Metalworking Machinery Manufacturing (United States)

NAICS 2022 code 33351 — a rollup investor primer covering five child industries: 333511, 333514, 333515, 333517, 333519

1. Overview

Metalworking machinery manufacturing is the tool-building layer that sits underneath almost all durable-goods production. Before a factory can stamp a car door, mold a detergent bottle, machine a jet-engine bracket, or roll a coil of sheet steel, someone has to build the machine, the tool, or the mold that does the work. NAICS (North American Industry Classification System, the U.S. government's official industry taxonomy) groups five of those tool- and machine-building trades under one five-digit code, 33351. Together they are the "picks-and-shovels" of American manufacturing. [1][2]

The distinctive thing about this level is that its five children look similar on the surface — all small, all skilled-labor-heavy, all cyclical, all overwhelmingly private — yet they run on three different economic clocks:

• Bespoke project tooling — industrial molds (333511) and special dies, jigs, and fixtures (333514). One-off, engineer-to-order tools bought before a new product goes into production. Demand is a leading signal of new-product launches.
• Consumables — cutting tools and machine-tool accessories (333515). Drills and inserts that wear out and get replaced during production. Demand tracks how hard existing machines are running — factory utilization right now.
• Capital-goods machine builders — machine tools (333517) and rolling mills and other metalworking machinery (333519). Big-ticket machines bought when customers expand or modernize capacity. Demand is a leading signal of manufacturing capital spending (capex).

For an investor, the through-line is the same across all five: there is no clean, liquid U.S.-listed pure-play in any of them, the real ownership is private, and the public routes are indirect (adjacent suppliers, foreign-listed builders, diversified industrials). But which clock a child runs on — launches, utilization, or capex — determines when it makes money and how to read it. A second, separate question is how much of each child is visible in public markets at all, and as §4 shows, that turns out not to line up with the clocks. This primer leads with both contrasts, then covers the level as a whole.

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

Each child is profiled in its own primer; the value here is the side-by-side. Federal figures below are for each specific code (receipts from the 2022 Economic Census; employment from County Business Patterns, "CBP," 2023). "Share of level" is that child's slice of the $33.87 billion rollup. [1][2]

Shares sum to ~100% (receipts 18+16+27+9+29 ≈ 99%; employment 23+18+21+8+30 = 100%, rounding).

How to read the table. Three fault lines run through this level:

1. Project vs. consumable vs. capex. Molds and dies (333511, 333514) are lumpy, one-off projects — revenue disappears when the tool ships unless the customer reorders. Cutting tools (333515) are the opposite: consumables that generate recurring, razor-and-blades revenue every time a factory runs. Machine tools and rolling mills (333517, 333519) are big-ticket capital goods — a single machine can run from $100,000 to well over $1 million, and a mill line or roll contract can be worth millions, sold to order with long lead times. A single custom injection mold spans $15,000 to $8 million, with U.S.-built custom molds averaging more than $100,000. [4][5][6][7]

2. Where in the cycle each turns first — and how long the lag runs. Because tooling is bought before production, molds and dies lead product launches. Because cutting tools wear during production, they lead current factory activity. Because machines are bought to add capacity, machine tools and rolling mills lead capital investment. An investor watching the whole level effectively has three different leading indicators, not one — and the capex clock has the longest lag between the signal and the cash. Hurco, the one U.S.-listed machine-tool pure-play, posted negative operating margins in both fiscal 2024 (−4%) and fiscal 2025 (−6%) while U.S. machine-tool orders were rising 22.5%, because orders book months ahead of shipments and its revenue was still working through the prior trough. Do not read an order series as a current-quarter earnings signal. [4][6]

3. Domestic fragmentation vs. global concentration — and the two do line up. All five are hyper-fragmented inside the U.S. (see §3, §8). But the global competitive picture splits, and the domestic concentration numbers turn out to be a weak signal of it: the two children that are fragmented worldwide have the two lowest domestic CR4s (molds 9.1%, dies 16.6%), while the three dominated by a short list of large foreign majors have the three highest (cutting tools 17.6%, rolling mills 19.3%, machine tools 28.6%). Sandvik, Berkshire's ISCAR, and Kennametal lead cutting tools; DMG Mori, Mazak, Haas, and DN Solutions lead machine tools; SMS, Danieli, Primetals, Andritz, and Fives lead rolling mills. Even so, the low U.S. numbers badly understate how concentrated those three global markets really are, because foreign giants appear only through their U.S. establishments. [2][4][6][7]

