between 2024 and 2034, with real output declining as well; the near-term casting upcycle is more than half pricing.[6][11]
• Substitution — aluminum, plastics, and composites against steel and copper; PEX in plumbing and fiber in telecommunications against copper; additive manufacturing against some precision castings (though printed sand molds strengthen the foundries that adopt them); EV adoption redistributing rather than lifting casting demand.[5][6]
• Asset, customer, and single-point concentration — a handful of substitutable plants in several children (two primary copper smelters, one operating alumina refinery, 11 secondary-lead smelters), plus real customer concentration at individual operators — one precision caster's two largest customers were 38% of 2025 revenue, and a 2025 fire at one plant affected more than 700 sole-sourced aerospace parts.[4][5][6]
• Labor and safety — union negotiations, an aging skilled-trades base, and recordable injury rates that run above the private-industry benchmark in every child that publishes one.[2][3][5][6]
• Environmental and legacy liability — inherited contamination at older smelting and foundry sites can outlast the asset and turn a cheap private acquisition expensive.
• Investability and data risk — outside the steel-mill core, limited pure-play public options; the private tail is illiquid and cyclical; every listed proxy bundles this activity with something larger; and, as Section 3 shows, no reliable current revenue or concentration data exists below the five children to underwrite against.[3][4][5][6]

---

10. How to invest & outlook

Public-market investors should choose which metal and which rung of the chain they want, because there is no single vehicle for the subsector. The cleanest direct exposure is the iron-and-steel-mill core — listed producers Nucor (NUE), Steel Dynamics (STLD), and Cleveland-Cliffs (CLF), or a steel-industry ETF for a diversified basket — and there the first decision is production route rather than ticker, since the electric-arc and integrated models behaved very differently in 2025.[2] Beyond that, exposure is partial and diluted: Kaiser (KALU), Constellium (CSTM), Tredegar (TG), and Alcoa (AA) or Century (CENX) for aluminum; Mueller Industries (MLI), Materion (MTRN), ATI (ATI), and Carpenter (CRS) for copper and specialty metals; Howmet (HWM) and the newly listed Doncasters/DPC (DPC) for aerospace-precision casting; and small pure-plays like Insteel (IIIN) for steel-wire products. Treat all of these as cyclical industrials tied to a metal spread, not income or growth compounders — there is no dividend-yield or utility-style thesis here (Howmet's precision premium is the exception, and at a price-to-earnings ratio near 63 and a ~0.2% dividend yield it already prices in much optimism, even against guidance for roughly $9.1 billion of 2026 revenue).[6] Value them on mid-cycle earnings and normalized conversion earnings — enterprise value to earnings before interest, taxes, depreciation, and amortization (EV/EBITDA), free-cash-flow yield, leverage, and replacement cost per ton — not on peak-spread quarters and not on metal-inflated headline sales.

Private-market investors get the larger and often cleaner opportunity set, because the biggest assets in four of the five children are private. The clearest theses are the roll-ups and carve-outs in the fragmented children — regional coil processors, toll operations, and sub-scale wire, mesh, and reinforcement plants in 3312 (now best framed as plant rationalization at modest ticket sizes, since the headline corporate M&A has moved to rolled shapes and service centers); regional extruders and recyclers in 3313, where upstream the private menu has narrowed to three named situations rather than a market; family copper mills and permitted brownfield recovery plants in 3314; and family-, employee-, and PE-owned foundries in 3315.[3][4][5][6] Diligence centers on the same items everywhere: metal-purchase pricing formulas and escalator mechanics, customer contracts and product qualifications (and whether those qualifications transfer on a change of control), utilization by line or furnace size, normalized yield and scrap, direct tons versus toll tons, inventory accounting and working-capital lags, maintenance capital (frequently higher than book depreciation implies), permits and environmental history, skilled-labor retention, and freight radius. Separate a commodity inventory windfall from sustainable processing margin, and discount backlog that lacks firm quantities.

