U.S. Mining (except Oil and Gas) — An Investor's Primer (Subsector Level)

NAICS 2022 code 212 — Mining (except Oil and Gas) (United States)

A plain-language guide for public- and private-market investors. NAICS (the North American Industry Classification System) is the federal code system that defines industries. This is a rollup page for the three-digit subsector 212, which pools the three big "solid-minerals" mining industries — coal, metal ore, and nonmetallic minerals. The value of reading it at this level is the contrast across those three: which is biggest, what they dig, which way price and demand are pointing, how their costs work, who owns the mines, and how you would buy in. Business figures (firms, revenue, jobs, concentration) are U.S. Census Bureau ground truth for this exact level. Physical production and reserves come from the U.S. Energy Information Administration (EIA, the federal energy-statistics agency, for coal) and the U.S. Geological Survey (USGS, the federal nonfuel-minerals agency, for metals and nonmetallics), labeled by source. Figures carry numbered citations to the Sources list; where an official figure is absent at this level, that is stated, not filled in.

---

1. Overview

Mining (except Oil and Gas), NAICS 212, is the part of the U.S. economy that digs solid minerals out of the ground and does the first-stage processing — crushing, washing, concentrating (beneficiation) — that turns rock into a shippable product. It stops at the mine gate. The name's parenthetical is the key to its boundary: it excludes liquid and gaseous hydrocarbons (crude oil and natural gas, which sit in the sibling subsector 211 Oil and Gas Extraction), and it excludes the mining-support services — contract drilling, exploration, site prep — that sit in 213. Coal is in 212 because coal is a solid; oil and gas are out. Downstream steps — smelting and refining metal, burning limestone into cement, converting phosphate rock into fertilizer or lithium brine into battery chemicals — are manufacturing and sit in other codes entirely [2].

In one federal snapshot, 212 booked about $98 billion of receipts across 2,872 firms and 6,121 mine sites, employing roughly 174,000 people [1]. But "Mining (except Oil and Gas)" is not one business — it is three genuinely different commodity businesses stapled together by the classification system, and they behave very differently:

• Coal (2121) — a single, structurally shrinking energy commodity; a price-taker split between domestic electricity and steel-making exports.
• Metal Ore (2122) — four metals in one (iron, gold/silver, copper/base, and critical minerals), all price-takers, but each riding a different cycle — gold near record highs, copper on the electrification growth story, critical minerals a national-security policy story.
• Nonmetallic Mineral Mining (2123) — the tonnage giant of American mining, itself split between local price-makers (crushed stone, sand and gravel — the freight-protected backbone of construction) and global price-takers (phosphate, potash, soda ash, lithium — the chemical and fertilizer minerals).

The single most important idea at this level — and the reason a rollup view earns its keep — is that 212 is where the usual rule of mining breaks. Coal and metals are pure price-takers: they sell undifferentiated material into globally quoted markets at prices they cannot set, so revenue and margin ride the commodity cycle. But roughly two-thirds of the nonmetallic child — the construction-aggregates business — is the exception in all of mining: because heavy, cheap rock cannot travel far before freight destroys its value, each quarry is a local price-maker with a gentle cycle. So under one three-digit code you own a declining energy commodity, a spread of globally-priced metals pointing in every direction, and the one corner of mining that behaves like a local toll road. Where they diverge — which cycle, how concentrated, who owns them, and how (or whether) you can invest — is the whole point of this page.

---

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

NAICS nests from broad sectors (2-digit — here Sector 21, Mining, Quarrying, and Oil and Gas Extraction) down through subsectors (3-digit, like 212), industry groups (4-digit), industries (5-digit), and national industries (6-digit). Subsector 212 has three industry-group children. The table below is the heart of this primer — it is where they genuinely contrast.

Share of group = 2022 Economic Census receipts and 2023 County Business Patterns employment; rows may not sum to 100% due to rounding. HHI = Herfindahl-Hirschman Index, a 0–10,000 national-concentration score where below 1,500 is "unconcentrated." CR4 = the share of revenue held by the four largest firms. Sources: [1][3][4][5].

The contrasts that matter most

1. Relative size — dollars vs. jobs vs. sites. By revenue the three are surprisingly even: nonmetallic ~37%, metals ~35%, coal ~28%. But jobs and mine sites tell a different story: nonmetallic is ~52% of employment and 86% of all mine sites (5,293 of 6,121) [1] — a swarm of small pits and quarries — while coal (~24% of jobs) and metals (~24% of jobs) are far more capital-intensive per worker. Pay per worker averages ~$94,000 across 212 — well above the U.S. private-sector average — but coal and metals run ~$106,000 (skilled heavy-equipment and underground crews) against nonmetallic's ~$82,000 [1][3][4][5].

