Other Nonmetallic Mineral Product Manufacturing (NAICS 3279): An Investor's Primer

NAICS (North American Industry Classification System, the standard code set the U.S. government uses to group businesses) 2022 code 3279 — United States. This is a NAICS industry group (the four-digit level), a rollup of two child industries: Abrasive Product Manufacturing (32791) and All Other Nonmetallic Mineral Product Manufacturing (32799). [1]

1. Overview

NAICS Sector 32 (manufacturing) has named homes for the big nonmetallic-mineral products — cement, ready-mix concrete, glass, lime, gypsum, clay and ceramics. NAICS 3279 is where everything else that starts with rock ends up. It bundles two very different residual bins: the abrasives business (grinding wheels, sandpaper, cut-off discs, diamond tools — the consumables that cut, grind and finish everything else that gets manufactured), and the sprawling "all other" nonmetallic bin (stone countertops, ground mineral powders for paint and plastics, mineral-wool insulation, fracking proppants, stucco and plaster) [2][3].

For an investor, the useful way to read 3279 is as two loosely related picks-and-shovels businesses filed under one code. They rhyme in four ways — both take rock or mineral grit and turn it into an industrial or building input; both sell consumable or commodity materials whose demand tracks how much cutting, building and drilling the wider economy is doing, not one-off equipment purchases; in both, the durable growth is now in value per unit rather than tonnage (engineered grain that lasts longer, specialty mineral grades that sell for multiples of bulk filler) [2][3]; and both are thinly represented in public markets, with no U.S.-listed company that is simultaneously a pure play and a leader in either child. Beyond that, size, growth direction, ownership mix and concentration diverge sharply — which is exactly what a rollup view exists to expose [2][3].

Why care: at about $31.3 billion of manufacturing receipts [1], 3279 is a mid-sized slice of industrial and building materials, but it sits at chokepoints in much larger flows — U.S. factory output, the countertop and insulation trades, the mineral-filler supply chain, and shale completions. The public menu is thin and mostly indirect; the private and private-equity (PE, investment firms that buy whole companies) menu is deep. Two hazards now run through both children: respirable crystalline silica dust, from grinding-wheel and blasting work to engineered-stone fabrication, and an active trade-remedy regime — as of 2025–2026 each child has its own antidumping and countervailing-duty (AD/CVD) orders in force [2][3][8][9].

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

The level's two children are wildly unequal in size and opposite in structure. One (32799) carries ~88% of the dollars across four unrelated sub-markets; the other (32791) is a single, tightly concentrated product line carrying the remaining ~12%. Read together, they show that "how concentrated is this industry?" has no answer at the four-digit level — you must name the child.

Table A — the numbers (who is how big, and how concentrated). Receipts, firm counts and concentration ratios are 2022 Economic Census (EC); establishments and employment are 2023 County Business Patterns (CBP). Shares are of the 3279 total. Level figures are our ground-truth federal statistics [1]; child figures come from the child primers [2][3].

CRn = combined revenue share of the largest n firms; HHI = Herfindahl-Hirschman Index, a standard market-concentration gauge (higher = more concentrated). The abrasives (32791) HHI is suppressed in federal data, so we state no value for it; its top-heavy CR4 of 50.5% signals a concentrated market all the same [2]. (Establishments, employment and payroll sum exactly across the two children — the children's payrolls of $785.3 million and ~$4.88 billion reconcile to the level's ~$5.66 billion; only the firm counts differ slightly, with the level tallied separately at 2,979 against the children's 242 and 2,745, as Census firm counts do at each level [1][2][3].) Note also that establishments are physical employer locations, not independent competitors — the firm count is the better proxy for how many businesses compete [2].

Table B — the strategic contrast (direction, ownership, how to invest).

Three structural readings jump out:

• Concentration is a mirage at this level — and abrasives is the reason. The level looks hyper-competitive (CR4 19.3%, HHI 159), lower than even its larger child 32799 (CR4 20.9%). That is backwards from intuition: adding a genuinely concentrated market (abrasives, where the top four firms take ~half the revenue) to the mix pushes measured concentration down. The CR50 ladder makes the mechanism plain — the fifty largest firms hold 59.1% of receipts at the level [1], but 87.4% inside abrasives [2] and 60.7% inside 32799, which in turn runs as high as 97.6% inside its mineral-wool child [3]. Measured concentration falls at every step up the tree because the leaders in one product market do not compete in the others; lumping them together dilutes every firm's measured share. Both children make this argument independently about their own aggregates, which is the strongest evidence that the level's low HHI is an aggregation artifact, not evidence of a competitive industry — real pricing power lives in the sub-markets (abrasives, and the mineral-wool oligopoly inside 32799 at HHI 1,106) [2][3].
• The two children run on overlapping but distinct cycles. Abrasives is a near-pure read on industrial production and metal fabrication, with kickers from autos, aerospace and (structurally) semiconductors. The 32799 bin adds housing and remodeling (countertops, stucco, insulation), building-energy and fire codes (mineral wool), industrial end-markets for mineral fillers (paint, plastics, paper, filtration), and an oil-and-gas drilling cycle (barite, proppants) that touches neither abrasives nor most of the rest [2][3].
• Growth in both children is in value per unit, not tonnage. In abrasives, engineered ceramic and shaped grains cut faster and last longer, lifting supplier revenue and customer productivity while reducing the number of discs and wheels consumed — so unit consumption is a poor measure of demand [2]. In minerals, the same rock spans an enormous price range by grade: diatomite runs from roughly $10 per metric ton as lightweight aggregate to more than $1,000 per metric ton in specialty applications [3][7]. The investable thesis in both children is mix, not volume.

