Machinery Manufacturing (United States) — NAICS 333

A Histometrics rollup primer for public-market and private investors. NAICS (the North American Industry Classification System) is the standard code the U.S. government uses to group businesses by activity. Code 333 is a three-digit subsector inside the manufacturing sector — the machinery-building layer of American industry. It contains seven four-digit industry groups: 3331, 3332, 3333, 3334, 3335, 3336, and 3339. Figures for this level are our ingested U.S. federal statistics; forward-looking statements are labeled as judgments, not facts.

---

1. Overview

NAICS 333 is the part of American manufacturing that builds the machines other people use to make, move, grow, extract, cool, and power everything else. A farmer's combine, a semiconductor fab's etch tool, a restaurant's fryer, a data center's chiller, a factory's stamping die, a power plant's gas turbine, a warehouse's forklift — all of them are made by firms inside this subsector. It is, almost end to end, a "picks-and-shovels" business: with a handful of consumer-facing edges (power tools, residential air-conditioning, lawn-and-garden gear), these companies sell equipment to other producers, so their fortunes ride the industrial capital-spending cycle rather than the shopper.

The reason this is a coherent subsector — and not just an accounting bin — is that all seven children run on the same economic skeleton. Four traits repeat everywhere:

• The razor-and-blades installed base. A machine is sold once but runs 7 to 40 years, throwing off a decades-long, higher-margin stream of spare parts, consumables, service, rebuilds, and increasingly software. The revised children now put hard numbers on it in every corner: aftermarket is roughly two-thirds of mining-equipment revenue (Epiroc's 2025 mix was 66% aftermarket to 34% equipment) and ~80% of one oilfield maker's revenue[3]; 71% of Kadant's sales and ~50% of JBT Marel's[4]; 28% of Carrier's 2025 sales, with Trane splitting revenue $13.98B product to $7.34B service[6]; ~40% of Regal Rexnord's sales move through distributors and RBC's industrial segment sells $751.9M through distribution and aftermarket against $331.0M to equipment makers[8]; and Otis's service segment is roughly two-thirds of its revenue[9]. That annuity — far less cyclical than the original sale — is the single most valuable asset a firm here can own, and the quiet engine of quality across the whole subsector.[3][4][5][6][7][8][9]
• Deep cyclicality with high operating leverage — and less internal hedging than the structure suggests. Demand is derived, and heavy fixed factory costs make profits swing hard with volume. When orders fall, makers deliberately build fewer machines than customers buy, drawing down channel inventory and crushing throughput on the way down. The sharpest revision on this page comes from 3331, which now warns that its own three cycles — farm income, construction, commodity prices — did not de-synchronize in 2025: rates, tariffs, and channel destocking compressed all three at once, and it tells readers not to underwrite that group as internally hedged.[3] Read the same warning up one level.
• A shared input and tariff exposure, now measured. Steel, aluminum, copper, motors, and electronics dominate every bill of materials, so the same Section 232 metals tariffs and Section 301 China duties feed straight into every child's costs at once — and the bills are disclosed: Deere absorbed ~$600 million of tariff cost in 2025 and guided to ~$1.2 billion for 2026, Caterpillar reported $2.1 billion of unfavorable manufacturing costs[11], GE Vernova estimated $250–350 million for 2026 after contractual protections, and Hyster-Yale identified roughly $100 million on 2025 inventory purchases.[3][8][9][11]
• Standards, not rate-setting. None of this is an economically regulated industry — there is no rate base, no price-setting authority, no reimbursement regime. Regulation shapes product design and creates replacement cycles instead.

The distinctive value of a rollup, though, is the contrast. These seven children are bound by that shared model but ride seven different cycles, sit at very different points on their cycles right now, span a fifteen-to-one range in how concentrated they are, and — the fact that matters most to an investor — differ enormously in who owns them and how you can buy in, from clean large-cap pure-plays to industries that are almost entirely private. Section 2 lays them side by side; the rest of the page treats the subsector as a whole.

---

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

The table orders the children by NAICS code. "Share of level" is each child's slice of the subsector's 2022 shipments and 2023 employment; "concentration" is each child's four-firm share (CR4, the share of receipts held by the four largest firms) and Herfindahl-Hirschman Index (HHI, the standard concentration gauge — below 1,500 is "unconcentrated"). Tickers are held back for Sections 4 and 10.

CBP = County Business Patterns, an annual Census dataset; EC = the five-yearly Economic Census; OEM = original equipment manufacturer; PE = private equity; ETF = exchange-traded fund (a basket of securities that trades like one share).

Three contrasts do most of the analytical work.

• Size is lopsided, and it runs opposite to concentration. Two children — 3339 (~28% of receipts) and 3331 (~25%) — are together more than half the subsector, yet they sit at opposite ends of the concentration scale (3339 is the most fragmented, 3331 among the most concentrated). The most concentrated child, 3336 (engines/turbines, HHI 589), is a mid-sized slice; the most fragmented, 3335 (metalworking, HHI 37), is small. Bigger does not mean more concentrated, and neither one predicts how easy the child is to own. The same inversion nests: inside 3339, the largest child is the least buyable as a single theme and the smallest sub-market (elevators) has the best listed franchise.[9]
• The children ride seven different clocks — but "de-synchronized" is a weaker claim than it looks. Right now semiconductor tools (in 3332) and gas turbines (in 3336) are booming on the same force — the artificial-intelligence (AI) data-center build-out — while farm equipment (in 3331) is in a downturn and residential HVACR (in 3334) is working through a 20% volume decline. That is genuine divergence, and it is the reason to look at the subsector rather than one child. But the revised 3331 primer is explicit that its own three cycles compressed together in 2025 under a shared rate-and-tariff shock, and tells readers not to treat the group as internally hedged.[3] The honest version of the claim: the children's demand drivers are independent, but their cost bases and financing conditions are not, so the diversification protects against an end-market slump and does not protect against a rates-or-tariffs shock.
• Investability is a spectrum, and in 2026 it moved. At one end, several children hand you clean, liquid, large-cap pure-plays — Deere and Caterpillar (3331), the semiconductor-tool leaders (3332), GE Vernova and Cummins (3336), Otis and the welders (3339). At the other end, an entire child — metalworking (3335) — is essentially unbuyable on a U.S. exchange, and large pockets of others (HVAC fans and boilers, food and paper machinery, packaging, furnaces, conveyors) live in private equity, family firms, and foreign listings. The direction of drift is not uniform: 3333 gained a listing (Alliance Laundry's October 2025 IPO) and 3332 gained one by spin-off (Midera, July 2026), while 3339 lost four routes in a single stretch — Chart into Baker Hughes, Toyota Industries private at ~$39 billion, Mitsubishi Logisnext and Fujitec delisted, Honeywell's conveyor business sold to private equity.[4][5][9] Who owns each child is where the seven differ most sharply, and the deepest, most focused assets keep moving toward private hands.