2. The master split — price-taker vs. price-maker. This is the deepest divide in the subsector, and it does not run neatly along child lines. Coal, all of metals, and the chemical-mineral slice of nonmetallic are price-takers — their margin is the thin, swinging residual between a globally set price and a largely fixed cost base. But the ~70%-of-nonmetallic construction-aggregates business is a price-maker: a cheap, heavy ton of crushed stone cannot travel more than ~25–50 miles before freight eats its value, so each quarry is a local mini-monopoly whose prices rise even through recessions. When people say "mining is a price-taker business," 212 is the code where that generalization fails — and the exception is its single biggest slice by tonnage [4][5].

3. Demand is pointing three different ways at once. This is unusual, and it is the strongest argument for treating 212 as three industries rather than one. As of 2025–26: coal is in structural decline (its share of U.S. electricity fell from ~51% in 2001 to ~15% in 2024), even though it firmed tactically in 2025; metals point every direction (gold at record highs, copper structurally bullish on electrification, iron soft, lithium collapsed ~80% from its peak); and construction aggregates are stable-to-rising on infrastructure, reshoring, and data-center build-out with near-zero energy-transition risk [3][4][5][6][7]. The three children rarely peak or trough together — which is why the subsector's aggregate numbers hide more than they reveal.

4. Concentration inverts intuition. Each child is meaningfully concentrated in its own way — coal moderately (HHI 523), metals as a set of near-oligopolies (74–99% top-four shares inside each metal), and the chemical-mineral sub-markets tightly (a handful of phosphate, potash, and soda-ash firms each). Yet the group looks almost perfectly competitive: HHI 187.5, CR4 just 21.1% [1]. That is a composition effect, and it is a trap. The subsector pools three mining businesses that do not compete with each other — a coal producer is not bidding against a gold miner is not bidding against a limestone quarry — and the fragmented aggregates base (86% of all sites) drags the whole index down. Do not read 21.1% as a competitive market. Real market power lives one or two levels down, commodity by commodity and, for aggregates, town by town.

5. Ownership mix diverges sharply. Coal is a handful of small public producers plus private/tribal/foreign owners — none of them large-cap. Metals are dominated by global majors and, newly, by the U.S. government itself (the Treasury holds preferred equity in rare-earth producer MP Materials; the Department of Defense is a major holder). Nonmetallic is a mix of large-cap aggregates majors over a vast private and private-equity tail (a Martin Marietta management estimate puts roughly two-thirds of U.S. aggregate production in private hands) [5]. Foreign ownership is pervasive across all three.

Scope note — what is not in 212 (so you don't double-count): crude oil and natural gas (subsector 211); contract drilling, exploration, and mine-site prep ("mining support activities," subsector 213); and all downstream manufacturing — steel mills, copper/lead/zinc smelters and refineries, cement and lime kilns, fertilizer plants, lithium conversion, magnet and battery making (NAICS 331/325/327) [2].

---

3. How big it is

Two federal yardsticks measure two different universes. Both are correct; keep them separate — mixing them is the most common error in reading this subsector.

3a. The business — U.S. Census, our ground truth for 212

CBP (County Business Patterns) is the annual employer-business series (counts, employment, payroll — no revenue); the Economic Census is the five-year revenue and concentration benchmark.

The rollup reconciles almost perfectly — a useful data-quality check. Adding the three children:

Establishments and employment tie exactly; receipts and payroll tie to rounding. The only line that does not add is firms — the children list 2,882 but the parent shows 2,872, because a company operating in two of these industries is counted once at the subsector level but in each child. That ~10-firm de-duplication is a normal feature of Census counts, not an error [1][3][4][5].

On concentration, the headline number is misleading — read it carefully. The group HHI of 187.5 is lower than every child except nonmetallic [1]. That is not because the industry is genuinely competitive; it is because the subsector's firm count is 86% nonmetallic (the most fragmented child) and because the three commodity businesses don't compete. Genuine market power is a level down: coal's national tonnage is ~85% controlled by its top 20 companies; each metal is a near-oligopoly; and aggregates, though nationally fragmented, are locally concentrated (a town may have only two deliverable quarries). See §8.

3b. The physical commodity — EIA and USGS, a different lens

Our 212 ground truth is business statistics only. It contains no group-level physical tonnage, reserve total, unit value, per-ton cost, or Small Business Administration size standard — those are simply not defined at the three-digit level, and we do not invent them. For the physical picture we draw on EIA (coal) and USGS (metals and nonmetallics), by commodity, and label it as such. These figures cover different years than the Census dollars and count captive and non-employer output the business census misses — so do not add or average USGS/EIA physical values against Census receipts.