3. How big it is (the rollup, and why the total understates the real economy)

Ground-truth federal figures for NAICS 3279 as a whole (our ingested statistics [1]):

Derived level-wide averages: roughly 25 workers per establishment, about $68,000 of payroll per worker, and about $10.5 million of receipts per firm [1]. Treat these as near-meaningless: they blend a top-heavy $14-million-per-plant abrasives industry with a 32799 bin that spans $3.7-million labor-heavy cut-stone shops and $35-million capital-heavy insulation plants (Section 2). Pay illustrates the same fiction — abrasives works out to roughly $71,500 a worker on the child's own payroll and headcount [2], sitting inside a 32799 range that runs from about $57,900 in cut stone to about $85,000 in ground minerals [3]. Any single "typical firm" in 3279 is a statistical artifact.

Undercount and boundary caveats — the $31.3 billion is narrower than it sounds.

1. The nonemployer undercount is entirely a 32799 phenomenon. Federal employer statistics omit self-employed businesses, which bites hard in the artisan tail of 32799 — thousands of one-person cut-stone fabricators, monument dealers, and plaster/statuary makers — inflating how much the firm count understates reality [3]. Abrasives, by contrast, is factory manufacturing with essentially no sole-proprietor tail, so the Census captures it cleanly; the child primer says so explicitly [2]. Even inside 32799, mineral wool and ground minerals are plant-scale and fully captured [3]. The undercount affects headcount and firm counts far more than the dollar total, which larger plants dominate.
2. The mining tonnage sits in a different code. Quarrying dimension stone (NAICS 212311), mining industrial and frac sand (212322), and mining the raw minerals that feed the grinders (2123) are all classified as mining, not manufacturing; abrasives likewise excludes mined natural grindstones (212399) as well as the machine tools that use abrasives (333517) [2][3]. So 3279 captures only the processing/fabrication value-add, not the rock underneath it. This is most distorting in 32799's miscellaneous bin, where raw frac sand is roughly 80%+ of all proppant tonnage and none of it is counted here [3].
3. Imports and embedded reporting shrink the abrasives figure specifically. A large share of U.S. abrasive consumption is imported — finished product and, above all, raw grain, where the USGS puts China at 96% of U.S. crude fused-aluminum-oxide imports and 97% of crude silicon-carbide imports for 2025 [2][7] — and the biggest sellers book abrasives inside much larger segments. Private market-research firms that add distribution and imports put the U.S. abrasives market nearer $5 billion against the ~$3.79 billion of federal production, and the global market at roughly $41–49 billion in 2025 [2].
4. Third-party "market" numbers cut both ways — check the scope before comparing. Some run larger than the federal figure because they add layers Census excludes: the Natural Stone Institute's dimension-stone ecosystem of 9,689 firms, 85,966 employees and $3.153 billion of payroll spans quarrying, fabrication, wholesale and installation across several codes, and countertop-market estimates that include imported slab and installation run to tens of billions [3]. Others run smaller because they are narrower: research valuing the "U.S. mineral wool insulation market" near $2.7 billion counts rock and slag wool only, while Census's $7.92 billion for that child folds fiberglass in [3]. Products that compete head-to-head are also scattered across codes — precipitated calcium carbonate is a chemical (325180), foam plastic insulation is 326140/326150, and even inside 32799 engineered quartz is claimed by two children at once, so no single "engineered stone" figure maps cleanly onto one code [3].

Net: treat $31.3 billion as U.S. manufacturing value-add sitting at the center of much larger flows of mined rock, imported material, and downstream industrial and construction spend.

4. The investable universe — where value concentrates across the children

The single most important fact for a public-market investor: there is no U.S.-listed company that is both a pure play and a leader anywhere in 3279 — in either child. Value concentrates in private companies, PE portfolios, and the divisions of diversified multinationals. This pass forces one refinement to the old view: 32799 does now have a genuine listed near-pure-play in ground minerals, and a whole downstream layer of insulation names worth owning by proxy. Abrasives still has nothing of the kind. (Tickers, revenues and margins are reserved for this section and Section 10.)