What this level excludes (adjacent codes an investor should not conflate): the molders who use molds to make plastic parts (NAICS 326), the foundries that pour metal into dies (NAICS 331), and the machine shops (NAICS 332710) that operate machine tools to make parts. Those customer industries are far larger; the 33351 group sells to them. [3][5]

3. How big it is

Federal statistics (our ground-truth figures for this five-digit level):

Those figures imply an average of roughly $6.4 million of receipts per firm, about 23 employees per establishment, and average pay near $69,000 per worker — well above the manufacturing norm for low-skill assembly, reflecting a machining- and toolmaking-heavy workforce. [1][2] Pay is remarkably uniform across the children, from ~$62,000 in molds to ~$75,400 in machine tools, and the Bureau of Labor Statistics puts the 2024 median wage for tool and die makers at $63,180. [3][4][5][6] For scale, ~$34 billion of receipts and ~134,000 jobs make the entire metalworking-machinery group a small corner of U.S. manufacturing — smaller than most people expect for something so strategically central.

A striking structural fact: the rollup is more fragmented than any of its children. Its HHI of 37.2 sits below every child's — the lowest of the five is industrial molds at 50.4, and the highest is machine tools at 314 — and its four-firm share of 9.4% is below four of the five children, level with the fifth (molds, 9.1%). An HHI under 100 is far below the 1,500 threshold antitrust regulators treat as even "moderately concentrated." The reason is intuitive: combining five separate competitive fields into one union means no single firm can dominate the whole — a die-shop giant is a nobody in cutting tools, and vice versa. This is one of the most dispersed manufacturing categories in the country. [2]

Undercount caveat — every child's federal figure understates something, but for two different reasons. The counts capture establishments primarily engaged in each code, and they measure U.S. producers, not the U.S. market:

• Captive production (all five, worst for tooling). A large volume of tooling is built inside the in-house tool rooms of automakers, appliance makers, molders, and aerospace firms, and is classified under the parent plant's industry rather than 333511/333514/333515. The smallest one- and two-person shops are also easy to miss. So total U.S. mold-, die-, and tool-building activity is larger than the statistic. [3][5][6]
• Imports (worst for capital goods and consumables). Federal figures count only what is built in the U.S., and these markets are import-dominated. U.S. machine-tool consumption was roughly $10.5 billion in 2022 against ~$5.9 billion of domestic production (a 5th-place global ranking), and in 2024 the U.S. imported ~$1.6 billion of machine tools against ~$550 million of exports — a ~$1.1 billion deficit. Third-party consumption estimates put the U.S. cutting-tool market near $15 billion in 2024, roughly three times the $5.55 billion of domestic shipments. And the largest, most advanced rolling-mill lines are engineered abroad by SMS, Danieli, and Primetals, whose sales never enter this code at all. [4][6][7]

Two further reconciliation notes. First, the child firm counts sum to slightly more than the rollup's 5,301 (the five add to ~5,361), because a company active in two child industries is counted once at the five-digit level — a normal deduplication, not an error. Employment, establishments, and receipts sum cleanly. Second, employment depends on which federal survey you use. The table above uses CBP throughout, for comparability; the BLS establishment and industry-productivity series run materially higher for the same codes (machine tools 39,600 in March 2025 versus CBP's 28,524; dies 49,000 in 2025 on an all-persons basis versus CBP's 40,467), because they capture self-employment and use different methodology and timing. Use one series consistently rather than mixing them. [3][4]

Finally, broad market-research figures use much wider scopes than the Census codes and are not contradictions of the federal data: a "U.S. tooling market" at roughly $49–51 billion bundles molds, cutting tools, machine tools, and consumables together, and the consumption estimates above bundle imports. Treat them as different measurements of different things. [3][4][6]

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

The single most important fact for a public-market investor is unchanged: there is no clean, liquid, U.S.-listed pure-play in any of the five children. What has changed is the shape of the indirect menu. Ranked by how directly the listed exposure sits inside the code — not by how many names exist:

• Cutting tools (333515) — the only child with a listed leader whose core business is the industry. Kennametal (NYSE: KMT), ~$1.97 billion in FY2025 sales, runs a Metal Cutting segment of $1.22 billion squarely inside this code (7.1% operating margin in FY2025, down from 10.4%; 60% of consolidated sales are outside the U.S.). Beyond it, the global majors are foreign-listed or private: Sandvik (Nasdaq Stockholm: SAND), whose Machining segment is ~SEK 48.6 billion, Kyocera (Nasdaq ADR: KYOCY), Mitsubishi Materials (Tokyo: 5711), Sumitomo Electric (5802), and OSG (6136), plus Berkshire Hathaway (NYSE: BRK.B), which wholly owns IMC/ISCAR (~$4.1 billion of global revenue in 2025, up 3.9%). Distributors MSC Industrial (NYSE: MSM) and W.W. Grainger (NYSE: GWW) are a "picks-and-shovels-on-the-picks-and-shovels" route. [6]
• Rolling mills (333519) — a small-cap sitting on the in-scope consumable. Ampco-Pittsburgh (NYSE: AP) is the nearest U.S. name and is genuinely in scope: mill rolls are explicitly part of this code, and its Forged & Cast Engineered Products segment (~67% of $418 million of 2024 sales; $273 million of mill-roll sales; $251 million backlog) makes work and back-up rolls, with the remainder an unrelated air-handling business. Park-Ohio (NYSE: PKOH) adds partial exposure via induction heating, forging presses, and forming machines. The global mill-builders are foreign: SMS (private, €4.03 billion of 2024 sales), Danieli (Milan: DAN), Andritz (Vienna: ANDR), Primetals via Mitsubishi Heavy Industries (Tokyo: 7011), and Fives (private). [7]
• Machine tools (333517) — one loss-making micro-cap, then foreign. Hurco (Nasdaq: HURC) is the lone U.S.-listed pure-play builder, at ~$179 million of FY2025 sales and negative operating margins in both of the last two fiscal years — single-name and liquidity risk in full. The better builders are foreign-listed (DMG Mori, Frankfurt/OTC; Okuma, Makino, Amada, JTEKT, Tokyo; Georg Fischer, SIX) or private (Haas, Mazak, Trumpf, DN Solutions, Gleason, and Hardinge's machine-tool lines under Centre Lane Partners). Higher-quality adjacent exposure comes through controls/robotics leader FANUC (Tokyo/OTC: FANUY). [4]
• Industrial molds (333511) — the broadest proxy list, none of them a mold builder. This is where the public menu has grown most: Husky Technologies is now public through GPGI (via SPAC), a global integrated supplier with ~13,500 installed systems and ~65% of revenue from aftermarket molds, hot runners, service, and tooling — broader than the U.S. code and economically unlike a job shop. Alongside it: Hillenbrand (NYSE: HI) for mold components and hot runners (DME, Mold-Masters) and a retained ~49% of Milacron; Proto Labs (NYSE: PRLB), ~$533 million of 2025 revenue with injection molding at $191.5 million (35.9%); Xometry (NASDAQ: XMTR), a marketplace that added automated injection-molding quotes in late 2025; and Core Molding (NYSE American: CMT), a custom molder with ~$30–40 million of tooling revenue. Milacron itself passed to Bain Capital (~51%) in March 2025. [5]
• Special dies and fixtures (333514) — still the thinnest. No U.S. pure-play. Exco Technologies (TSX: XTC), ~C$638 million of FY2024 revenue, remains the nearest listed tooling-focused proxy, with roughly half its business in auto components. MISUMI (TYO: 9962) gives die-component exposure through Dayton Lamina, Dayton Progress, and Anchor Lamina; Mayville Engineering (NYSE: MEC) and Cleveland-Cliffs (NYSE: CLF) build progressive dies and run tooling-and-stamping operations inside far larger fabrication businesses. The archetypal scaled owner is a private-equity roll-up such as Tooling Tech Group. [3]

The pattern across all five — and it is not the one the clocks predict. Public access does not thin out along the launches–utilization–capex axis. It thins out along a different axis: what is listed is almost always the layer beside the shop, not the shop. Mold components and machinery are listed; mold builders are not. Machine-tool controls and cutting tools are listed; the American builder base barely is. Mill rolls are listed; mill lines are not. Only cutting tools has a listed company whose core is the code itself. Everywhere else the bulk of the child is thousands of private, family-owned shops plus captive OEM tool rooms — which is why the real ownership opportunity in this level is private (see §10). None of the public proxies disclose revenue sensitivity to their specific NAICS code. [3][4][5][7]

5. How the money works

Because the level runs on three economic clocks, the money works three ways — and the smartest operators blend them:

• Project tooling (molds, dies) sells projects, not products. Each tool is separately quoted, engineered, and billed — often on a progress schedule (deposit, milestone, delivery) that ties up working capital during a build of weeks to months. Revenue is lumpy and largely non-recurring: when the tool ships, that revenue is gone unless the customer reorders. The cost stack is dominated by skilled labor (toolmakers, mold designers, CNC — computer numerical control — and EDM — electrical-discharge machining — operators), machine time, and tool steel, over heavy fixed overhead. Owners watch backlog and quote-to-win rate (the pipeline), machine/spindle utilization (top shops push above 75%; idle high-cost iron destroys margin), on-time delivery, and sales per employee. Healthy independent shops target roughly 30–60% gross margin on direct job cost; on the bottom line, one benchmark puts top shops near 13.5% net against about 8% for the rest. A rising share of mold shops now rate their own margins "unacceptable." [3][5]
• Consumables (cutting tools) sell the razor-and-blades cycle. A carbide drill or insert has a finite tool life; when it dulls it is replaced (or reground and recoated). Revenue is recurring and tracks machine-spindle-hours running in U.S. factories — i.e., capacity utilization, not machine purchases. The proposition sold is cost per finished part and machine uptime, not price per tool, which is why the majors spend on coatings and geometries. Key inputs are tungsten and cobalt (volatile, China-concentrated critical minerals). Operating leverage is high — carbide-pressing and grinding plants are fixed-cost-heavy — and inventory across tens of thousands of SKUs is central to returns. [6]
• Capital goods (machine tools, rolling mills) sell the iron plus the aftermarket. A big-ticket machine is moderate-margin "iron"; profitability improves with automation content (robot loading, in-machine measurement, software) and, above all, the recurring aftermarket — spare parts, service, retrofits, and consumable rolls. Because machines are built to order, orders (bookings) lead shipments (revenue) by months, so book-to-bill and backlog are the leading health indicators — though not every builder works that way: Hurco builds to stock and normally carries only ~45 days of backlog, which is why forecasting errors and dealer destocking consume its cash quickly. Operating leverage is severe. Contracts often carry raw-material surcharge pass-throughs, so headline sales can move without profit changing — Ampco-Pittsburgh grew Forged & Cast operating income ~38% in 2024 on better pricing and absorption even as volume and surcharge sales fell. Watch base pricing and operating income. [4][7]

Recurring revenue is the real quality gradient, and the children sit at very different points on it. The public proxies disclose recurring shares from roughly a fifth to roughly two-thirds of revenue: Husky ~65% from aftermarket molds, hot runners, service and tooling; Park-Ohio's induction/melting business ~49% from replacement parts and field service; Andritz's Metals segment 27% service against 73% capital equipment; Hurco's service parts and fees $33.9 million, or 19% of FY2025 revenue. Ampco-Pittsburgh is a limiting case — its product is the consumable. These are different companies at different scopes, so read them as a gradient, not a league table. [4][5][7]

Margins do not average into an "industry margin." The disclosed figures across the level are measured differently and at different scopes, and range widely: Danieli's plantmaking EBITDA margin 12.3% (FY2024/25), Kennametal's Metal Cutting operating margin 7.1% (FY2025), Andritz's Metals EBITDA margin 6.7% (2025), Ampco-Pittsburgh's Forged & Cast operating margin 3.7% (2024), Hurco's company operating margin −6% (FY2025) — against the 30–60% gross-on-direct-cost convention used by private job shops, which is not comparable to any of them. Do not blend these into a single number. [3][4][6][7]

The common thread: in every child, customer capacity utilization is the master variable. When factories run hot, they launch products (buy molds and dies), burn through consumables (buy cutting tools), and expand capacity (buy machines and mills). When utilization sags, all three demand streams cool — tooling and capex first, consumables a beat later.

6. What drives demand

Demand across the whole level is derived — it comes from the capital-spending and new-product cycles of the industries that buy from it — which makes the group inherently cyclical. The shared drivers:

• New product launches and model changeovers. The core driver for molds and dies, and it tracks part redesign rather than production tonnage: a long run may reuse a tool for years, while a change in geometry, resin, cavity count, tolerance, or production location forces a new one. Automotive is the dominant end market — roughly 60% of U.S. metal-stamping demand, over 40% of global tooling demand, and ~39% of 2025 metalworking-machinery demand. [3][7]
• Manufacturing capex and the industrial cycle. The core driver for machine tools and rolling mills. Machine-tool orders are one of the cleanest leading indicators of manufacturing capex, and interest rates gate order timing directly because the machines are financed. [4]
• Factory utilization. The core driver for cutting-tool consumables — a near-real-time proxy for how hard existing machines are running, and explicitly treated by AMT as a leading indicator of U.S. manufacturing activity. [6]
• Reshoring and trade policy. Tariffs on imported tooling, longer overseas lead times, and intellectual-property (IP) concerns have steered some sourcing back to domestic shops — a tailwind that cuts across all five children. It is not automatic: hybrid strategies (build the tool overseas, run production domestically, or have a low-cost supplier rough-machine while a domestic shop owns tryout and validation) are common. [3][5]
• Aerospace and defense — the standout in the current up-cycle. Aerospace manufacturing-technology orders rose 45.1% in 2025 versus 2024 and ran 233% above prior-year levels in the first two months of 2026; hard-to-machine alloys (titanium, nickel superalloys) wear cutting tools faster and reward premium grades and coatings, lifting consumable demand as well. [4][6]
• The U.S. steel and aluminum build-out. A specific pull for rolling mills: electric-arc-furnace (EAF) "mini-mill" steel reached ~69% of U.S. output in 2024, Nucor alone committed roughly $3–4 billion of capex across 2024–2026, and Section 232 tariffs have coincided with more than $10 billion in announced new domestic mill investment. Every new melt shop needs rolling and finishing lines and consumable rolls. The World Steel Association projects U.S. steel demand growth of 1.7% in 2026 and 2.0% in 2027 — a precondition for that investment, not a measure of machinery revenue. [7]
• Automation and labor scarcity as a demand driver, not just a constraint. A shortage of skilled machinists pushes buyers toward automated, multi-tasking machines, so order value has been rising faster than order units — buyers are loading machines with robots, material handling, inspection, and software. Automation also raises the premium on predictable tool life, since an unattended cell needs consistent wear behavior. [4][6]
• The EV (electric-vehicle) transition — two-sided, and the children agree on the magnitude. Electrification seeds new tooling programs (battery trays, structural "gigacasting" dies, electrical-steel processing lines) and is projected to reach ~28% of automotive tooling demand by 2030, while retiring others. But independent estimates from two different children converge on the same haircut per vehicle: EV programs spend roughly 30% less on tooling than combustion models, and cemented-carbide cutting-tool use per EV runs about 25–30% lower than for a gasoline vehicle. The mix shift is real; it is not additive. [3][6][7]
• Additive manufacturing — complement more than substitute. All five children now describe 3D printing the same way: it displaces some low-volume tooling and rough machining, while creating new work (conformal-cooling inserts, printed fixtures) and demanding finishing operations, since printed metal parts still need milling, turning, or grinding for mating surfaces and tolerances. [3][5][6]

Where the children diverge is timing and current temperature — and, in one case, they genuinely disagree. As of the latest available data, the capital-goods children (333517, 333519) are in a clear order up-cycle and the consumable child (333515) is modestly growing (+2.5% in 2025). The bespoke-tooling children point in opposite directions: the die primer reports quoting up 20%-plus and utilization climbing toward the high-70s in 2025, while the mold primer reports AMBA's 2026 survey describing shipments, backlog, and quoting as broadly stagnant, with automotive a stalled source of work for some respondents. Both draw on AMBA surveys, but of different vintages and different member populations, and both are sentiment rather than shipments. The hard data cut against the optimistic read: BLS real output for the die industry fell back to 78.9 in 2023 from 88.9 in 2022 (2017 = 100), with unit labor cost at 132.2 — pricing rising while physical output and absorption deteriorated. Treat project tooling as soft-to-flat with improving quotation activity, not as recovering. [3][4][5][6][7]

7. Regulation

None of the five is a licensed or rate-regulated industry; the regulation that moves them is trade policy, export control, and workplace/environmental rules — and it applies unevenly across the children:

• Trade tariffs — the biggest single policy lever, cutting both ways, and now moving. For molds and dies, Section 301 tariffs on Chinese tooling (generally 25%) act as a protective floor; more than 60 U.S. mold manufacturers formally petitioned to keep them. That floor is no longer stable: a late-2025 U.S.–China agreement included a partial (~10 percentage-point) rollback of certain Section 301 duties. For machine tools and rolling mills, Section 232 (national-security) tariffs on steel and aluminum rose from 25% to 50% in June 2025, and in August 2025 Commerce added 407 derivative product categories to the 50% list — including HTS 8457.10.00, "machining centers for working metal" — both raising input costs and pulling customer investment onshore. A separate Section 232 investigation into robots and industrial machinery could extend tariffs further. For cutting tools, tariffs protect finished-tool pricing while raising the cost of imported tungsten-carbide inputs. Policy is in flux in both directions. [3][4][5][7]
• Export controls (dual-use). High-precision machine tools — especially five-axis simultaneous CNC machines — are controlled under U.S. Export Administration Regulations (Export Control Classification Number, or ECCN, 2B001), administered by the Commerce Department's Bureau of Industry and Security (BIS), to prevent weapons and nuclear/missile proliferation; related software and technology can also be controlled. Tooling for defense programs can fall under ITAR (International Traffic in Arms Regulations), and advanced metallurgical equipment can face licensing when sold abroad. Unrecognized export-control exposure is a named diligence trap in machine-tool acquisitions. [3][4][7]
• Critical minerals. Tungsten and cobalt are essential to cutting tools and are China-concentrated critical minerals: USGS reports U.S. net import reliance of 50% for tungsten, with China-sourced imports an estimated 14% of U.S. consumption, and China placed tungsten products under export control in 2025. This is a supply-security and procurement issue more than a product regulation. [6]
• Standard manufacturing oversight — with two enforcement specifics worth knowing. OSHA (Occupational Safety and Health Administration) machine-guarding rules apply throughout (29 CFR 1910.212 names milling machines, shears, presses, and forming rolls), alongside lockout/tagout and metalworking-fluid exposure. OSHA's National Emphasis Program on manufacturing amputations expressly names NAICS 333514, and OSHA's permissible exposure limit for cobalt metal, dust, and fume — a live hazard in carbide grinding, where chronic exposure causes hard-metal lung disease — is 0.1 mg/m³ as an eight-hour time-weighted average. BLS reported a 2022 total-recordable injury rate of 4.2 per 100 full-time workers for 333519 (days-away/restriction/transfer 2.1), elevated relative to manufacturing overall. EPA (Environmental Protection Agency) Metal Products and Machinery effluent guidelines apply to direct dischargers generating oily machining wastewater — process- and facility-specific, not automatic. [3][5][6][7]

Rate-of-return, REIT (real-estate investment trust), or reserve-based regulatory frameworks do not apply to this level — these are lightly product-regulated machinery businesses whose competitive fate is set mainly at the border.

8. Consolidation

Domestically, every child is fragmented and stays that way. The rollup HHI of 37.2 and CR4 of 9.4% (§3) sit at the extreme low end of U.S. manufacturing, and each child individually is unconcentrated — HHIs of 50.4 (molds), 99.9 (dies), 130.2 (cutting tools), 190.5 (rolling mills), and 314 (machine tools), all far under 1,500. Competition is local, relationship-driven, and reputation-based, with barriers built on engineering know-how, installed base, and customer qualification rather than scale. [2]

But fragmentation is not stasis — the domestic shop base has been shrinking for two decades. Industrial mold manufacturing had 2,468 establishments and 46,722 employees in the 2002 Economic Census against roughly 1,250 establishments and 30,855 employees today: about half the plants and two-thirds the employment. The median AMBA-responding mold shop fell from 36 shop employees and 6 design employees in 2015 to about 21 and 4 by 2024, with shop hours barely down (50 to 47). The U.S. machine-tool builder base contracted even more violently: more than two-thirds of U.S. machine-tool firms closed after the early-1980s downturn, ceding global leadership to Japan, Germany, and now China. Low concentration in this level means many survivors, not a healthy field. [4][5]

The shape of consolidation differs sharply across the level, and this is the key contrast:

• Bespoke tooling (molds, dies) consolidates slowly, mainly through succession. The workforce is aging — roughly three-quarters of tool-and-die makers are over 45, about 40% are eligible to retire within 5–7 years, and only ~2% are under 35 — with too few apprentices entering. Many owner-operated shops have no internal successor, producing a steady stream of acquisition targets and outright closures. That drives private-equity roll-ups: Tooling Tech Group in dies (assembled under successive PE owners), Tooling Systems Group (12 companies, 600-plus employees), and SyBridge Technologies in molds (15 acquired manufacturing companies across several countries). Consolidation is far more advanced around moldmaking than within it — Hillenbrand's DME and Mold-Masters, Husky (Platinum Equity, now public via GPGI), and Milacron (Bain Capital, ~51%, March 2025) have concentrated the components-and-machinery layer and continue to rationalize plants. [3][5]
• Consumables and capital goods are already globally consolidated at the top. Cutting tools are a global oligopoly (Sandvik, Berkshire's ISCAR/Tungaloy, Kennametal, Kyocera, Mitsubishi Materials, Sumitomo, OSG, Ceratizit), machine tools are led by a short list of Japanese, German, and Korean OEMs, and large rolling-mill lines are an oligopoly of SMS, Danieli, Primetals, Andritz, and Fives. In all three, advantage concentrates in materials science, controls/software, application engineering, and service networks — favoring scale — while niche domestic shops survive on specials, fast turnaround, and regrind/aftermarket work. Deals keep coming: Korea's DN Solutions completed its acquisition of Germany's Heller in early 2026; Berkshire Partners' Harvey Performance bought ATA in 2024, adding 25,000-plus products and seven plants; Hardinge's machine-tool and workholding lines passed from Privet Fund to Centre Lane Partners in 2024, becoming Kellenberger and Forkardt Hardinge. [4][6][7]