The physical picture is wildly inverse to the revenue picture:

• Coal (EIA): ~512.5 million short tons produced in 2024 (a short ton = 2,000 lb), the lowest since 1964, rebounding to ~533 million in 2025; ~90% of domestic use is burned for electricity, with ~100 million tons exported (much of it high-value met coal) [3][7].
• Metal ore (USGS): small tonnages at high unit value — ~1.0 million metric tons of copper, ~160 metric tons of gold (~5.1 million troy ounces), ~38 million tons of iron ore, plus rare earths, uranium, and lithium. At 2025 prices, gold-plus-silver (~$18 billion of mine value) actually edged ahead of copper-plus-base-metals (~$14 billion), reversing the 2022 revenue order — a direct consequence of every metal being a price-taker [4].
• Nonmetallic (USGS): the physical giant — roughly 2.4 billion tons of construction rock a year (crushed stone ~1.5 billion tons worth ~$27 billion, plus construction sand and gravel ~870 million tons), the largest extractive flow in the country by weight, yet modest in dollars because the material is a "penny-a-pound," freight-dominated product [5].

Total U.S. nonfuel mineral production (metals + nonmetallics, i.e. 2122 + 2123 but not coal, which is a fuel) was worth about $112 billion in 2025 [6]. On reserves, the honest picture differs by child: the U.S. is not short of coal (a ~468-billion-ton demonstrated reserve base, but only ~10.6 billion tons at producing mines — it is short of demand, not rock) [3]; USGS calls stone, sand, gravel, and clay resources "plentiful" and publishes no national reserve tonnage (the binding scarcity is permitted material near a market) [5]; and the strategic story is metals and critical minerals, where import reliance runs high (silver ~77%, zinc ~73%, uranium ~92%, potash ~92%) [4][5].

---

4. The investable universe

There is no single "Mining (except Oil and Gas)" stock, and the routes differ sharply by child. The names and tickers below are pointers, not valuations — market caps swing hard with the commodity cycle, so check live quotes. Company figures come from SEC filings and disclosures, not the federal ground truth. This section, and §10, are the only places this primer carries tickers.

Coal (2121) — a short list of small-cap, high-beta producers. Every listed name is a leveraged bet on the coal price; none is a large-cap. The public routes are producers — Peabody Energy (BTU, mostly Powder River Basin surface thermal), Core Natural Resources (CNR, the Jan-2025 Arch–CONSOL merger; met + thermal + export terminals), Warrior Met Coal (HCC, pure premium met), Alpha Metallurgical Resources (AMR, met), Alliance Resource Partners (ARLP, thermal-income partnership), and Ramaco Resources (METC, met) — plus one royalty owner (Natural Resource Partners, NRP) and a niche ETF (COAL). Major private/tribal/foreign owners: Navajo Transitional Energy (NTEC), American Consolidated Natural Resources (ACNR), Foresight Energy [3].

Metal Ore (2122) — the deepest menu, by child. Iron: the only clean U.S. listing is Cleveland-Cliffs (CLF, a leveraged steel-plus-captive-ore bet); Japan's Nippon Steel bought U.S. Steel (and its Minnesota ore) in 2025. Gold/silver: producers Newmont (NEM), Barrick (GOLD), Kinross, Coeur, Hecla, plus the Barrick/Newmont Nevada Gold Mines joint venture (~half of U.S. gold) and royalty/streaming financiers Franco-Nevada, Wheaton Precious Metals, Royal Gold. Copper/base: Freeport-McMoRan (FCX, the U.S. copper bellwether), Rio Tinto (Kennecott), Teck (Red Dog zinc). Critical minerals: near-monopoly names — MP Materials (MP, rare earths, with the U.S. Treasury as a preferred holder and the DoD as a price-floor buyer), Energy Fuels and Cameco (uranium), Lithium Americas and Albemarle (lithium) [4].

Nonmetallic (2123) — quality compounders plus cyclical producers. Aggregates (the primary vehicle): Vulcan Materials (VMC) and Martin Marietta (MLM) are the liquid large-cap pure-plays — local-pricing-power reserve annuities — with CRH (CRH), Amrize (AMRZ), and Knife River (KNF) adding diversified building-materials exposure. Frac sand (high-beta energy proxy): Atlas Energy Solutions (AESI), Smart Sand (SND). Clay: Minerals Technologies (MTX). Chemical/fertilizer minerals (cyclical price-takers): Nutrien (NTR, potash), Mosaic (MOS, phosphate), Intrepid Potash (IPI), Compass Minerals (CMP, salt), Albemarle (ALB, lithium). Much of the biggest tonnage — salt, soda ash, and roughly two-thirds of aggregates — is private/PE/foreign (Quikrete, Cargill, WE Soda, J.R. Simplot) [5].