Where the listed money actually is, by child:

• All Other Nonmetallic (32799) — the deepest, if still indirect, listed menu. Mineral wool is the cleanest theme: Owens Corning (NYSE: OC), the largest U.S. insulation maker, did $3.7B of insulation sales and $848M of EBITDA (a 23% margin) in 2025 — about a third of a company that also does roofing and doors — and Rockwool A/S (Nasdaq Copenhagen: ROCK-B) is the closest listed pure-play on stone wool at €3.88B group revenue, a 21.3% insulation EBITDA margin and €737M of North American revenue, but foreign-listed [3]. One layer downstream, TopBuild (NYSE: BLD, $4.3B of 2025 insulation-and-accessories revenue) and Installed Building Products (NYSE: IBP) give the same demand without furnace and energy risk [3]. Ground/treated minerals: Minerals Technologies (NYSE: MTX; ~$2.07B of 2025 revenue at a 25.0% production margin) is the largest listed U.S. position, with a talc-litigation overhang, and Oil-Dri (NYSE: ODC; ~$486M FY2025 revenue at a 29.5% gross margin) is the closest thing to a U.S.-listed pure-play anywhere in 3279 — small, and not the industry leader; Imerys (Euronext Paris: NK) is the European leader [3]. Cut stone exposure is thin and troubled — Caesarstone (Nasdaq: CSTE) has shrunk to ~$397M of 2025 revenue at an 18.4% gross margin, with Matthews International (Nasdaq: MATW) offering a memorial slice and the best operators (Cambria, Cosentino, MS International, Polycor) private [3]. Miscellaneous is really a shale-sand bet — Atlas Energy Solutions (NYSE: AESI) and Smart Sand (Nasdaq: SND) — with the construction-materials value embedded in Sika, RPM International and CRH [3].
• Abrasives (32791) — no direct U.S. play at all. Public-market exposure comes through diversified industrials — 3M, which disclosed $1.34 billion of Abrasives division sales for 2025 inside a $24.9 billion company [2][11], and France's Saint-Gobain, owner of Norton and generally regarded as the world's largest abrasives maker, which does not break abrasives out of its €46.6 billion (2024) group revenue [2]. Foreign-listed specialists give a more concentrated bet: Japan's Fujimi (net sales ¥62.5 billion FY3/2025, up 22% on semiconductor polishing), Noritake (industrial products ¥55.7 billion) and Asahi Diamond; India's Carborundum Universal (CUMI) (standalone abrasives revenue ₹11.95 billion FY3/2025) and Grindwell Norton, a Saint-Gobain affiliate that reports abrasives separately [2]. The standout corner is semiconductor polishing — Fujimi and Entegris in CMP (chemical-mechanical planarization, the slurry-and-pad step that flattens chip wafers) — the fastest-growing, highest-value slice of the whole level [2].

The through-line across both children: to own 3279 through public markets you buy a diversified industrial or building-products/specialty-chemicals compounder where the code is one slice (3M, Saint-Gobain, Owens Corning, Minerals Technologies, Sika, RPM, CRH), a foreign-listed leader (Rockwool, Imerys, Carborundum Universal, Noritake, Fujimi), a downstream demand proxy (TopBuild, IBP), or a small, stressed or hybrid name (Oil-Dri, Caesarstone, Matthews, Atlas Energy, Smart Sand). The pure, clean, U.S.-listed leader does not exist in either child — which is precisely why the private-market opportunity is the larger one.

5. How the money works

Both children are materials businesses, not equipment or service businesses: owners make money on volume, mix and unit economics, not big-ticket sales. Three mechanics recur across the level.

Consumable and commodity demand — tied to activity, not capex. An abrasive is designed to be consumed and reordered — a razor-and-blades revenue stream tied to customers' activity levels (how much they are cutting and grinding) rather than their capital spending [2]. Ground minerals, stucco, insulation and proppants are likewise inputs bought by the ton as work proceeds. That makes the whole level cyclical and operating-leveraged, tracking industrial production, the manufacturing PMI (Purchasing Managers' Index, a monthly activity gauge), housing turnover and drilling — and distributor destocking can make reported sales swing more than underlying consumption [2].

The buyer's cost structure is what lets the premium end charge. A supplier illustration the abrasives child now documents puts abrasives at under 2% of cost at a large metal fabricator while the complete grinding-and-finishing process runs 10–15% — so labor time, rework and throughput matter far more to the customer than the consumable's sticker price [2]. The 32799 bin monetizes the same asymmetry differently: mineral fillers that let a coatings maker cut costly titanium-dioxide (TiO₂) pigment, and Minerals Technologies' satellite PCC model of building the mineral plant inside the customer's paper mill on a ~10-year requirements contract [3]. In both children, the pricing argument is about the customer's total process cost, not the material's unit price.

Freight and geography set the competitive map (strongest in 32799). The 32799 products are heavy with low value per pound, so shipping can cost as much as the product — forcing plants to sit on top of the raw material or next to the customer, limiting imports, and handing incumbents a geographic moat rather than a patent one: shipping a $100–$200/tonne mineral powder more than a few hundred miles can double its delivered cost, mineral wool is "mostly air" (which is why Rockwool and Knauf build U.S. plants instead of exporting), and proppant winners now compete on last-mile logistics to the wellsite [3]. Abrasives is the exception that proves the rule: higher value per pound means it trades and imports freely, so its competitive pressure comes from global low-cost producers and tariffs rather than local freight [2].