Note the structure buried in the codes: five of the seven children (3332, 3333, 3334, 3335, 3336) are "single-child" groups — each equals exactly one five-digit industry, so their internal diversity lives one level down among their six-digit sub-industries. Only 3331 and 3339 split into multiple children at the four-digit level. For the deepest company-by-company detail, read the individual child primers; this page carries their conclusions up.

---

3. Size — the subsector as a whole

Our ground-truth federal figures for NAICS 333:

So NAICS 333 is a ~$445 billion, ~1.09 million-worker manufacturing base paying skilled-trade wages (~$78,300 average, well above the all-industry mean — these are machinists, welders, assemblers, and controls, thermal, and mechanical engineers). Average pay varies by child, from ~$69,000 in metalworking (3335) up to ~$98,000 in industrial machinery (3332), pulled up by high-wage semiconductor work — where pay runs about $151,000 per worker against ~$79–82,000 in that child's other three sub-industries. Engines/turbines and other general-purpose machinery both sit near ~$80,700.[4][7][8][9]

The children add up cleanly — the arithmetic proves the rollup. Receipts, employment, establishments, and payroll each sum exactly to the subsector totals:

• Receipts: $109.92B + $40.23B + $32.06B + $55.03B + $33.87B + $49.67B + $124.74B = $445.52B.[3][4][5][6][7][8][9]
• Employment: 193,869 + 111,497 + 80,504 + 147,794 + 134,208 + 96,251 + 322,023 = 1,086,146.[3][4][5][6][7][8][9]
• Establishments (2,805 + 2,836 + 1,731 + 1,685 + 5,772 + 899 + 5,940) and payroll ($14.57B + $10.95B + $6.65B + $9.83B + $9.31B + $7.77B + $25.98B) also reconcile to the level totals.

The only line that does not add up is the firm count: the seven children list 19,168 firms between them, but the subsector shows 18,779 — because roughly 390 companies operate plants in more than one child and are counted once here. Every child reports the same artifact at its own level, and the overlap is exactly the set of diversified industrials that recur across the tree: Caterpillar (construction and mining in 3331, and reciprocating engines and Solar Turbines in 3336), Illinois Tool Works (food equipment in 3333 and welding in 3339), Dover and Vontier (3333 and 3339), Nordson (3332 and 3339), Baker Hughes (oilfield equipment in 3331 and, since absorbing Chart in July 2026, process-gas compression in 3339), and Ingersoll Rand and IDEX (multiple sub-markets inside 3339) each show up in every child they touch but count once at the parent.[3][4][5][8][9]

The concentration figure, and why it is misleading. The subsector's CR4 of 12.5% and HHI of 67.7 look like one of the most fragmented, wide-open industries in America — but that is largely a statistical artifact of stapling seven unrelated product markets together. No single firm can dominate "machinery" when the machines range from a wafer etcher to a combine to a gas turbine. All seven revised children now make this point independently, and several quantify it: 3331's group CR4 of 40% is lower than two of its three children (agriculture 54.6%, construction 54.5%) because Deere leads one and Caterpillar the other; 3335's group HHI of 37.2 sits below every one of its five sub-industries; 3334's CR4 of 22% sits below two of its three grandchildren; 3339 is less concentrated than every child, each of which is less concentrated than almost all of its children.[3][4][5][6][7][8][9] Tellingly, the 333 level is not less concentrated than every child: metalworking (HHI 37) and other-GP machinery (HHI 39) are individually more fragmented than the subsector, because each contains thousands of small shops. But it is far less concentrated than the children with strong leaders (engines/turbines 589, HVACR 199, industrial machinery 153).

The 3339 primer adds a refinement worth carrying up: the federal ratio errs in both directions. It understates power where the real product market is narrower than the code — the top four elevator vendors hold roughly 55% of the U.S. market inside a child whose rollup HHI is 142, and the semiconductor sub-industry's domestic HHI of 1,153 badly understates true product-market shares (ASML holds essentially all of extreme-ultraviolet lithography, Lam ~45% of etch, KLA ~52% of process control). And it overstates coherence where the code is broader than the market — Census files aerial work platforms inside the crane code, a rental-fleet business with almost nothing in common with an installed bridge crane.[4][9] The lesson: never read competitive intensity or pricing power off the 333 number — or even off a child's number. Go two or three levels down, then check what the code actually contains.

Undercount and read-carefully caveats. Unlike sectors dominated by sole proprietors, gig workers, or informal operators, this is corporate-scale factory manufacturing that federal statistics measure well — every child reaches that conclusion independently, and there is little small- or individual-ownership undercount to flag. The caveats run the other way:

• These are U.S. factory figures, not the U.S. market — but do not assume every code understates its market. The $445B is the value of machinery shipped from American plants (foreign-owned ones included), not what U.S. buyers spend. Most of it is net-imported: roughly 85% of power tools, 2024 plastics-equipment imports equal to 68.8% of domestic shipments (molds 93%), machine-tool consumption of ~$10.5 billion against ~$5.9 billion of domestic production, U.S. pump imports at an estimated 25–30% of consumption. But two pockets run the other way — domestic forklift shipments of $13.18 billion exceed a demand-side U.S. forklift market estimated near $9.1 billion, and domestic industrial-furnace shipments of $3.45 billion exceed one private estimate of that end market.[3][4][6][7][9]
• Global brand revenue dwarfs the domestic line. Deere alone reported worldwide sales near $45.7 billion in fiscal 2025 — roughly the whole engine/turbine child — without any contradiction, because most of its revenue is booked abroad.[10] Caterpillar reported about $65 billion in 2025, of which Construction Industries alone was $25.1 billion, more than half the entire construction sub-industry's U.S. output.[11] Carrier's $21.7 billion of 2025 sales is roughly two-fifths of the HVACR child's entire domestic shipments, and Applied Materials' $28.4 billion is more than double the semiconductor sub-industry's ~$13.75 billion of domestic shipments.[4][6] A single company's global figure can exceed a whole child's U.S. line.
• Aftermarket and services leak out of the count. Installation, inspection, maintenance, repair, and modernization — the highest-margin activity in every child — are often classified outside manufacturing (elevator service under a construction code, crane repair under a repair code, oilfield services, much pump and compressor work). Otis alone booked ~$14.4 billion of global 2025 sales, roughly two-thirds of it service — more than three times the entire U.S. elevator manufacturing code.[9]
• Captive tooling is undercounted the other way. In metalworking (3335), a large volume of molds and dies is built inside the in-house tool rooms of automakers and aerospace firms and filed under the parent plant's code, so that child's domestic production is understated — even as its market is understated too, by imports.[7]
• Half of two children's history is not splice-able. The 2022 NAICS revision re-cut the detail beneath 3332 (merging sawmill with paper machinery, and combining two all-other codes), 3333 (merging three six-digit codes into one), and 3339 (folding scales into the residual bucket). Group totals are continuous; the detail beneath them is not.[4][5][9]

Read ~$445.52 billion as an accurate floor on U.S. domestic machinery production, not as the size of the end markets these firms serve, and never as the denominator for a global company's market share.