The royalty/streaming layer — deep in metals, thin everywhere else. Precious-metals investors have a mature menu of royalty/streaming companies (Franco-Nevada, Wheaton, Royal Gold) that own a slice of a mine's revenue without operating risk — historically the best risk-adjusted way to own mining cash flow. Coal and nonmetallic have no equivalent large pure-play (nearest listed proxies: NRP and Mesabi Trust in coal/iron; FRP Holdings, LandBridge in aggregates/frac sand) [3][4][5][14]. Across all three children, the classic private-capital angle is mineral-and-royalty ownership — owning the ground and collecting a per-ton or percentage royalty as material is extracted.

---

5. How the money works (and where the children diverge)

The economics rhyme across 212 — every ore body or coal seam is a wasting asset that depletes and must be replaced; a federal percentage-depletion allowance shelters some cash flow from tax; and returns come from volume × per-unit margin plus, for metals, byproduct credits. And it is worth stating what this is not: none of 212 is a regulated utility or a REIT. There is no rate base, no allowed return, and no funds-from-operations (FFO) story here. The right lens throughout is reserves, per-unit margin, cost-curve position, royalties, and depletion — commodity economics, not utility or real-estate math. The differences are where the money is made:

• What sets the price — and whether the producer has any. Coal is a price-taker on thermal benchmarks and volatile seaborne met markets [3]. Metals price off global exchanges — gold/silver on London and COMEX (the New York futures exchange), copper/zinc/lead/nickel on the London Metal Exchange (LME), iron ore off China-delivered fines — while critical minerals trade in thin, opaque markets where China is often the marginal price-setter [4]. Chemical minerals (phosphate, potash, soda ash, lithium) are global price-takers too [5]. The exception is construction aggregates, the one part of 212 with genuine pricing power: because freight dominates, the low-cost supplier is simply the one closest to the job, and prices compound through the cycle — Vulcan's 2025 aggregates ran roughly a $21.98/ton price against ~$10.65/ton cash cost, ~$11.33/ton cash gross profit (~51% cash margin), extraordinary for anything labeled "mining" [5].

• The cost yardstick differs. Metals use all-in sustaining cost (AISC — cash cost plus sustaining capital and royalties per unit), the closest thing to a standard [4]. Coal has no single standard cost metric (reported "cash cost per ton" excludes different items), and freight is often decisive for low-value Powder River Basin coal, where rail can cost as much as the coal itself [3]. Aggregates have no national break-even at all — the useful gauge is the freight-adjusted price-minus-cash-cost spread per ton [5]. Note the deliberate contrast with the sibling oil-and-gas subsector, which lives on lifting cost and finding-and-development (F&D) cost — different metrics for a different (fluid) resource.

• Royalties split by federal-land regime (see §7). Hardrock metals on federal land pay no federal production royalty; coal and the leasable chemical minerals pay one; and common-variety aggregates pay none but are gated by local zoning [8][9][10].

• Reserve replacement urgency differs. For coal and aggregates the scarce input is permission, not geology — operators harvest existing permitted leases rather than explore, and reserve lives run decades [3][5]. Metals are the opposite: multi-billion-dollar capex and ~29-year discovery-to-production timelines mean growth is bought, not found, and reserve replacement is a live pressure [4].

The upshot in a downturn: a coal or chemical-mineral producer sees thin margins whipsaw hard; a gold miner rides a metal that often rises when the economy wobbles; a copper miner leans on byproduct credits and a structural demand story; and an aggregates quarry keeps pushing price even as volumes soften. Same subsector, four different survival mechanics.

---

6. Demand drivers

Three largely unrelated demand universes under one code:

• Coal → electricity and steel. ~90% of U.S. coal use is power generation, a market in structural decline (coal's electricity share fell from ~51% in 2001 to ~15% in 2024, firming to ~17% in the colder, higher-gas 2025) as cheap shale natural gas and renewables displace it [3][7]. The other leg is met coal for blast-furnace steelmaking, which survives by exporting into a roughly flat global steel market because U.S. steel is now mostly scrap-fed. Growing electricity demand (data centers, electrification) does not translate one-for-one into coal demand [3].

• Metals → four stories. Iron tracks U.S. steel output (pressured by the shift from blast furnaces to scrap-fed electric-arc furnaces). Gold is a monetary metal driven by central-bank buying and safe-haven flows — often counter-cyclical, which is why it is at record highs while iron is soft. Copper is the electrification growth engine (grids, motors, EVs, data centers; the International Energy Agency projects ~30% demand growth by 2040). Critical minerals ride the security-and-transition theme — rare-earth magnets, battery lithium, reviving nuclear — with de-risking supply chains away from China as the cross-cutting driver of new government demand [4][15].

• Nonmetallic → construction, farming, and a little battery. Roughly 70% of the child is construction aggregates, whose master driver is U.S. building activity — the most durable leg being federal infrastructure (the 2021 Infrastructure Investment and Jobs Act, IIJA, whose post-2026 reauthorization is the single biggest demand-swing variable), plus reshoring, data centers, and grid build-out [5]. The chemical-mineral slice serves agriculture (phosphate and potash fertilizer — no substitute for the nutrients), de-icing (salt), glass and chemicals (soda ash), and batteries (lithium and lithium-iron-phosphate chemistry). Frac sand alone is an energy bet, ~81% tied to shale completions [5].