Outlook — selective, and structurally supported at both ends. The demand backdrop is favorable for reasons that cut across the subsector: reshoring, domestic-content rules, electrification and data-center buildout, and a durable preference for domestic strategic-metal supply, plus 50% Section 232 rates backstopping the primary producers.[12][13][14][15] But the protection is a transfer from the reshaping children to the primary ones, not free money,[9] and the five children point in different directions — steel mills consolidating against 640 million tonnes of global excess capacity, with electric-arc production already at 70% of domestic output;[7][8] steel products splitting into a pipe-and-tube volume story, a reshoring-and-electrification rolled-shape story, and a wire-drawing consolidation story with exactly one pure-play; aluminum pulling four ways at once, with recycling rising, extrusion falling 3.1%, sheet absorbing a new-supply wave, and primary a reshoring-policy wildcard; copper and specialty metals riding electrification, data centers, and aerospace/defense, while merchant smelting is squeezed by treatment charges that collapsed toward zero; foundries led by a strongest-tailwind precision corner over a cyclical die-casting bulk, inside a base projected to shed 14.7% of its jobs by 2034.[2][3][4][5][6][11] The through-line is the same in every child: the winners are low-cost, well-located, qualified operators that pass metal costs through quickly, keep their plants full, and sell where service and certification beat the lowest quote — not simply the largest nameplate capacity, and rarely the greenfield ambition over the permitted, qualified existing asset. Read as one idea, "primary metals" is a leveraged, late-cycle bet on the price of metal and the pace of construction; read as five, it is a menu — and the aggregate HHI of 234.7 should never be mistaken for a single level playing field.

For the full company-by-company detail, economics, and diligence checklists, read the five child primers: 3311, 3312, 3313, 3314, and 3315.

---

Sources

Level figures (receipts, establishments, employment, payroll, firm count, and concentration) are our own ingested federal ground truth for NAICS 331 [1]. Per-child size, concentration, ownership, margin, and company facts are synthesized from the five child primers [2]–[6], which in turn cite the primary federal and company sources. Level-spanning primary sources that no single child owns are listed separately at [7]–[15]. Tickers and multiples are provided only in the investable-universe and how-to-invest sections, per house style. Two confidence caveats carry up from the children: no child could establish a current concentration ratio (all are 2022 Economic Census readings), and federal revenue data largely disappears below the four-digit children.