Capacity utilization, energy and the value ladder set the margin. For the plant-scale operations — abrasive-grain fusing in electric-arc furnaces, mineral-wool furnaces, mineral calcining and grinding — fixed costs are high and the swing factors are utilization and energy [2][3]. Mineral wool is the measured case: the Energy Information Administration puts it at 3,239 million Btu per employee, a new stone-wool plant costs $150–175 million or more, and Owens Corning's 2025 arithmetic shows how fast it turns — roughly 5% lower volume, $50 million of production-downtime cost and $42 million of input inflation against only $27 million of price and $30 million of manufacturing improvement [3]. In abrasives, fusing energy is the analogous exposure, softened somewhat by recycling — as much as 30% of fused aluminum oxide and about 5% of silicon carbide is recycled [2][7]. Margin then splits sharply by grade in both children:

• In abrasives, commodity sandpaper and cut-off wheels are low-margin and price-competitive, while engineered ceramic-grain belts, precision vitrified wheels and superabrasive (synthetic diamond and cubic boron nitride, or CBN) tools carry much higher margins protected by intellectual property and technical service [2].
• In minerals, the same rock sells across a wide price range by fineness, brightness and coating — ultrafine and surface-coated grades earn far more than bulk filler [3]. Insulation and branded quartz add brand, spec and code-approval pricing power on top [3].

An asymmetry in what can actually be measured. The 32799 child now supports a real margin ladder — Owens Corning's insulation EBITDA margin ~23%, Rockwool's ~21.3%, CRH's Americas Building Solutions 20.7%, Minerals Technologies' 25.0% production margin, Oil-Dri's 29.5% gross margin, against Caesarstone fighting to hold an 18.4% gross margin at the fabrication bench (these are different margin definitions; read the ordering, not the levels) [3]. Abrasives offers no equivalent, because no listed company discloses abrasives economics separately — and the child primer is explicit that Saint-Gobain's High Performance Solutions and 3M's Safety & Industrial segment margins should not be used as proxies [2]. The level's one genuine structural outlier remains the cut-stone fabrication shop inside 32799: labor-heavy, low-capital, thin per-job margins on the spread between slab cost and installed price — the reason 32799 holds thousands of tiny establishments while abrasives and mineral wool concentrate similar revenue in a few hundred capital-intensive plants [3].

6. What drives demand

Because the two children serve overlapping but distinct end-markets, the level's demand is a bundle of loosely correlated drivers rather than one cycle:

• Industrial production and metal fabrication — the single biggest driver of abrasives, and the master read for the whole level's factory-facing half [2].
• Housing turnover and repair-and-remodel (R&R) — the biggest driver inside 32799: countertops, stucco, dry-mix mortar and reinsulation. Residential uses are roughly 70%+ of countertop demand, and Caesarstone attributes 60%–70% of its U.S., Australian and Canadian revenue to renovation rather than new build. Rate-sensitive and discretionary [3].
• New construction (residential and non-residential) — feeds countertops, cladding, insulation, stucco and dry-mix [3].
• Building-energy and fire-safety codes, plus the retrofit runway — the structural tailwind unique to mineral wool: each tightening of the International Energy Conservation Code (IECC) or California Title 24 raises required insulation per building, and adoption is state-by-state so the ratchet is uneven. Behind it sits a long retrofit runway — the Department of Energy puts opaque building envelopes at 28% of building energy use and expects nearly 93% of today's homes still to exist in 2050 [3].
• Autos and aerospace — body finishing, powertrain machining and turbine-blade grinding pull premium abrasives [2].
• Semiconductors and electronics — CMP slurries and wafer polishing, a structural tailwind from the AI-driven chip build-out and CHIPS-Act fabs, visible in Fujimi's 22% sales growth [2].
• Automation — newly a driver in its own right for abrasives: robotic finishing cannot compensate for erratic consumables, so it raises the premium on predictable wear and consistent cut rate, and suppliers are shifting from selling consumables to selling process solutions (Mirka's 2022 purchase of robotics specialist Flexmill; 3M's robotic and fixed-automation finishing systems) [2].
• Industrial end-markets for mineral fillers — paint/coatings, plastics/rubber, paper/packaging and filtration pull carbonate, talc, silica and diatomite (U.S. talc end-uses run 36% plastics, 19% paint, 17% ceramics, 12% paper; filtration is about 60% of diatomite use); a standing tailwind is TiO₂ extension [3][7].
• Oil-and-gas drilling — barite weighting agent (more than 90% of which goes into drilling mud) and engineered proppants rise and fall with rig counts and crude prices, the most volatile demand in the level [3].
• Data centers — the AI build-out now reaches both children through different doors: CMP and wafer polishing in abrasives, and insulation plus construction chemicals in 32799, where Sika called data centers the bright spot in 2025 U.S. demand while commercial building softened [2][3].
• Purification, environmental and horticultural niches — bleaching clay for edible-oil and renewable-diesel pretreatment, bentonite liners and water treatment, plus perlite and stone-wool growing media for controlled-environment agriculture [3].
• Reshoring, infrastructure and low-carbon procurement — forward-looking supports on both sides: CHIPS-Act fabs and IIJA construction for the factory-facing half [2], and the Federal Highway Administration's up-to-$2 billion low-carbon materials program pulling dry-mix makers toward lower-clinker formulations [3].

7. Regulation

Both children are lightly regulated as products but sit at the center of workplace-safety and trade regimes — and one hazard runs through most, though not all, of the level.