Notably, most of 212 is not a critical-minerals or EV story — the thesis for the large majority of receipts is electricity, steel, construction, and agriculture. The genuine green-transition winners (lithium, battery-grade phosphate, copper) and the clear loser (thermal coal, and to a lesser extent frac sand) are specific corners, not the whole.

---

7. Regulation

The regulatory stack shares a spine — every mine, pit, and quarry in 212 is a workplace under the Mine Safety and Health Administration (MSHA), whose April 2024 respirable-crystalline-silica rule (halving the permissible limit) is the shared active item, currently delayed in the metal/nonmetal sector after a court stay [3][4][5][12]. Environmental permitting runs through the Environmental Protection Agency (EPA), the National Environmental Policy Act (NEPA) review, and Clean Water Act (CWA) discharge rules across all three [4][5].

But the sharp divide is land tenure — and 212 is unusual in containing all three federal-mineral regimes at once:

• Leasable minerals (Mineral Leasing Act of 1920). Coal is leasable — the U.S. keeps title and leases development rights through the Bureau of Land Management (BLM), owning most Powder River Basin coal in the ground, and collects a production royalty (a 2025 law capped it at ≤7% through Sept. 30, 2034 and directed new leasing) [3][9][11]. So are the chemical minerals phosphate, potash, and sodium (salt/soda ash/trona), where the BLM collects a ~2–5% royalty — making the U.S. Treasury effectively a royalty owner on that output [5][9].
• Locatable minerals (General Mining Law of 1872). Hardrock metals — gold, silver, copper, and most critical minerals including lithium — are claimed on federal public-domain land and pay no federal production royalty, a real U.S. cost advantage [4][8]. (Iron is the metals exception: it sits on state/fee land in Minnesota and pays a taconite production tax [4].)
• Saleable minerals (Materials Act of 1947). Common-variety aggregates (stone, sand, gravel, clay) on federal land are sold at fair market value, not claimed — but the binding constraint is really local zoning and conditional-use permitting, the aggregates business's chief chokepoint (what blocks new entrants protects incumbents' pricing) [5][10].

Two more forces cut across: downstream rules (EPA power-plant standards decide whether a thermal miner's customer keeps burning coal — often mattering more than mine-site rules) [3]; and trade and industrial policy, now first-order for metals and critical minerals — a 2025 Section 232 copper tariff, Inflation Reduction Act (IRA) §45X manufacturing credits, and Department of Defense (DoD) and Department of Energy (DOE) capital that put a partial floor under favored critical-mineral producers [4][15]. The investor implication is consistent: model policy support contract-by-contract; do not capitalize an indefinite subsidy or an unreversed royalty regime.

---

8. Consolidation

The strategic motion is consolidation in all three children — but for different reasons, and the through-line is that buyers grow by acquiring permitted reserves because greenfield permitting is so hard:

• Coal was forged in the 2015–16 bankruptcy wave; survivors emerged deleveraged and export-focused, most recently the Jan-2025 Core Natural Resources (Arch–CONSOL) merger. Antitrust caps further combination within the Powder River Basin (the FTC blocked a Peabody–Arch tie-up in 2019), so producers now compete on cost and return cash rather than grow [3].
• Metals consolidate by depletion-driven M&A (mines run down, so growth is bought — Newmont-Newcrest; the Nevada Gold Mines JV), by "copper is the prize" mega-bids (BHP's failed ~$49 bn run at Anglo American), and by down-cycle distress plus vertical integration in critical minerals — with government offtake itself becoming a competitive edge [4].
• Nonmetallic shows a concentrated top rolling up a fragmented tail: Quikrete's ~$11.5 bn take-private of Summit Materials (2025), pending CRH–Arcosa (~$8.5 bn) and Martin Marietta–Lhoist (~$13.5 bn) in aggregates; distress consolidation in frac sand; and chemical minerals passing into foreign and PE hands (WE Soda's ~$1.4 bn Genesis Alkali deal) [5].

The apparent contradiction — "2,872 firms, HHI 188" yet real pricing power — resolves the same way in every child: national fragmentation masks local concentration (aggregates: a town with two quarries), commodity-market concentration (five phosphate firms; each metal an oligopoly), or tonnage concentration (coal's top-20 move ~85% of output). Antitrust review (the Federal Trade Commission, FTC, and Department of Justice, DOJ) accordingly focuses on local overlap for aggregates and commodity-market overlap for metals and chemicals. Barriers to entry — capex, permitting, ~29-year metal timelines — are extreme, which is why a handful of incumbents dominate each corner [4][5].