---

4. Investable universe — where value concentrates across the children

The single most important fact for a stock investor: there is no NAICS-333 pure-play and no "machinery" ETF that maps to this subsector. You assemble exposure child by child, and the available form changes completely depending on which cycle you want. Value concentrates in three layers.

1. The large-cap anchors (where a child has a dominant listed name). A handful of household-name industrials carry most of the subsector's liquid, buyable value:

• Agriculture/construction/mining (3331): Deere (NYSE: DE) — #1 in farm equipment plus a ~$11.4 billion construction business — and Caterpillar (NYSE: CAT) — #1 in construction and the top U.S.-listed mining-equipment name. Between them they anchor two of that child's three end markets; the mining half has no U.S.-listed pure-play at all.[3][10][11]
• Industrial machinery (3332): the semiconductor-tool leaders — Applied Materials (Nasdaq: AMAT), Lam Research (LRCX), KLA (KLAC) — plus foreign giants ASML and Tokyo Electron. This is the deepest, most liquid public menu anywhere in the subsector, and it dwarfs the child's other three sub-markets combined.[4]
• Engines & turbines (3336): GE Vernova (NYSE: GEV) for gas turbines and Cummins (NYSE: CMI) for engines — the two clean large-cap proxies for the electricity-load-growth theme, with the caveat that neither is a pure play: GEV's Power and Wind segments earned 14.7% and negative 6.6% in the same year.[8]
• Commercial & service machinery (3333): three near-pure-plays now — Middleby (Nasdaq: MIDD) in commercial foodservice (becoming a pure-play after spinning off food processing and selling most of residential), Alliance Laundry (NYSE: ALH) in commercial laundry, and Tennant (NYSE: TNC) in commercial floor-cleaning, a product family this page previously did not name.[5]
• Other GP machinery (3339): the richest pure-play set outside 3331 — Otis (NYSE: OTIS) in elevators, Columbus McKinnon (Nasdaq: CMCO) in cranes and conveyors, Hyster-Yale (NYSE: HY) in forklifts, and Lincoln Electric (Nasdaq: LECO) and ESAB (NYSE: ESAB) in welding.[9]
• HVACR (3334): clean listed exposure concentrates in the air-conditioning/refrigeration slice — roughly three-quarters of the child's shipments — led by Carrier, with Trane, Lennox, and AAON. Fans have no U.S. pure-play; boilers have real but minority listed exposure inside diversified parents.[6]
• Metalworking (3335): the one correction to "no pure-play anywhere" — Kennametal is a listed company whose core business genuinely is the code, with a $1.22 billion Metal Cutting segment squarely inside it. Everything else listed sits beside the shop, not in it.[7]

2. The diversified compounders (the connective tissue). A cluster of flow-, motion-, and fluid-handling industrials each own a slice of two or more children, so one stock buys diffuse exposure across the subsector: Dover (DOV), Illinois Tool Works (ITW), Vontier (VNT), Parker Hannifin (PH), IDEX (IEX), Ingersoll Rand (IR), Nordson (NDSN), Graco (GGG), Hillenbrand (HI), Baker Hughes (BKR), and the power-transmission names Regal Rexnord (RRX), Timken (TKR), and RBC Bearings (RBC). These are the same ~390 firms that count across children in the firm reconciliation above — you are buying these product lines inside a larger industrial story, and the relevant segment is often a minority of the parent.[4][5][8][9]

3. Private markets and foreign listings (the deepest, most focused assets). Whole swathes of the subsector are largely closed to U.S. public investors, and 2026 closed more of them:

• Metalworking (3335) is overwhelmingly private — thousands of independent mold, die, tool, and machine shops, with no clean pure-play for the level.[7]
• The best assets in several children are foreign or private: construction and mining equipment (Komatsu, Volvo, Epiroc, Sandvik, Weir, Metso); food, sawmill, and paper machinery (GEA, Krones, Valmet, Andritz, Dürr/HOMAG); HVAC fans and boilers (Greenheck, Twin City Fan, Mestek — family and ESOP firms); packaging, furnaces, and the deepest conveyor and forklift assets (ProMach, Duravant, Syntegon, Hytrol, Intralox, KION, Crown).[3][4][5][6][9]
• Four listed routes closed in 2026 alone — Chart Industries absorbed into Baker Hughes, Toyota Industries taken private at roughly $39 billion, Mitsubishi Logisnext and Fujitec delisted, and Honeywell's conveyor businesses sold to private equity — while no child gained an equivalent pure-play.[9]

The through-line across the subsector: the more concentrated a sub-market, the more its value is locked inside one or two large listed or foreign brand owners; the more fragmented and industrial it is, the more the value sits in private equity and family firms. In every child, the durable profit pool is the installed base — whoever has placed the most machines and runs the densest dealer or service network earns the most defensible, least-cyclical money. (Company-wide revenues for the diversified names span far more than their NAICS-333 lines; treat all scale figures as recent and approximate, and reserve valuation multiples and yields for security-specific work in §10.)

---

5. How the money works

All seven children run the same core model — a cyclical capital-goods maker with a large, higher-margin recurring aftermarket — which is exactly what justifies grouping them. These are capital-goods economics; the regulated-utility rate base, the real-estate trust's funds-from-operations, and the miner's all-in sustaining cost do not apply here. What does:

• New equipment is the "razor." Machines — from a $50 fryer part to a $100-million-plus wafer-fab tool or gas turbine — are cyclical, often engineered-to-order, high in steel content, and sold at modest margins against long lead times and a backlog. Backlog quality — cancellation rights, deposits, escalation clauses, remaining engineering content — matters as much as its size, a caution three children now make explicitly.[4][8][9]
• The aftermarket is the "blade," and the real prize. Parts, consumables, service, rebuilds, and software recur for 7 to 40 years off the installed base. It appears in a different costume in each child — welding consumables, elevator maintenance, engine and turbine service, pump reseals, farm-equipment parts, cutting-tool consumables — but the logic is identical. The revised children put the sharpest anchors yet on it: Otis's service segment earned a 25.1% operating margin in 2025 against 4.8% on new equipment — roughly five times, not four; Konecranes' Industrial Service earned 21.8% against 9.4% on equipment; ESAB's consumables were 66% of sales; Kadant's parts and consumables reached 71%; and mining aftermarket runs about two-thirds of equipment revenue.[3][4][9] The clearest price anyone has put on the annuity sits in 3339: Parker agreed to pay $9.25 billion for Filtration Group, a ~$2.0 billion-revenue business with 85% of revenue from aftermarket — the largest deal named anywhere in the subsector.[9]
• The margin ladder — the subsector's single most useful economic signal, and a finding neither the parent nor any one child previously carried. Lay the children's disclosed segment margins end to end and one rule explains almost all of them. At the top sit consumables, brand, and aftermarket content: ITW's Welding segment at 32.9%, Dover's Pumps & Process Solutions at 30.3%, Ingersoll Rand's industrial segment at 28.9%, ITW's Food Equipment at 27.9%, Alliance Laundry at 25.5% EBITDA, Otis service at 25.1%. In the middle sit component and catalog businesses. At the bottom sit large, one-off, engineered-to-order projects: Konecranes' equipment arm at 9.4%, Otis new equipment at 4.8%, Andritz Metals at 4.5%.[5][9] Critically, concentration does not buy margin. The 3336 primer states it directly: Timken's fragmented Industrial Motion business earned a 19.0% EBITDA margin against GE Vernova's oligopolistic Power segment at 14.7% and Cummins' Engine segment at 12.7%.[8] Across seven independently researched children, the ordering of profitability tracks aftermarket and consumable content almost perfectly — and tracks concentration hardly at all.
• Operating leverage and capacity utilization are the master margin dial. High fixed factory costs mean profits expand fast when lines run full and compress fast when they run soft — visible in 2025 across the whole subsector, from Deere's Production & Precision Ag margin falling to 15.4% from 21.7% and Caterpillar's Construction Industries to 18.7% from 24.2%, to Hyster-Yale swinging to an operating loss and Lennox's Home Comfort volume falling 17%.[3][6][9] And in downturns, makers build fewer machines than customers buy to draw down channel inventory — the defining move of a machinery downcycle.
• Steel, copper, and aluminum are the dominant inputs, so margins live on the spread between selling price and metal cost — tying the whole subsector to one input cycle and one tariff regime (§7). Pass-through works but never cleanly: Lincoln Electric's 2025 pricing added 6.2 points of growth against 3.7 points of volume decline and gross margin still fell.[9]
• Captive finance and the roll-up model. The heavy-equipment majors (Deere, Caterpillar, CNH, AGCO) run in-house lenders that keep machines moving when bank credit tightens; and because most sub-markets are fragmented, the highest-return strategy — public and private alike — has been the decentralized compounder / buy-and-build: acquire niche leaders, keep the aftermarket, compound the cash into the next deal.

Metrics to watch, common to all seven: order intake and book-to-bill (new orders ÷ revenue; above 1.0 means the pipeline is growing — the leading cycle indicator), backlog and its quality, aftermarket / recurring-revenue mix (the higher, the more durable the earnings), capacity utilization and incremental margins, and channel inventory (which can mask true end demand). One caution the metalworking child raises applies everywhere: orders are not earnings — its lone listed machine-tool pure-play posted negative operating margins in two consecutive fiscal years while U.S. machine-tool orders rose 22.5%, because bookings lead shipments by months.[7] EBITDA — earnings before interest, taxes, depreciation, and amortization — is the usual proxy for operating cash generation.

---

6. Demand drivers

Demand across the subsector is derived — it follows how much the economy is farming, building, extracting, processing, cooling, powering, or moving — but the specific trigger differs by child, and that divergence is the point of the rollup:

• Shared master cycle: industrial capital spending and factory utilization. The manufacturing PMI (Purchasing Managers' Index; above 50 signals expansion) is the standard gauge; it read 48.2 in November 2025 after a soft year, then returned to expansion at 52.7 in March 2026.[9] Interest rates move the whole subsector at once, since most end demand is financed capital equipment — which is precisely why the children's independent cycles can still compress together.
• The AI / data-center build-out is the loudest structural force of 2026, and it lands on multiple children at once: gas turbines and gensets (3336) to generate power, chillers and air handling (3334) to cool the racks, wafer-fab tools (3332) to build the chips, and cranes, pumps, and fluid power (3339) to construct and run the facilities. The anchors are now concrete: U.S. electricity demand grew ~1.7% a year between 2020 and 2025 against ~0.1% a year from 2005 to 2019; GE Vernova's gas-turbine backlog reached ~100 GW against ~10 GW of annual output, with turbine prices projected up ~195% from 2019 levels and lead times of five to six years; U.S. data-center diesel generator capacity roughly tripled from ~20 GW in 2018 to ~55 GW in 2024; and U.S. data-center construction rose from ~$9.5 billion to ~$47 billion annualized.[8][9] But the children genuinely disagree about how deep the pull reaches. Inside 3334, the AC side treats liquid cooling as its fastest-growing slice while the fan side warns that rising rack densities and direct-to-chip cooling may reduce fan content per unit of computing; inside 3339, the pump-and-compressor child warns the theme "is real but easily overstated" because compressors are a small slice of any data-center bill of materials, while the material-handling and all-other children treat it as a live order-book driver today.[6][9] Treat it as a mix shift that moves content between sub-markets, not a rising tide.
• Reshoring and new-factory construction — a real tailwind, but the evidence is more mixed than this page previously implied. Roughly $1.66 trillion of U.S. manufacturing investment has been announced since January 2025, including nearly $450 billion in semiconductors and electronics — with the standing caveat that announcements are not completed spending.[4] Against that, the 3339 primer notes U.S. factory construction has fallen since 2024, so the data-center wave, not reshoring, is doing more of the visible work right now.[9]
• Replacement of an aging installed base — the most durable, counter-cyclical driver in every child: roughly two-thirds of hydronic-heating demand and "most" of AC demand, 35–40% of mechanical power-transmission demand, a global stock of modernization-ready elevators projected to grow from ~9 million toward ~13 million by 2030, and an eight-million-machine commercial-laundry base on a 7-to-13-year cycle.[5][6][8][9]
• The skilled-labor picture — real, but narrower than the slogan. The subsector's automation thesis rests on labor scarcity, and the revised children split on how strong that claim is. Genuine scarcity: semiconductor, where one study projected 67,000 expected new U.S. jobs could go unfilled by 2030; the licensed elevator trades, with only ~2,000 openings a year against high licensing barriers; and 95% of surveyed consumer-packaged-goods companies reporting trouble hiring skilled operators.[4][9] But 3339 is explicit that the welding "shortage" headline is overstated — BLS counted 457,300 welders in 2024 and projects 467,200 in 2034, just 2.2% net growth, with openings driven mainly by replacement — and recommends underwriting automation on payback math, not a shortage narrative.[9] Machinists and tool-and-die makers show the same pattern: ~34,200 annual openings against an 11% projected decline in tool-and-die employment.[3][7]
• Child-specific drivers: net farm income (USDA forecasts 2026 net farm income of $153.4 billion, down 0.7%, with farm working capital down 9.2%) and the ~$1.2 trillion Infrastructure Investment and Jobs Act for 3331; chip capex — projected near $200 billion in 2026, about 79% concentrated in China, Taiwan, and Korea — for 3332; restaurant, hotel, and retail formation for 3333 (U.S. food-away-from-home spending hit $1.52 trillion in 2024); weather, building codes, and refrigerant rules for 3334; automotive model changeovers for 3335 (autos are ~39% of 2025 metalworking-machinery demand); electricity load growth and freight for 3336; water and energy infrastructure (~$50 billion of IIJA money against EPA-assessed 20-year needs of $625 billion for drinking water and $630 billion for clean water) and e-commerce warehouse automation for 3339.[3][4][5][6][7][8][9]