Respirable crystalline silica — the defining cross-cutting hazard, with one clear exception. Cutting or grinding stone, minerals and (in use) abrasive tools releases fine silica dust that causes silicosis, an incurable lung disease, and lung cancer. The federal Occupational Safety and Health Administration (OSHA) sets a permissible exposure limit (PEL) of 50 micrograms per cubic meter over an 8-hour shift, with a 25 µg/m³ action level plus exposure assessment, engineering controls and medical surveillance [4]; in April 2024 the Mine Safety and Health Administration (MSHA) matched that limit for mines, effective June 2024 with metal/non-metal compliance phasing through 2025–2026, so operators running both a mine and a plant sit under both regimes [6]. The exposure shows up at both ends of 3279 — abrasive-blasting and grinding operations, and stone/mineral fabrication — but the children now let us be precise about where it does not: mineral wool is exempt, because its hazard is fiber irritation rather than crystalline silica, and international and U.S. bodies reclassified glass, rock and slag wool as IARC Group 3, "not classifiable as to carcinogenicity" [3]. Roughly a quarter of the level's dollars therefore sit outside the silica story entirely.

• The engineered-stone silicosis crisis (inside 32799). Engineered quartz can be >90% crystalline silica, and cutting it has produced a wave of severe, sometimes fatal disease among fabricators. California confirmed 542 silicosis cases and 29 deaths as of April 2026, concentrated among young workers and more than half in Los Angeles County [3][10]. Its STOP Act (SB 20) banning dry-cutting and requiring shop certification took full effect January 1, 2026, and Cal/OSHA has moved toward prohibiting fabrication of engineered stone above 1% silica — a first-in-the-nation step echoing Australia's July 2024 ban at the same threshold [3][10]. OSHA also runs a focused inspection initiative expressly targeting engineered-stone fabricators [3]. The liability tail is already visible: Caesarstone booked a $50 million provision at year-end 2024 for silicosis claims [3]. This is simultaneously a safety mandate, a product-liability wave, and an existential threat to what had been the fastest-growing product line in the bin.
• Abrasive-specific safety. Beyond silica, abrasives is governed by OSHA machine-guarding rules and the ANSI B7.1 code (ANSI = American National Standards Institute) because a bursting grinding wheel is a lethal hazard — the child primer frames formulation control, reinforcement, speed testing, labeling and channel traceability as economic requirements rather than compliance line items [2][5]. One second-order effect worth noting: the silica rule has pushed many blasting operators off silica sand and onto alternative media that carry heavy-metal hazards of their own, so the regulation moved the risk rather than removing it [2].

Trade remedies and tariffs — now active in both children. Section 301 tariffs of up to 25% raise the cost of Chinese abrasives and abrasive grain, and abrasive material costs rose an estimated 15–25% in 2025 on trade-policy changes; newly, in September 2025 the U.S. International Trade Commission found that subsidized and dumped Chinese sol-gel alumina ceramic abrasive grain materially injured the domestic industry, leading to AD/CVD orders [2][8]. On the other side, domestic quartz makers hold steep AD/CVD orders on imported quartz surface products — China in 2019 at combined rates of roughly 265%–340%, then India and Türkiye at roughly 3.8%–80.8%, continued January 15, 2026 after sunset reviews — propping up domestic slab pricing [3][9]. Both regimes cut the same way: protection for domestic grain and slab makers, higher input costs for the downstream converters and fabricators that buy from them.

Product-liability litigation. Talc mined near tremolite can carry trace asbestos; mass-tort suits drove talc units into bankruptcy (Imerys Talc America, 2019; Minerals Technologies' Barretts Minerals, October 2023, with a proposed ~$450 million injury trust), and Minerals Technologies recorded a further $215 million provision in 2025 for current and future talc claims — a liability that can dwarf the underlying operation [3].

Environmental and other. Clean Air Act permits for particulate, furnace, kiln and grain-fusing emissions are real plant-siting hurdles in both children — abrasive grain fusing, curing ovens and coated-abrasive lines carry particulate, combustion and volatile-organic-compound obligations [2], while 32799 sits under EPA's New Source Performance Standards for nonmetallic mineral processing and, for mineral wool, binder-chemistry (formaldehyde) and NESHAP air rules; Rockwool's West Virginia plant drew years of community opposition [3]. Building and seismic codes govern stone cladding and stucco/EIFS moisture performance, and ASTM fire tests plus the FTC R-value Rule govern insulation labeling [3].

8. Consolidation

The direction of travel across 3279 is toward more private ownership, more private equity, and fewer public pure-plays — the opposite of most maturing industries — but the mechanism differs by child:

• Abrasives (32791) — top-heavy pyramid with a fragmented base. The four largest firms already take just over half of U.S. revenue and the top 50 take ~87%, yet ~240 firms exist [2]. Saint-Gobain built its lead by acquiring Norton (1990) and dozens of smaller businesses. The 3M retreat has now closed out: it agreed in September 2025 to sell its Precision Grinding & Finishing business (about $130 million of sales) and completed the sale in April 2026 — so 3M's $1.34 billion of 2025 Abrasives sales includes an operation it no longer owns, with no pro-forma disclosed for what remains [2][11]. The fragmented small-firm base remains fertile ground for PE roll-ups, and many of the strongest specialists (Tyrolit, with more than €740 million of turnover; Klingspor; PFERD; Mirka) stay family- or privately controlled [2].
• All Other Nonmetallic (32799) — four different consolidation stories, in an unusually heavy 2024–2025 deal stretch. Cut stone is an atomized field (HHI 52) that is a textbook PE and search-fund roll-up target, with Polycor's 2022 purchase by Birch Hill and Investissement Québec the scaled version. Ground minerals sees majors prune liability-laden lines while PE buys — Apollo took U.S. Silica private for ~$1.85 billion in July 2024. Mineral wool consolidates via capacity build-out by well-capitalized incumbents (Owens Corning bought Thermafiber and Paroc; Knauf acquired USG; Rockwool built Ranson, WV and broke ground on Wallula, WA at $175 million for 2028), with capital, permitting and freight moats blocking newcomers. Miscellaneous combined a brutal proppant shakeout with a packaged-concrete megadeal — QUIKRETE closed its acquisition of Summit Materials at an enterprise value of ~$11.5 billion in February 2025 — plus a European construction-chemicals roll-up (Sika's ~$5.9 billion purchase of BASF's construction-chemicals business) [3].