Two structural features add volatility everywhere. Captive OEM tool rooms mean large manufacturers can pull tooling in-house or push it out to merchant shops depending on the cycle. And in the capital-goods children the revenue itself is project-lumpy at a scale that swamps trend: SMS group's metallurgical plant-engineering order intake fell 37.2%, from €3.68 billion to €2.31 billion in 2024, simply because one unusually large prior-year contract did not repeat. [3][7]

9. Risks

Shared across the level:

• Deep cyclicality. All five ride customers' capital-spending and new-product cycles; a pause in auto or industrial investment hits order books quickly and lumpily. Capital-goods children (333517, 333519) have the most operating leverage — Hurco's revenue fell from $227.8 million (FY2023) to $178.6 million (FY2025) while gross margin fell from 25% to 18% and operating margin went from +3% to −6%. Even the bespoke-tooling children show 10%-plus swings in BLS real output year over year. [3][4][7]
• Import competition and trade whiplash. The children put the China cost gap at different levels for different products — 40–70% below U.S. prices for molds, up to roughly 40% for dies — but agree it is held back only by tariffs that could be rolled back, and a partial Section 301 rollback already occurred in late 2025. Machine tools and rolling mills face larger, better-capitalized foreign OEMs at the high end; cutting tools face import-heavy supply chains. Tariffs swing in both directions, protecting finished products while raising input costs. [3][5][6][7]
• Customer and end-market concentration. Heavy dependence on automotive means an auto downturn flows straight through the whole level — and the EV transition cuts per-vehicle content by roughly a quarter to a third on both the tooling and the cutting-tool side, so a mix shift toward EVs is a demand headwind even if unit volumes hold. [3][6]
• Skilled-labor scarcity and succession — the hard ceiling. An aging toolmaking/machinist workforce with too few apprentices constrains how fast any child can grow into a reshoring boom, and threatens the continuity of individual shops. BLS projects tool-and-die-maker employment to fall 11% between 2024 and 2034 (55,200 to 49,300) as CNC equipment absorbs traditional tasks, while still generating about 34,200 annual openings across machinists and tool-and-die makers over the same period — essentially all from replacement need. The deeper risk is loss of tacit process knowledge: springback, heat-treatment distortion, press behavior, wear, and how to correct a tool under deadline. Skilled-labor scarcity also hits the capital-goods children, where Park-Ohio reports a shrinking applicant pool causing higher costs, inefficiencies, and downtime. [3][6][7]
• Input-cost volatility. Tool steel and mold-base material (tooling), tungsten and cobalt (cutting tools), and scrap/alloy/energy (rolling and machine building) all hit cost lines that fixed-price quotes cannot always pass through; surcharge mechanisms help the capital-goods children but do not fully insulate margins. [3][6][7]
• Technology disruption. Additive manufacturing and AI-assisted quoting/machining are both a threat to slow adopters and an efficiency lever for leaders; better tool life can even reduce consumable volumes per part. Digital quoting platforms (Xometry, Proto Labs) are moving closer to the customer in project tooling. [3][5][6]
• Execution and contract risk (specific to the project children). Fixed-price quoting errors, late customer design changes, uncompensated engineering changes, validation failures, acceptance disputes, warranty claims, work-in-process write-downs, and disputes over ownership of tooling or CAD files. A backlog should be discounted for cancellation rights, remaining engineering risk, and the amount already billed. [3][5]
• Safety and product liability. Presses, rolls, stored energy, high voltage, and heavy moving assemblies produce elevated injury rates (333519 total-recordable rate 4.2 per 100 workers in 2022) and continuing design, guarding, and service obligations for OEMs. [4][7]
• Thin public-market liquidity (an investor-specific risk). With at most one micro- or small-cap pure-play per child, none of which discloses NAICS-level revenue sensitivity, the theme is hard to own cleanly and easy to own only via proxies. [3][4][7]