Three demand headwinds run against the grain. Electrification slowly erodes the internal-combustion engine (3336) and gasoline fuel dispensers (in 3339). Electric actuators are gradually substituting for hydraulic fluid power (in 3339) — though the revised child is more careful than this page previously was: electrification in heavy mobile equipment increasingly means an electric prime mover paired with more sensorized electro-hydraulics, so dollar content per machine can rise even as unit substitution proceeds. And a headwind the parent previously lacked entirely: the electric-vehicle transition cuts metalworking content per vehicle, with EV programs spending roughly 30% less on tooling and using about 25–30% less carbide cutting tool per vehicle — two independent measurements of the same haircut inside 3335.[7][8][9]

---

7. Regulation

At the subsector level this is a lightly regulated, standards-and-codes-driven set of industries. There is no economic price or entry regulation, no licensing regime, and no rate base — the utility, REIT, and mining frameworks simply do not apply. Oversight is indirect and, tellingly, pushes most of the children the same way: toward cleaner, more efficient, safer machines.

• Product and worker safety (universal). U.S. Occupational Safety and Health Administration (OSHA) rules and third-party listings (UL, NSF, CE marking) apply across the subsector — lockout/tagout for hazardous stored energy is the one standard that reaches into essentially every child — layered with child-specific consensus codes: ASME (American Society of Mechanical Engineers) A17.1 for elevators, B20.1 for conveyors, the B30 series for cranes and B56.1 for forklifts; NFPA (National Fire Protection Association) 86 for furnaces and 13 for sprinklers; AGMA standards for gears; MSHA approval under 30 CFR Part 7 for underground mining gear and API blowout-preventer standards for offshore oil-and-gas equipment; 3-A sanitary standards for food-contact machinery. Compliance mandates recurring inspection and rebuilds, which feed the aftermarket annuity and raise the entry bar for incumbents.[3][4][5][6][8][9]
• Emissions and efficiency — a demand catalyst, and the federal push has retreated unevenly. On the retreating side: the Department of Energy's proposed 95% AFUE all-condensing boiler standard was withdrawn in January 2025, its fan-efficiency standard was withdrawn and it has proposed dropping fans and blowers as "covered equipment" entirely, and the One Big Beautiful Bill (July 2025) terminated the Section 25C consumer credit for equipment placed in service after December 31, 2025, alongside 45L and 179D.[6] On the advancing side: the EPA kept its model-year 2027 heavy-duty NOx standard (0.035 g/hp-hr) on schedule despite industry requests for delay, finalized tighter new-source performance standards for stationary combustion turbines in January 2026, and the AIM Act refrigerant phase-down took effect on January 1, 2025 with a global-warming-potential limit of 700 (variable-refrigerant-flow systems following in 2027).[6][8] DOE efficiency minimums for pumps, compressors, and circulator pumps run out across the decade on a staggered schedule.[9] The parallel 95% AFUE furnace rule was upheld on appeal and is now the subject of a Supreme Court petition — unresolved, not settled.[6] Net: the mandatory push has weakened in buildings and strengthened in engines and turbines, while state codes, energy prices, and payback economics keep the efficiency mix-shift intact everywhere.
• Semiconductor is the one child where policy directly gates revenue: U.S. Bureau of Industry and Security (BIS) export controls from October 2022 through August 2025 progressively restricted advanced tool sales to China, with the Netherlands adding metrology and inspection systems in January 2025 and the EU incorporating those controls in November 2025 — while the CHIPS and Science Act (~$52.7 billion, with ~$33.7 billion in direct awards finalized by early 2025 plus a 25% equipment tax credit) pulls domestic fab construction and tool orders forward (3332).[4]
• Trade policy — the one lever that moves the whole subsector together, now fully quantified. Because every child is metal-intensive, Section 232 tariffs on steel and aluminum — raised from 25% to 50% in June 2025, extended to copper in July 2025 and to hundreds of derivative product categories in August 2025 (including machining centers for working metal), and from 2026 assessed on the full customs value of covered derivatives rather than their metal content — raise input costs across all seven at once, with 25% on derivative articles and a further strengthening action in April 2026. Section 301 duties running from 7.5% to 100% on various Chinese categories layer on top, hitting hoists, forklifts, batteries, hydraulic components, power tools, pulp-and-paper machinery, and tooling; a late-2025 U.S.–China agreement rolled back roughly 10 percentage points of certain Section 301 duties, and targeted Section 232 relief for construction and agriculture equipment arrived in mid-2026. A cost headwind on inputs, but demand support for domestic fabrication.[3][4][5][6][7][8][9]
• Child-specific fights include agriculture's right-to-repair settlement with Deere — an FTC-and-states agreement concluded in July 2026 requiring ten years of equal access to diagnostic software and repair tools for farmers and independent shops, which reshapes high-margin farm-equipment service economics.[3] Packaging is the one place where regulation most directly buys equipment: the Drug Supply Chain Security Act forces unit-level serialization on every pharma line, seven U.S. states have enacted Extended Producer Responsibility laws, and the EU's Packaging and Packaging Waste Regulation applies from August 2026.[9]

Autonomy and AI — driverless tractors, autonomous haul trucks (roughly 3,832 running by mid-2025), robotic welders and warehouses — run ahead of settled regulation everywhere in the subsector.[3][9]

---

8. Consolidation

The subsector's headline concentration says "wide open," but that is the aggregation artifact of §2 and §3. The real story is child by child, and it is active consolidation everywhere, driven by one identical strategic logic: own more of the aftermarket, because service, parts, and consumables annuities are worth more than one-time machine sales. The structure repeats — concentrated at the top, fragmented at the bottom — a handful of majors owning the hardest-to-build, highest-value machines above a long tail of small component, tooling, and service shops.

Consolidation has been unusually heavy in 2025–26, and a second pattern runs alongside it that only a rollup can see: much of the resulting ownership landed in private hands.