A rollup-level reading: the acquirers ignore the code boundaries. The same names recur across children — Saint-Gobain owns Norton in abrasives, CertainTeed in insulation, and figures in ceramic-proppant history; Apollo's U.S. Silica spans both the ground-minerals and proppant narratives; Imerys appears in specialty minerals and as a ceramic-proppant competitor [2][3]. Ownership consolidates faster than the statistics suggest, and 3279's low measured HHI (Section 2) misses it entirely, because Census counts markets while acquirers buy capabilities.

9. Risks

• Cyclicality across multiple engines. The level rides industrial production (abrasives, ground minerals), housing and remodeling (cut stone, stucco, insulation) and oil-and-gas drilling (barite, proppants). Different cycles, but each half is fully exposed to a downturn in its own demand [2][3].
• Silica regulation and litigation — the biggest single swing factor. A spreading engineered-stone ban would gut the highest-growth product line in 32799 and trigger liability (Caesarstone's $50 million provision is the first installment); tighter enforcement raises compliance costs across abrasive-blasting and fabrication alike; talc-asbestos suits remain a bankruptcy-scale risk in ground minerals, with a fresh $215 million provision booked in 2025 [2][3][10].
• Input concentration and trade policy — sharper, and worse, than the previous reading. The 2026 USGS data shows U.S. dependence on China at 96% of crude fused-alumina imports and 97% of crude silicon-carbide imports, though only 15% and 57% respectively for ground and refined grades — dependence is near-total upstream and much lower for processed grain, and the earlier assumption of merely "majority" reliance understated it [2][7]. Slab, graphite and barite lean on foreign sources on the 32799 side [3]. Tariffs and AD/CVD orders cut both ways, protecting domestic grain and slab producers while inflating input costs for downstream converters and fabricators [2][3][8][9].
• Energy and freight inflation. The plant-scale operations depend on cheap power for melting and fusing and cheap transport for heavy product; sustained increases compress margins and idle expensive furnaces, and a badly timed furnace outage or rebuild can hurt as much as commodity inflation [2][3].
• Overcapacity and lumpy additions (32799). Mineral-wool lines are large, indivisible investments and several plants are due to commission in 2027–2028; if they land into soft demand, today's scarcity pricing goes with them [3].
• Substitution and fashion risk. Laser, waterjet and plasma cutting, alternative blasting media, and longer-lasting engineered products that reduce units sold per job all nibble at abrasives [2]; quartz vs. granite vs. porcelain, mineral wool vs. foam plastics, and untreated sand vs. ceramic proppant can strand a firm levered to yesterday's product [3].
• Thin, stressed public float. For public-market investors the shortage of clean listed names in either child concentrates risk in a few diversified, foreign, small or troubled stocks [2][3].
• Customer and channel concentration — big-box order flow pressures cut-stone fabricator pricing and large distributors hold buying power over insulation volumes [3], while abrasives makers carry distributor and end-market concentration in cyclical autos, aerospace and construction plus a high-value but historically volatile semiconductor niche [2].
• Permitting and community opposition for new plants, and safety liability and labor scarcity for skilled fabricators, installers and furnace/grinding operators [2][3].

10. How to invest, and the outlook

Public-market routes (limited, indirect, and clustered).

• Cleanest listed themes sit inside 32799: Owens Corning (NYSE: OC) for diversified U.S. insulation exposure, or Rockwool (Copenhagen: ROCK-B) for the closest listed pure-play on stone wool's code-and-fire-safety tailwind; TopBuild (BLD) and Installed Building Products (IBP) for the same demand without furnace and energy risk; Minerals Technologies (NYSE: MTX) for specialty minerals (mind the talc overhang) and Oil-Dri (NYSE: ODC) as the nearest U.S.-listed pure-play in the level; Imerys (Euronext Paris: NK) for the European mineral leader [3].
• The fastest-growing corner sits inside 32791: Fujimi and Entegris for semiconductor polishing (CMP) — the highest-value slice of the whole level [2].
• Diversified industrials spanning both children: 3M and Saint-Gobain (abrasives plus, for Saint-Gobain, CertainTeed insulation and stone), plus Sika, RPM International and CRH for the construction-materials half [2][3].
• Foreign-listed abrasive specialists: Carborundum Universal (CUMI), Grindwell Norton, Noritake, Asahi Diamond [2].
• Lower-quality, stressed or hybrid: Caesarstone (CSTE) (a shrinking turnaround under regulatory siege), Matthews International (MATW), foreign-listed quartz makers LX Hausys and Vicostone, and Atlas Energy Solutions (AESI) and Smart Sand (SND) (really shale-sand bets) [3].