---

9. Risks

1. Commodity-price (and volume) cyclicality — the central risk, in three flavors, and getting the flavor right is the whole underwriting task. These are commodity businesses whose equities swing more than the underlying material. Coal: thin, variable per-ton margins mean small price moves swing earnings, dividends, and share prices hard [3]. Metals: price-takers levered to volatile global prices — recent moves span iron's realized price falling ~$156→~$89/t, lithium collapsing ~80%+ and rare earths >50% from 2022 peaks, while gold ran to records [4]. Nonmetallic: aggregates are the gentler cycle in all of mining (local price stays sticky-to-rising; the cycle shows up in volumes), but frac sand is the sharpest (price hostage to shale completions and oversupply — the Hi-Crush/Covia bankruptcies are the cautionary tale) and chemical minerals swing violently with world prices [5]. The unifying lesson: peak-cycle earnings are not permanent, and mining dividends and buybacks are variable distributions, not bond-like income. You get diversification across these cycles only by owning different children — not by owning "the subsector."
2. Secular demand / stranded-asset risk — concentrated, not universal. Thermal coal is the acute case (reserves and rail spurs can be stranded if the customer plant retires first); frac sand carries a milder fossil-fuel overhang. Construction aggregates, clay, copper, and fertilizer minerals face essentially none — they are needed under every decarbonization scenario [3][5].
3. Cost inflation and execution — diesel, power, labor, explosives, reagents, and steel often rise in the same boom that lifts prices [4][5].
4. Permitting, zoning, litigation, and social license — multi-decade metal timelines, local aggregates zoning fights, and project-specific blocks can strand capital [4][5].
5. Depletion and reserve replacement — every mine is a wasting asset; high near-term cash flow can just be the harvest of a finite resource [3][4][5].
6. Regulatory reversal, reclamation/bonding, and long-tail liabilities — two-directional policy risk (the coal royalty cap and §45X credits can be repealed) plus obligations that outlast production (mine reclamation, black-lung, water treatment, phosphogypsum stacks) [3][5][11].
7. Ownership, foreign dependence, and geopolitics — pervasive foreign control (Nippon in iron; Grupo México, Trafigura, Glencore in base metals; WE Soda in soda ash) and structural U.S. reliance on foreign processing for most refined metals and on imports for potash, uranium, and more [4][5].
8. Policy dependence, concentrated in critical minerals — the critical-mineral bull case increasingly rests on subsidies, price floors, and offtake that politics can remove [4].

---

10. How to invest, and the outlook

There is no one-ticket way to own "Mining (except Oil and Gas)," and you probably shouldn't want one — the three children are different bets pointing different directions. Match the route to the thesis:

• Coal — a tactical, cash-harvesting trade. Small-cap producer equities (met: HCC, AMR, METC; diversified/export: CNR; thermal income: ARLP; largest surface: BTU) plus one royalty (NRP) and a niche ETF (COAL). Expect a boom-bust dividend-and-buyback pattern priced as "melting ice cubes" — low earnings multiples, high free-cash-flow yields. This is a capital-return story, not a growth story [3].
• Metals — the deepest and most varied menu. Producer equities give leveraged exposure (NEM/GOLD for gold, FCX for copper, MP/Energy Fuels/Cameco/Lithium Americas for critical minerals); royalty and streaming companies (Franco-Nevada, Wheaton, Royal Gold) own margin without capital calls — historically the best risk-adjusted mining exposure; thematic and physical-metal ETFs are the cleanest one-ticket route to gold, copper, uranium, lithium, or rare-earth themes [4].
• Nonmetallic — compounders, not high-yielders. Aggregates majors (VMC, MLM, plus CRH/Amrize/Knife River) are quality cyclicals — local-pricing-power reserve annuities with steadily growing dividends and premium multiples, not boom-bust payers, because the product price does not crash. Frac sand (AESI) is cyclical energy-beta; chemical/fertilizer minerals (NTR, MOS, IPI, CMP, ALB) are cyclical price-takers to size against mid-cycle, not peak, free cash flow. There is no pure-play nonmetallic ETF or royalty stock — this is a stock-selection and private-diligence corner [5].
• Private-market routes — the only way into much of 212. Direct or private-equity ownership of coal mines, single-asset metal developers, and the vast fragmented aggregates tail (local-monopoly cash flows at lower entry multiples than the majors); government-adjacent structured capital in critical minerals (DOE loans, DoD/Defense Production Act preferred equity, transferable §45X credits); and the classic mineral-and-royalty ownership angle across all three children — the lowest-operational-risk exposure, but illiquid, appraisal-driven, and title-dependent. Underwrite mine-by-mine: reclamation, bonding, black-lung, water-treatment, permit, and end-market-life liabilities are the real risk [3][4][5][14].