• Cross-market strategic deals: Deere's ~$5.2B purchase of Wirtgen moved it into construction; Caterpillar's Bucyrus and Komatsu's Joy Global moved both deep into mining (3331); GE Vernova and Siemens Energy were created by splitting diversified parents, and Midera was spun out of Middleby in July 2026 (3332, 3333, 3336).[3][4][5][8]
• The largest recent transactions: KONE's ~€29.4B agreement to combine with TK Elevator (April 2026, a landmark private-equity exit for Advent and Cinven, closing no earlier than Q2 2027) and Parker's $9.25B agreement to buy Filtration Group — the biggest deal named anywhere in the subsector, and a direct bet on aftermarket economics, since 85% of the target's revenue is aftermarket (3339). Also: Columbus McKinnon's ~$2.8B Kito Crosby deal, completed February 2026 only after a DOJ-mandated divestiture — the first real antitrust bite in this subsector; Johnson Controls' $8.1B sale of residential HVAC to Bosch (completed August 2025) and Carrier's $775M sale of commercial refrigeration to Haier (3334); Baker Hughes' completion of the Chart Industries acquisition in July 2026, after Flowserve's competing ~$19B merger was terminated; Axcelis-Veeco at ~$4.4B and the JBT–Marel merger (3332); RBC Bearings' ~$2.9B purchase of ABB's Dodge business and Regal Rexnord's ~$5B Altra deal (3336).[4][6][8][9]
• Assets leaving public markets: Toyota Industries — the world's largest lift-truck maker — taken private at roughly $39 billion (delisting effective June 2026), Mitsubishi Logisnext and Fujitec delisted in the same year, Honeywell's Intelligrated, Trew, and Transnorm conveyor businesses sold to American Industrial Partners, and Cantaloupe agreed to be acquired out of the vending-technology market.[5][9]
• Private-equity roll-ups dominate the fragmented tails — packaging and furnace builders (ProMach, Duravant, Syntegon, IMA), plastics machinery (Bain's majority in Milacron), independent equipment dealers and rental fleets, HVAC and elevator service networks, and the succession-driven wave of aging-owner mold, die, and tool shops in metalworking, where roughly three-quarters of tool-and-die makers are over 45 and about 40% are retirement-eligible within five to seven years.[3][4][5][6][7][9]

Barriers to entry are steep and widening: capital-heavy plants, decades-deep dealer and service networks, brand loyalty, certification hurdles, and a growing technology moat as incumbents buy GPS, sensor, autonomy, and controls capability to avoid disintermediation by software players. The sharpest competitive challenge across the subsector is low-cost foreign entrants, above all Chinese OEMs climbing the global rankings in construction (XCMG ~$12.8B, SANY ~$10.8B against Komatsu's ~$26.6B), forklifts, and machine tools — and the metalworking child is blunt that low concentration there means many survivors, not a healthy field: its domestic mold base has roughly halved in plants since 2002.[3][7] U.S. antitrust agencies have signaled heightened scrutiny of serial "roll-up" strategies, and the Kito Crosby divestiture shows the teeth are real, though many individual deals fall below merger-notification thresholds. Net: the subsector will stay statistically fragmented for years simply because it spans so many separate niches, but within those niches consolidation is the persistent direction of travel — and increasingly the buyer is not a listed company.

---

9. Risks

Shared across all seven children:

• Deep cyclicality (the defining risk) — and the diversification is weaker than the rollup implies. As derivatives of industrial capex, construction, energy, farm income, and commodity prices, new-equipment orders swing hard, and operating leverage magnifies the profit hit. The children usually trough at different times, which is the core reason to hold the subsector rather than one child — but 3331's revised primer is the cautionary case: its three supposedly independent cycles compressed together in 2025 under a common rate-and-tariff shock, and it explicitly tells readers not to underwrite it as internally hedged.[3] Apply the same discount here: interest rates and tariffs can pressure all seven at once, and in 2025 they did.
• Input-cost and tariff whiplash, now with named bills. Steel, copper, aluminum, motors, and electronics dominate every cost base; Section 232/301 tariffs squeeze margins on fixed-price, engineered-to-order backlogs faster than firms can reprice, and raise customers' project costs, delaying orders. Disclosed 2025–26 impacts run from Deere's ~$600M rising to ~$1.2B and Caterpillar's $2.1B of unfavorable manufacturing costs, to GE Vernova's $250–350M estimate, Hyster-Yale's ~$100M, Otis's ~$20M, and Alliance Laundry's ~$12M in North America.[3][5][6][7][8][9][11]
• Import competition and currency. Formidable German, Japanese, Swiss, Nordic, Dutch, and increasingly Chinese makers set the bar in every child, and a strong dollar whipsaws competitiveness; import dependence is extreme in power tools (~85%) and material in plastics equipment (imports at 68.8% of domestic shipments), machine tools, pumps, forklifts, and construction equipment. Even the "clean domestic play" carries currency risk — ESAB books roughly 80% of sales outside the United States.[3][4][7][9]
• Channel-inventory whiplash. Overbuilding into a downturn forces painful destocking and production cuts across every child — visible in the farm-maker inventory drawdown from a ~$7.2 billion peak to ~$5.7 billion, and in the forklift backlog falling from ~$1.93 billion to $1.28 billion.[3][9]
• Aftermarket erosion. The whole quality thesis rests on retaining the service/parts/consumables annuity; generic "will-fit" parts, independent service (independents hold an estimated 30–55% of North American elevator service units), non-proprietary controllers, and right-to-repair — now a settled ten-year obligation on Deere rather than a pending threat — attack the highest-margin stream.[3][9]
• Customer, channel, and supply concentration. The children now quantify what the parent previously left general: in power tools, two customers were ~27% of Stanley Black & Decker's 2025 sales and one customer ~45.4% of Techtronic's revenue; Applied Materials reported two customers at ~19% and ~15%; the aftermarket distribution channel is itself consolidating; and one instrument maker sourced ~29% of global production from China.[4][8][9]
• Technology-transition risk cutting both ways. Value is shifting from steel toward software, controls, robotics, and electrification; hardware-only makers risk commoditization, and electrification and automation can reduce how many engines, dispensers, hydraulic units, cutting tools, or trucks a customer needs — with the EV transition cutting per-vehicle tooling and cutting-tool content by roughly a quarter to a third.[7][8][9]
• For the public-market investor specifically: the toolkit is thin, uneven, and in places shrinking. An entire child (metalworking, 3335) has no clean pure-play; several others offer at most one or two, with the best packaging, furnace, boiler, fan, food, paper, and conveyor assets private or foreign — and four listed routes closed in 2026 alone. Single-name exposure carries concentrated company-specific risk (integration and leverage at Columbus McKinnon, deep cyclicality at Hyster-Yale, a diluted theme inside every diversified compounder) plus currency and governance risk abroad.[9]
• Reading the federal data literally. Sizing "machinery" from the ~$445B factory figure, or inferring competition from the group HHI of 67.7, mistakes the domestic-production slice for the market and pooled codes for real concentration (§3) — and a code is never the denominator for a global company's market share.