Private-market routes (where most of the value and deal flow actually sit).

• Abrasives — most dedicated makers (Tyrolit, Klingspor, PFERD, Mirka, Weiler, Camel Grinding Wheels) are private, and the ~240-firm U.S. base is fragmented and roll-up-friendly; the child's diligence list is worth reading first, above all whether the differentiation sits in the target or in its grain supplier, plus imported-grain dependence, energy exposure, SKU-level profitability and wheel-testing records [2].
• Cut stone — buy and professionalize local fabrication shops; the HHI-52 structure is the textbook roll-up, with diligence on builder concentration, slab yield, callbacks, silica controls and workers'-compensation history [3].
• Ground minerals — buy a regional grinder or single-mineral processor, or back a PE roll-up (the Apollo/U.S. Silica model), allocating reclamation, silica and talc liabilities site by site [3].
• Mineral wool — direct plant, joint-venture or greenfield investment (the Rockwool/Knauf model), or the more accessible layer of regional installation, specialty distribution, mechanical-insulation fabrication and energy-retrofit contracting [3].
• Miscellaneous — cash-generative regional stucco/dry-mix platforms (freight-protected local near-monopolies) and distressed proppant assets in downturns, with engineered-stone silica-liability tail risk flagged [3].

Near-term drivers to watch (forward-looking judgment, not guarantees):

1. Industrial production and the manufacturing PMI — the master switch for the factory-facing half (abrasives, ground minerals) [2].
2. Interest rates and housing turnover — the master switch for the construction-linked half (cut stone, stucco, insulation) [3].
3. The silica-regulation wave — a California engineered-stone fabrication ban above 1% silica is the swing factor for the stone-heavy parts of 32799: bearish for high-silica quartz, potentially bullish for natural stone, porcelain and low-silica formulations — and largely neutral for abrasives-grade materials and for mineral wool, which the science now places outside the silica story altogether [3][10].
4. Trade remedies on both sides of the level — the new sol-gel alumina grain orders and the continued quartz orders each protect a domestic producer while taxing a domestic converter; watch which side of that trade any given holding sits on [2][3][8][9].
5. The semiconductor and building-code ratchets — the two durable structural growers: CMP and superabrasives on the chip build-out (32791) and mineral wool on tightening energy and fire codes plus the retrofit runway (32799), both of which lift content regardless of the broad cycle [2][3].
6. New mineral-wool capacity timing — Rockwool expects North American demand to exceed capacity until 2028; the plants commissioning in 2027–2028 decide whether today's tightness converts to margin or to price competition [3].

Net. NAICS 3279 is best understood not as an industry to "buy" but as two picks-and-shovels mineral-processing businesses reported together: a small, top-heavy, globally-traded abrasives industry (~12% of the dollars) and a large, freight-protected, mostly-private "all other nonmetallic" bin (~88%). Both are cyclical materials plays with thin public menus and deep private ones; both now sit under active trade remedies; most of the level sits under a silica-safety reckoning, with mineral wool the notable exception; and in both, the durable growth is in the premium end — engineered grain, superabrasives and semiconductor CMP in abrasives, code-driven mineral wool and specialty mineral grades in the bin, where the story is rising value per unit rather than rising tonnage. Public investors get a thin, indirect menu clustered in a handful of diversified compounders, one or two foreign leaders and a couple of small near-pure-plays; private investors get the real opportunity — and the clearest thesis in either child remains "own the premium, safer end," not "buy the code broadly."