Outlook (forward-looking judgment, not reported fact). The strongest single observation about 212 is that its three cycles are out of phase, which is the whole argument for viewing it as three industries, not one:

• Coal: more resilient near-term than a straight-line decline suggests (2025 production rebound, slowest plant-retirement pace since 2010, royalty cut, data-center power demand) — but the structural verdict is unchanged. Policy can extend the runway and fatten interim cash returns; it does not restore thermal coal's competitiveness against cheap gas and renewables. A high-cash-return, high-volatility, finite-life exposure [3][7][11].
• Metals: out of phase within themselves — gold/silver at a cyclical peak on record prices (record margins are the top of a cycle, not a baseline); copper structurally constructive but cyclical near-term; iron soft; critical minerals constructive but policy-dependent, with the sharpest risk being financing a high-cost project at a price peak and reaching production after the market turns [4][15].
• Nonmetallic: the ~70%-of-child construction-aggregates core is structurally intact — freight-protected local pricing, long reserve lives, permitting barriers that entrench incumbents, and infrastructure/reshoring/data-center tailwinds with near-zero transition risk; its real risks are cyclical (volume) and political (highway reauthorization), not demand destruction. Frac sand is soft on price, resilient on volume; chemical minerals split into a stable industrial base (salt, soda ash) and a volatile frontier (phosphate/LFP, potash, lithium) consolidating into foreign and PE hands [5].

Bottom line. NAICS 212 is a ~$98-billion, ~174,000-worker rollup of three mining businesses that share the mine-gate boundary and a wasting-asset, depletion-shielded economic frame — and little else. Nonmetallic is the tonnage and jobs giant and the home of mining's one true pricing-power exception; metals are the highest-value, most-varied, most policy-touched slice; coal is the shrinking energy commodity. The subsector's HHI of 187.5 makes it look competitive — an illusion of aggregation, because it pools three oligopolistic (or locally monopolistic) businesses that never bid against each other. Do not buy "the subsector." Buy the child — and within it the commodity, the cost-curve position, and the ownership structure — whose cycle you actually want to own.

---

Data-quality notes

• Business statistics (firms, establishments, receipts, payroll, concentration) are U.S. Census Bureau ground truth for NAICS 212: 2022 Economic Census (receipts $97.99 bn; 2,872 firms; CR4 21.1% / CR8 32.1% / CR20 48.7% / CR50 64.8%; HHI 187.5) and 2023 County Business Patterns (6,121 establishments; 173,729 employees; $16.28 bn payroll; $4.20 bn Q1 payroll) [1]. The two programs use different years and definitions; do not subtract across them to infer openings/closures.
• Rollup arithmetic reconciles. The three children's 2023 CBP establishments (491 + 337 + 5,293 = 6,121) and employment (42,347 + 41,280 + 90,102 = 173,729) sum exactly to the parent; 2022 receipts ($27.43 + $34.23 + $36.33 bn ≈ $98.0 bn) and payroll (~$16.28 bn) reconcile to rounding. Firm counts do not add (children 2,882; parent 2,872) because a firm operating in multiple children is de-duplicated (~10 firms) at the subsector level — a normal Census feature [1][3][4][5].
• Not stated because unavailable. Our 212 ground truth contains no group-level physical tonnage, reserve tonnage, unit value, per-unit cost/AISC-equivalent, or single SBA size standard — those are undefined at the three-digit level. Physical production, prices, and reserves are drawn per commodity from EIA (coal) and USGS Mineral Commodity Summaries 2026 (metals, nonmetallics) and are labeled as such; USGS/EIA physical values cover different years and a different (whole-commodity) universe than Census receipts, so do not add or average the two frames [1][3][4][5][6][7].
• Company figures, tickers, and ownership in §4 and §10 are from SEC filings and company disclosures, not the federal ground truth; market-cap, yield, and multiple references are order-of-magnitude anchors that move with the commodity cycle.