1. Histometrics ground-truth federal statistics, NAICS 331 — U.S. Census Bureau 2023 County Business Patterns (establishments 3,570; employment 354,914; annual payroll $29.70B; Q1 payroll $7.84B) and 2022 Economic Census: Concentration and Firm Size Statistics (receipts $323.1B; firms 2,387; CR4 26.4%, CR8 34.7%, CR20 49.2%, CR50 64.6%; HHI 234.7). https://www.census.gov/programs-surveys/cbp.html
2. Histometrics child primer, NAICS 3311 — Iron and Steel Mills and Ferroalloy Manufacturing (receipts $129.8B; CR4 59.5%; CR50 99.7%; HHI 1,071.9; 73 firms; 359 plants; 78,532 jobs; payroll $9.1B and ~$116,000 per worker; 2025 steel shipment mix; metallics 55%–65% of mill cost; Nucor utilization 76%→83%; Cliffs −5% steelmaking gross margin vs. Steel Dynamics 10.6% operating margin; >90% union-represented hourly workforce; ~86,000 alternative employment reading). /primers-preview/3311
3. Histometrics child primer, NAICS 3312 — Steel Product Manufacturing from Purchased Steel (receipts $45.998B; CR4 21.8%; CR50 69%; HHI 196; 457 firms; 736 plants; 55,836 jobs; payroll $4.493B; Worthington 7.0% adjusted EBITDA margin; Insteel 14.4% gross margin and $37.5M rod-cost sensitivity; U.S. Steel tubular 40%→14%; NWPX 66.7% and Friedman 9% segment shares; 2026 Worthington/Klöckner and Ryerson/Olympic deals; wire plants 252→198 and rolled shapes 204→244; 72.2% iron-and-steel-products utilization). /primers-preview/3312
4. Histometrics child primer, NAICS 3313 — Alumina and Aluminum Production and Processing (receipts $51.142B; CR4 38.3%; CR50 85.4%; HHI 500.4; 257 firms; 431 plants; 57,358 jobs; payroll $4.369B; Kaiser $3.37B sales vs. $1.45B conversion revenue; Century power and Midwest-premium figures; extrusion demand −3.1% and utilization 83.5%→67.7%; recycling output index 93.0→108.6; ~1.25Mt new sheet capacity; Hawesville sale; 2025 aluminum end-market shares; extrusion trade case ended without orders). /primers-preview/3313
5. Histometrics child primer, NAICS 3314 — Nonferrous Metal (except Aluminum) Production and Processing (receipts $63.353B; CR4 22.6%; CR50 76.3%; HHI 235.8; 465 firms; 709 plants; 55,915 jobs; payroll $4.505B; child receipts $12.29B/$35.87B/$15.19B; 2025 copper end-market shares; Freeport ~70% of U.S. refined copper; Mueller Industrial Metals $1.024B/16.0%; Carpenter SAO 23.0%; 2025 TC/RC benchmark; 11 secondary-lead smelters under six owners; ASM vs. Economic Census receipts gap). /primers-preview/3314
6. Histometrics child primer, NAICS 3315 — Foundries (receipts $32.76B; CR4 20.2%; CR50 63.7%; HHI 170.4; 1,211 firms; 1,335 plants; 107,273 jobs; payroll $7.22B and ~$67,000 per worker; ferrous/nonferrous split; steel-investment top four 73.5% vs. 59.2% in 2002; Howmet ~$8.3B and Engine Products 33.3%; Doncasters June 2026 IPO at $33 and $837M revenue; Air T 20.1% stake; margin ladder and injury rates; AFS 4.2% 2026 survey). /primers-preview/3315
7. U.S. Department of Energy, Iron and Steel Manufacturing, 2026 — electric-arc furnaces produce 70% of domestic steel; scrap remelting uses less than half the energy of the ore route. https://www.energy.gov/cmei/ito/iron-and-steel-manufacturing
8. Organisation for Economic Co-operation and Development, OECD Steel Outlook 2026, 2026 — global excess steelmaking capacity 640 Mt in 2025, projected 745 Mt by 2028. https://www.oecd.org/en/publications/2026/06/oecd-steel-outlook-2026_a79fb861/full-report/international-efforts-to-address-the-steel-crisis-are-intensifying_b6ca4947.html
9. U.S. International Trade Commission, Economic Impact of Section 232 and 301 Tariff Actions, 2023 — affected imports −24%, U.S. steel-product prices +2.4%, domestic production +1.9% over 2018–2021, with higher downstream prices and lower downstream production. https://www.usitc.gov/press_room/news_release/2023/er0315_63679.htm
10. Board of Governors of the Federal Reserve System / Federal Reserve Bank of St. Louis, Capacity Utilization: Manufacturing — Primary Metal (NAICS 331), 2026 — 66.7% in 2025, 67.0% in 2024, 73.0% in 2022. https://fred.stlouisfed.org/series/CAPUTLG331A
11. U.S. Bureau of Labor Statistics, Industry Employment and Output Projections — NAICS 3315 employment 106,000 (2024) to 90,400 (2034), −14.7%; real output $26.9B to $24.8B. https://www.bls.gov/emp/tables/industry-employment-and-output.htm
12. White House, Adjusting Imports of Aluminum and Steel into the United States, 2025 — Section 232 rate of 50% effective June 4, 2025. https://www.whitehouse.gov/presidential-actions/2025/06/adjusting-imports-of-aluminum-and-steel-into-the-united-states/
13. U.S. Department of Commerce, Section 232 Tariff Summary, 2026 — 50% on steel articles and 25% on qualifying derivative articles as of April 2026. https://www.trade.gov/press-release/what-they-are-saying-president-trump-strengthens-us-steel-aluminum-and-copper
14. White House, Strengthening Actions Taken to Adjust Imports of Aluminum, Steel, and Copper into the United States, April 2026, and Further Adjusting the Tariff Regimes for Imports of Aluminum, Steel, and Copper into the United States, June 2026. https://www.whitehouse.gov/presidential-actions/2026/04/strengthening-actions-taken-to-adjust-imports-of-aluminum-steel-and-copper-into-the-united-states/ · https://www.whitehouse.gov/presidential-actions/2026/06/further-adjusting-the-tariff-regimes-for-imports-of-aluminum-steel-and-copper-into-the-united-states/
15. White House, Further Strengthening Actions Taken to Adjust Imports of Aluminum into the United States, July 2026 — half-rate import allowance for approved investors in new or expanded U.S. primary aluminum capacity. https://www.whitehouse.gov/presidential-actions/2026/07/further-strengthening-actions-taken-to-adjust-imports-of-aluminum-into-the-united-states/
16. White House, Adjusting Imports of Copper into the United States, 2025 — 50% Section 232 tariff on covered semi-finished copper effective August 1, 2025. https://www.whitehouse.gov/presidential-actions/2025/07/adjusting-imports-of-copper-into-the-united-states/
17. Nippon Steel Corporation, Completion of U.S. Steel Acquisition, 2025 — closed June 18, 2025. https://www.nipponsteel.com/en/newsroom/news/2025/20250618_100.html