---

Sources

1. U.S. Census Bureau — 2022 Economic Census (receipts ~$31.33B; 2,979 firms; CR4 19.3%, CR8 31.9%, CR20 45.4%, CR50 59.1%; HHI 159.2) and 2023 County Business Patterns (3,341 establishments; 83,488 employees; ~$5.66B annual payroll; ~$1.43B first-quarter payroll) for NAICS 3279. (Histometrics ingested federal statistics, stats-3279.md.) https://www.census.gov/programs-surveys/economic-census.html; https://www.census.gov/programs-surveys/cbp.html
2. Histometrics child primer — NAICS 32791, Abrasive Product Manufacturing (and its single national industry 327910: receipts ~$3.79B; 242 firms; 264 establishments; 10,976 employees; $785.3M annual payroll and $203.2M first-quarter payroll; CR4 50.5%, CR8 59.8%, CR20 74.1%, CR50 87.4%, HHI suppressed; bonded/coated/superabrasive product families; abrasives <2% of fabricator cost vs. 10–15% for the grinding-and-finishing process; U.S. market ~$5B and global ~$41–49B versus $3.79B of federal production; 3M Abrasives $1.34B of 2025 sales inside $24.9B; Saint-Gobain €46.6B 2024 group revenue; Fujimi ¥62.5B, Noritake ¥55.7B, Carborundum Universal ₹11.95B FY3/2025; Tyrolit >€740M; Precision Grinding & Finishing divestiture ~$130M of sales; recycling ~30% fused alumina and ~5% silicon carbide; 15–25% abrasive material cost inflation in 2025). Draws on the 2022 Economic Census, 2023 CBP, USGS Mineral Commodity Summaries 2026, OSHA silica and abrasive-wheel standards, ANSI B7.1, USITC determinations, EPA AP-42, and company filings.
3. Histometrics child primer — NAICS 32799, All Other Nonmetallic Mineral Product Manufacturing (receipts ~$27.54B; 2,745 firms; 3,077 establishments; 72,512 employees; ~$4.88B annual payroll; CR4 20.9%, CR8 32.7%, CR20 46.5%, CR50 60.7%, HHI 172.3; four children — cut stone 327991 at ~$7.96B, HHI 52, ~$57,900 average pay; ground/treated mineral 327992 at ~$6.45B, HHI 503, ~$85,000 average pay; mineral wool 327993 at ~$7.92B, HHI 1,106, CR50 97.6%, ~$73,000 average pay; miscellaneous 327999 at ~$5.21B, CR4 48.8%; Owens Corning insulation $3.7B sales / $848M EBITDA / 23% margin 2025; Rockwool €3.88B revenue, 21.3% insulation EBITDA margin, €737M North America, Wallula WA $175M; TopBuild $4.3B; Minerals Technologies ~$2.07B revenue, 25.0% production margin, $215M talc provision; Oil-Dri ~$486M FY2025 revenue at 29.5% gross margin; Caesarstone ~$397M 2025 revenue, 18.4% gross margin, $50M silicosis provision; CRH Americas Building Solutions $7.122B at 20.7% adjusted EBITDA margin; Apollo/U.S. Silica $1.85B; QUIKRETE–Summit ~$11.5B; Sika/MBCC ~$5.9B; EIA 3,239 million Btu per employee; DOE envelope and retrofit data; ~85% dimension-stone net import reliance; Natural Stone Institute ecosystem 9,689 firms / 85,966 employees / $3.153B payroll). Draws on the 2022 Economic Census, 2023 CBP, USGS Mineral Commodity Summaries 2026, OSHA/MSHA/EPA sources, EIA MECS, California STOP Act and Cal/OSHA actions, USITC/Federal Register AD/CVD determinations, and company filings.
4. U.S. Occupational Safety and Health Administration — Respirable Crystalline Silica standard, 29 CFR 1910.1053 / 1926.1153 (PEL 50 µg/m³ 8-hour TWA; action level 25 µg/m³). (Cited via the child primers [2][3].) https://www.osha.gov/silica-crystalline
5. U.S. Occupational Safety and Health Administration — 29 CFR 1910.215, Abrasive wheel machinery. (Via child primer [2].) https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.215
6. Mine Safety and Health Administration — "Lowering Miners' Exposure to Respirable Crystalline Silica," final rule (PEL 50 µg/m³; effective June 2024), 2024. (Via child primer [3].) https://www.msha.gov/silica-final-rule-30-cfr-part-60-resources
7. U.S. Geological Survey — Mineral Commodity Summaries 2026 (abrasives import-source shares: crude fused alumina China 96%, ground/refined 15%; crude silicon carbide China 97%, ground/refined 57%; recycling rates; diatomite, perlite, talc and dimension-stone data). (Via child primers [2][3].) https://pubs.usgs.gov/periodicals/mcs2026/mcs2026.pdf
8. U.S. International Trade Commission — "Sol-Gel Alumina Ceramic Abrasive Grain from China," injury determination leading to antidumping and countervailing-duty orders, September 2025. (Via child primer [2].) https://www.usitc.gov/press_room/news_release/2025/er0903_67493.htm
9. Federal Register — "Certain Quartz Surface Products From India and the Republic of Türkiye: Continuation of Antidumping and Countervailing Duty Orders," January 15, 2026. (Via child primer [3].) https://www.federalregister.gov/documents/2026/01/15/2026-00739/certain-quartz-surface-products-from-india-and-the-republic-of-trkiye-continuation-of-antidumping
10. California engineered-stone silicosis record and regulatory response — Public Health Watch / Capital & Main (542 confirmed cases and 29 deaths as of April 2026; >90% silica content); KQED (Cal/OSHA ban vote and proposed >1% silica threshold; Australia's July 2024 ban); Stone World (STOP Act / SB 20 in full effect January 1, 2026). (Via child primer [3].) https://publichealthwatch.org/2026/04/14/california-silicosis-ban-lung-disease-stone-countertops/; https://www.kqed.org/news/12084910/california-steps-closer-to-ban-on-engineered-stone-after-silicosis-surge; https://www.stoneworld.com/articles/95545-californias-stop-act-sets-strictest-silica-rules-in-us-for-stone-fabricators
11. 3M Company — Form 10-K FY2025 (Abrasives division sales $1.340B; total company sales $24.948B) and First-Quarter 2026 earnings transcript (Precision Grinding & Finishing sale closed April 2026). (Via child primer [2].) https://www.sec.gov/Archives/edgar/data/66740/000006674026000014/mmm-20251231.htm; https://d1io3yog0oux5.cloudfront.net/_97a7bf5fea9d859971f982f6e8c7926e/3m/db/3222/30995/webcast_transcript/MMM-USQ_Transcript_2026-04-21.pdf