---

10. How to invest, and the outlook

How to invest — match the route to the child. There is no single ticker or fund for NAICS 333, so exposure is assembled, and the form is dictated by which cycle you want:

• For clean listed pure-plays and large-cap anchors: the heavy-equipment champions Deere (DE) and Caterpillar (CAT) for the ag/construction/mining cycle, with the equipment-rental names (United Rentals URI, Herc HRI, Ashtead) as a sideways route reaching roughly 60% of the U.S. construction fleet; the semiconductor-tool leaders (AMAT, LRCX, KLAC) — plus broad chip ETFs (SOXX, SMH) — for the AI-capex cycle; GE Vernova (GEV) and Cummins (CMI) for electricity load growth; Otis (OTIS), Columbus McKinnon (CMCO), Hyster-Yale (HY), Lincoln Electric (LECO), and ESAB for material handling and welding; Middleby (MIDD), Alliance Laundry (ALH), and Tennant (TNC) for commercial and service machinery; Carrier (CARR), Trane (TT), Lennox (LII), and AAON for HVACR; and Kennametal as the one listed name whose core business is genuinely inside metalworking.
• For diffuse, cross-child exposure through diversified compounders: Dover (DOV), Parker Hannifin (PH), IDEX (IEX), Ingersoll Rand (IR), Nordson (NDSN), Graco (GGG), Illinois Tool Works (ITW), Baker Hughes (BKR), and the power-transmission names Regal Rexnord (RRX), Timken (TKR), RBC Bearings (RBC) — each owning a slice of two or more children — plus the aftermarket distributors Applied Industrial Technologies (AIT) and Genuine Parts (GPC) for the recurring layer rather than the factory. Understand you are buying these product lines inside a larger industrial story, that the relevant segment is often a minority of the parent, and that quality compounders often trade at premium multiples (frequently 20–30× earnings), which is itself the main entry-price risk. Compare them on segment organic growth, aftermarket mix, price-cost realization, backlog quality, and segment return on capital rather than consolidated multiples.
• For the pockets that are mostly private or foreign: an entire child — metalworking (3335) — plus large parts of others (HVAC fans and boilers, food/sawmill/paper machinery, packaging, furnaces, the deepest conveyor and forklift assets, and construction and mining equipment) are reached only through private equity, buy-and-build of independent dealers and service networks, family-firm succession deals, private credit, or foreign listings (Komatsu, Epiroc, Sandvik, KONE, Konecranes, Interroll, Krones, Valmet, Andritz, Dürr, KION, Techtronic). Note that Toyota Industries, Mitsubishi Logisnext, and Fujitec are no longer traded and Chart and Honeywell are no longer standalone routes. Broad industrial ETFs (e.g., XLI) give diffuse, single-stock-risk-free exposure to the capital-goods cycle that drives the whole subsector. The diligence questions repeat across children: verify the installed base rather than the brand, split new-equipment from recurring aftermarket revenue, measure technician density near the installed base, test price-cost pass-through on steel and tariffs, and distinguish real orders from cancellable backlog.

Dividends across the subsector are modest to moderate — this is a total-return, cyclically-timed group, not an income one. As cyclicals, most of these stocks look "cheapest" on trailing earnings at cycle peaks and "expensive" at troughs, so through-cycle earnings power and entry point matter as much as company quality. Reserve valuation work (multiples, yields, free cash flow) for the individual name.

Outlook (forward-looking judgment, not fact). As of 2026 the seven children are out of phase on demand, though not on costs — the subsector's most useful feature, held with the caveat §1 and §9 attach to it. The strongest stories are the two riding the AI/data-center wave: semiconductor tools (3332), where global equipment sales are forecast to rise from ~$135B in 2025 to ~$139B in 2026 and ~$156B in 2027, and gas turbines (3336), where a ~100 GW backlog against ~10 GW of annual output has handed makers pricing power and five-to-six-year lead times — with cranes, pumps, and cooling (3339, 3334) pulled along behind them, though the children disagree about how deep that pull reaches. The recovering stories are the industrial-capex-levered children (3339 and the capital-goods clock inside 3335), turning up off a soft 2025 as the PMI moved from 48.2 to 52.7 — unevenly, with conveyor orders inflecting while forklift backlog still shrinks, and with the reminder that orders are not earnings. The lagging stories are agriculture inside 3331, sitting in a genuine downturn — historically the setup for the next up-leg as fleets age, not a permanent state — and, newly, HVACR (3334), which is not the uncomplicated defensive base this page previously described: residential unit shipments are down roughly a fifth year over year, the Section 25C credit expired at the end of 2025, and refrigerant-transition costs are still working through prices. Its replacement floor and data-center optionality are intact; its near term is an air pocket. Construction machinery is the one place the evidence genuinely conflicts, and we do not resolve it: global forecasts read 2025 as the bottom while U.S. factory shipments for the same code rose to $49.2 billion — read it as domestic volume holding with margins under pressure, not a clean turn.[3][4][6][7][8][9]

The structural case is stronger than any single cycle and points the same way for most of the subsector: data-center and power build-out, factory construction, automation whose payback math works on its own terms, water and energy infrastructure, code-mandated safety, efficiency, serialization and environmental equipment, the long runway in electrifying industrial process heat, and a fat, defensive aftermarket that cushions every downturn. Independent forecasts put broadly mid-single-digit annual growth on most of the underlying markets — packaging fastest at ~8% by 2027, pumps and compressors at 4–6%, furnaces at 4–5%, fluid power at ~4–4.5%, welding low-to-mid single digit with robotic welding compounding faster — with the commercial-and-service child slower still (BLS projects ~1.2% real annual output growth for it through 2034), pockets much faster (semiconductor tools, gas turbines, elevator modernization, data-center liquid cooling), and a few genuine structural drags (internal-combustion engines, hydraulic fluid power, gasoline dispensers, and metalworking content per electric vehicle).[3][4][5][6][7][8][9] Across all seven, the swing factor is the same: whether the incumbents convert one-time hardware buyers into recurring precision, autonomy, connected-machine, and software customers. The margin ladder in §5 says why that matters — profitability across this subsector tracks aftermarket and consumable content far more reliably than it tracks market share. If they convert, through-cycle margins and valuations re-rate upward; if low-cost imports, right-to-repair, and technology-transition costs win the margin fight, they do not.

Bottom line: NAICS 333 is a ~$445 billion, ~1.09 million-worker slice of U.S. manufacturing best understood as seven adjacent machine-building families sharing one business model — picks-and-shovels equipment, razor-and-blade aftermarkets, steel-and-tariff input exposure, and consolidation that keeps eating the fragmented tail — riding seven different demand cycles that rarely peak together but sharing one cost base that can squeeze them all at once. The recurring caveats are identical across all of them: read concentration two or three levels down and then check what the code actually contains, treat the ~$445 billion as domestic production rather than the served market, weight aftermarket mix over headline share when judging quality, and expect to assemble your exposure — because no single stock owns this subsector, one entire child of it is barely public at all, and the listed set is getting smaller in places rather than larger. These are judgments about direction, not guarantees; the industrial capex cycle, tariffs, rates, China, and antitrust rulings could all move the timeline.

---

Sources

Synthesized from the seven already-written child primers plus our ingested U.S. federal statistics for NAICS 333. Company-, market-, and policy-level facts are carried up unchanged from the child primers.