---

Sources

1. U.S. Census Bureau — 2022 Economic Census (EC2200BASIC, Summary Statistics & Concentration) and 2023 County Business Patterns, NAICS 212 (our ground truth): receipts $97.99 bn, 2,872 firms, 6,121 establishments, 173,729 employees, $16.28 bn payroll, CR4 21.1%/CR8 32.1%/CR20 48.7%/CR50 64.8%, HHI 187.5. https://data.census.gov/
2. U.S. Census Bureau — 2022 NAICS Definitions: Sector 21 and Subsector 212 (Mining except Oil and Gas), children 2121/2122/2123; sibling subsectors 211 (Oil and Gas Extraction) and 213 (Support Activities); downstream manufacturing 325/327/331. https://www.census.gov/naics/?input=212&year=2022
3. Histometrics child-industry primer — NAICS 2121 Coal Mining (synthesizing U.S. Census 2022 EC / 2023 CBP for 2121 and EIA Annual Coal Report 2024; production, met/thermal split, investable universe, leasable-mineral regime).
4. Histometrics child-industry primer — NAICS 2122 Metal Ore Mining (synthesizing U.S. Census 2022 EC / 2023 CBP for 2122 and USGS Mineral Commodity Summaries 2026; iron/gold/copper/critical-mineral detail, AISC economics, ownership and royalty layer).
5. Histometrics child-industry primer — NAICS 2123 Nonmetallic Mineral Mining and Quarrying (synthesizing U.S. Census 2022 EC / 2023 CBP for 2123 and USGS Mineral Commodity Summaries 2026; aggregates local-pricing economics, chemical-mineral price-taking, land-tenure split, investable universe).
6. U.S. Geological Survey — "Value of U.S. mineral production" (~$112 billion nonfuel mineral production, 2025), news release, 2026. https://www.usgs.gov/news/national-news-release/value-us-mineral-production-rose-last-year-driven-precious-metals-prices
7. U.S. Energy Information Administration — U.S. coal production and coal-fired generation increased in 2025 and Coal's share of U.S. electricity generation (~512.5→~533 MMst; ~51% 2001 → ~15% 2024 → ~17% 2025), 2026. https://www.eia.gov/coal/annual/
8. General Mining Act of 1872; Congressional Research Service, The General Mining Law of 1872: Issues and Legislation (2024) — locatable hardrock minerals (metals, lithium), no federal production royalty. https://www.congress.gov/crs-product/R48166
9. U.S. Bureau of Land Management — Nonenergy Leasable Minerals (Mineral Leasing Act of 1920); leasable coal, phosphate, potash, sodium; federal production royalties (43 CFR §3504.21). https://www.blm.gov/programs/energy-and-minerals/mining-and-minerals/nonenergy-leasable-materials
10. U.S. Bureau of Land Management — Saleable Minerals / Materials Act of 1947 (common-variety stone, sand, gravel, clay sold at fair market value, not locatable). https://www.blm.gov/programs/energy-and-minerals/mining-and-materials/saleable-minerals
11. 119th Congress — Public Law 119-21 — federal coal royalty ≤7% through Sept. 30, 2034 and expanded leasing; IRA §45X manufacturing-credit phase-down, 2025. https://www.congress.gov/119/plaws/publ21/PLAW-119publ21.htm
12. U.S. Mine Safety and Health Administration — Lowering Miners' Exposure to Respirable Crystalline Silica, Final Rule (PEL 50 µg/m³, Apr 2024; metal/nonmetal effective date delayed after court stay) and MSHA safety oversight of all mines. https://www.msha.gov/regulations/rulemaking/silica
13. Company disclosures / SEC filings (investable universe) — Peabody (BTU), Core Natural Resources (CNR), Warrior Met (HCC), Alliance (ARLP); Cleveland-Cliffs (CLF), Newmont (NEM), Barrick (GOLD), Freeport-McMoRan (FCX), MP Materials (MP); Vulcan Materials (VMC), Martin Marietta (MLM), CRH (CRH), Nutrien (NTR), Mosaic (MOS), Albemarle (ALB). SEC EDGAR, FY2025.
14. Royalty/streaming and government-capital layer — Franco-Nevada, Wheaton Precious Metals, Royal Gold (precious/base metals); Natural Resource Partners (NRP) and Mesabi Trust (MSB) (coal/iron); FRP Holdings (FRPH), LandBridge (LB) (aggregates/frac sand); U.S. Treasury preferred equity and DoD/DOE support in critical minerals. Company disclosures / SEC filings, FY2025.
15. International Energy Agency, Global Critical Minerals Outlook 2026 (copper demand growth, supply-shortfall risk); The White House, Section 232 copper tariff (2025); USGS 2025 List of Critical Minerals. https://www.iea.org/reports/global-critical-minerals-outlook-2026/executive-summary

---

Reference years: business statistics = 2022 Economic Census (receipts, firms, concentration) and 2023 County Business Patterns (establishments, employment, payroll) — U.S. Census Bureau, Histometrics' ingested ground truth for NAICS 212 ($97.99 bn receipts, 2,872 firms, 6,121 establishments, 173,729 employees, $16.28 bn payroll, HHI 187.5). Physical production/reserves = latest EIA (coal) and USGS (metals, nonmetallics) by commodity, noted inline. Policy = 2025–26. Subsector 212 contains three industry-group children (2121 Coal, 2122 Metal Ore, 2123 Nonmetallic Mineral Mining); establishment and employment figures sum exactly to the parent, receipts and payroll to rounding, and firm counts differ by ~10 due to multi-industry de-duplication. Forward-looking statements reflect cited EIA/USGS/IEA outlooks and editorial judgment, not guarantees of investor returns.