1. U.S. Census Bureau, 2022 Economic Census — Concentration by Largest Firms, NAICS 333 (Machinery Manufacturing): receipts $445.52B; 18,779 firms; CR4 12.5%, CR8 16.3%, CR20 24.4%, CR50 35.8%; HHI 67.7. Histometrics-ingested federal ground-truth (stats-333.md). https://www.census.gov/programs-surveys/economic-census.html
2. U.S. Census Bureau, County Business Patterns 2023, NAICS 333: employment 1,086,146; establishments 21,668; annual payroll $85.06B; Q1 payroll $21.83B. Histometrics-ingested federal ground-truth (stats-333.md). https://www.census.gov/programs-surveys/cbp.html
3. Histometrics child primer — NAICS 3331, Agriculture, Construction, and Mining Machinery Manufacturing (2022 receipts $109.92B; 2023 employment 193,869; CR4 40.0%, HHI suppressed; children 33311/33312/33313; the 2025 cycle-alignment finding — farm shipments −15.8% to ~$30.8B, Deere P&PA margin 15.4% vs 21.7%, CAT Construction Industries 18.7% vs 24.2%, U.S. upstream investment −6% to ~$420B; U.S. construction-machinery shipments $49.2B in 2025 against a global forecast calling 2025 the bottom; Epiroc 66/34 aftermarket mix and sponsored ADRs; rental firms owning ~60% of the U.S. fleet; Section 232 25%→50% and Deere/AGCO tariff costs; FTC right-to-repair settlement July 2026). Underlying: 2022 Economic Census / CBP 2023.
4. Histometrics child primer — NAICS 3332, Industrial Machinery Manufacturing (2022 receipts ~$40.23B; 2023 employment 111,497; CR4 20.8%, HHI 152.7; single child 33324; semiconductor ~34% of receipts at HHI 1,153 and ~$151k pay, all-other ~40% at HHI 66, food ~16%, sawmill/paper ~10%; AMAT $28.4B, SEMI equipment billings ~$135B/$139B/$156B, ~$200B 2026 chip capex, ~79% China/Taiwan/Korea; plastics-equipment imports at 68.8% of domestic shipments; Kadant 71% recurring; Midera spin July 2026; CHIPS Act and BIS export controls; ~$1.66T announced U.S. manufacturing investment). Underlying: 2022 Economic Census / CBP 2023.
5. Histometrics child primer — NAICS 3333, Commercial and Service Industry Machinery Manufacturing (2022 receipts ~$32.06B; 2023 employment 80,504; CR4 15.5%, HHI 111.1; single child 33331; three near-pure-plays — Middleby ~$2.35B segment revenue at 40.2% gross margin and becoming a foodservice pure-play, Alliance Laundry $1.71B at 25.5% adjusted EBITDA with an ~8-million-machine installed base, Tennant ~$1.29B; ITW Food Equipment ~$2.7B at 27.9%; Xerox pro-forma revenue −7.6%; food-away-from-home $1.52T; BLS ~1.2% real output growth to 2034). Underlying: 2022 Economic Census / CBP 2023.
6. Histometrics child primer — NAICS 3334, Ventilation, Heating, Air-Conditioning & Commercial Refrigeration Equipment Manufacturing (2022 receipts ~$55.03B; 2023 employment 147,794; CR4 22%, HHI 199.4; single child 33341 splitting ~73%/~15%/~12%; AHRI shipments −19.9% y/y through November 2025; Section 25C termination, withdrawn boiler and fan standards, AIM Act GWP-700 limit and SEER2; Carrier $21.7B 2025 sales with 28% parts and service, Trane ~$21.3B split $13.98B product / $7.34B service, AAON backlog $1.83B vs $867M; JCI residential HVAC sold to Bosch for $8.1B; the fan-versus-AC disagreement on data-center content). Underlying: 2022 Economic Census / CBP 2023.
7. Histometrics child primer — NAICS 3335, Metalworking Machinery Manufacturing (2022 receipts ~$33.87B; 2023 employment 134,208; CR4 9.4%, HHI 37.2; single child 33351 splitting into five sub-industries on three clocks; machine-tool orders $5.74B in 2025 (+22.5%) and ~+32% in early 2026, cutting-tool shipments $2.56B (+2.5%), bespoke tooling soft-to-flat; Kennametal ~$1.97B with a $1.22B Metal Cutting segment; EV tooling spend ~30% lower and carbide use ~25–30% lower; machine-tool consumption ~$10.5B vs ~$5.9B domestic production; Section 232/301 actions including 407 derivative categories; "orders are not earnings"). Underlying: 2022 Economic Census / CBP 2023.
8. Histometrics child primer — NAICS 3336, Engine, Turbine, and Power Transmission Equipment Manufacturing (2022 receipts ~$49.67B; 2023 employment 96,251; CR4 42.8%, HHI 589.4; single child 33361 splitting into engines ~57%, turbines ~24%, mechanical power transmission ~11%, gears ~8%; GE Vernova backlog ~83→~100 GW against ~10 GW output, Power 14.7% vs Wind −6.6% on a $598M loss, $250–350M 2026 tariff estimate; Cummins N.A. heavy-duty −27%, Caterpillar power generation +32% and Power & Energy $32.2B; Timken Industrial Motion 19.0% vs GEV Power 14.7% — concentration does not buy margin; U.S. electricity demand +1.7%/yr; EPA MY2027 NOx and January 2026 turbine NSPS; 45X wind repeal after 2027). Underlying: 2022 Economic Census / CBP 2023.
9. Histometrics child primer — NAICS 3339, Other General Purpose Machinery Manufacturing (2022 receipts ~$124.74B; 2023 employment 322,023; CR4 7.2%, HHI 39.4; children 33391 ~21%, 33392 ~33%, 33399 ~46%; the margin ladder tracking aftermarket content; Otis service 25.1% vs new equipment 4.8% and ~$14.4B global sales, Konecranes 21.8% vs 9.4%, ESAB 66% consumables, ITW Welding 32.9%; Parker/Filtration Group $9.25B at 85% aftermarket; KONE–TK Elevator ~€29.4B, Columbus McKinnon–Kito Crosby ~$2.8B with DOJ divestiture, Baker Hughes–Chart, Honeywell–AIP, Toyota Industries/Logisnext/Fujitec delistings; PMI 48.2→52.7; BLS welder projections qualifying the shortage narrative; the both-directions undercount correction). Underlying: 2022 Economic Census / CBP 2023.
10. Deere & Company, "Deere Reports Net Income of $5.027 Billion for Fiscal Year 2025" (net sales & revenues ~$45.7B) — via child primer [3]. https://www.prnewswire.com/news-releases/deere-reports-net-income-of-1-065-billion-for-fourth-quarter-5-027-billion-for-fiscal-year-302626652.html
11. Caterpillar Inc., 2025 Form 10-K (total revenue ~$65B; Construction Industries $25.1B at an 18.7% segment margin vs 24.2%; Power & Energy $32.2B with power-generation sales +32%; $2.1B unfavorable manufacturing costs and $817M unfavorable price realization, much attributed to tariffs) — via child primers [3] and [8]. https://www.sec.gov/Archives/edgar/data/18230/000001823026000008/cat-20251231.htm