10. How to invest and the outlook

Public-market routes (all partial). There is no way to own this level cleanly on a U.S. exchange. The realistic menu, best-to-thinnest:

• Cutting tools offer the closest thing to a listed leader — Kennametal (NYSE: KMT), a cyclical consumable business with pricing power on performance, paying a $0.20 quarterly dividend (~$0.80 annualized) in fiscal 2025; judge yield and valuation against where you are in the manufacturing cycle, not a static multiple, and note that its Infrastructure segment ($747 million of FY2025 sales) sits outside this level. Diversified majors Sandvik (SAND), Kyocera (KYOCY), Mitsubishi Materials (5711), Sumitomo Electric (5802), and OSG (6136), and Berkshire Hathaway (BRK.B) via ISCAR, add exposure bundled with much larger businesses. Distributors MSC (MSM) and Grainger (GWW) monetize tool consumption without factory risk. [6]
• Rolling mills and machine tools are mostly foreign or micro-cap: Ampco-Pittsburgh (AP) for in-scope mill rolls, Park-Ohio (PKOH) for partial U.S. exposure with heavy aftermarket content, foreign-listed Danieli, Andritz, and Mitsubishi Heavy Industries (Primetals) on the mill side, and Hurco (HURC), DMG Mori, Okuma / Makino / Amada / JTEKT, and adjacent controls leader FANUC on the machine-tool side. Standard equity metrics — dividend yield and enterprise-value-to-EBITDA (earnings before interest, taxes, depreciation, and amortization) — apply, but each gives only a slice, and for the foreign names currency, governance, and non-U.S. demand can dominate the exposure you are buying. An often simpler expression of the mill thesis is owning the steel and aluminum producers whose capex is this industry's revenue. [4][7]
• Molds and dies are indirect only — GPGI (Husky), Hillenbrand (HI), Proto Labs (PRLB), Xometry (XMTR), Core Molding (CMT), Exco (XTC), MISUMI (9962), Mayville (MEC), and Cleveland-Cliffs (CLF) — plus broad industrial/machinery ETFs (exchange-traded funds) to capture the capex cycle without pure-play risk. The list has grown, but not one of these is a job-shop mold or die builder. [3][5]

Private routes (where the real access is). Because the level is overwhelmingly private, direct ownership is the live opportunity — and the aging-owner succession wave makes it a buyer's field. The playbook: acquire and professionalize an individual shop (the search-fund/independent-sponsor model), back a private-equity roll-up (Tooling Tech Group and Tooling Systems Group in dies; SyBridge in molds; Privet- and Centre Lane-style take-privates in accessories and machine tools), or take a supplier/partner position serving reshored production. Scale, engineering depth, metallurgical IP, installed base, and a trained workforce are the assets that command a premium in a fragmented field. The diligence items recur across all five children and are worth treating as one checklist: separate recurring repair/service/aftermarket revenue from volatile new-build awards; test customer and platform concentration; scrutinize backlog conversion, cancellation and acceptance terms, milestone billing, and work-in-process accounting; check quoted-versus-actual hours, machine utilization, scrap and rework, and on-time delivery; assess workforce age and apprenticeship pipeline and dependence on senior technical personnel; and price replacement cost of CNC, EDM, press, and metrology assets alongside environmental and export-control history. [3][4][5][6][7]

Near-term outlook (forward-looking judgment, not reported fact). The level is not moving as one body — the three clocks read differently right now, and one of them is contested:

• Capital-goods children are in a genuine order up-cycle. U.S. machine-tool orders reached $5.74 billion in 2025 (+22.5%), capped by a record December of $814 million, and ran $2.77 billion in January–May 2026 (+~32%) on an aerospace-and-defense surge and reshoring; rolling-mill demand is buoyed by an EAF-led steel build-out, Nucor-scale capex, and Section 232 tariffs pulling investment onshore. These are the level's hot spots — but also its highest-beta, most capex-sensitive exposures, and the Hurco example is the caution: orders are not earnings, and the lag can run years. [4][7]
• The consumable child is growing modestly. Cutting-tool shipments were $2.56 billion in 2025, up 2.5%, with the industry "cautiously optimistic that 2026 will be slightly better than 2025" — the level's most defensive, recurring-revenue slice, still tied to factory utilization and now also to tungsten supply and tariff dynamics. [6]
• The bespoke-tooling children disagree, and the hard data lean soft. Die shops report quoting up 20%-plus and utilization climbing toward the high-70s; mold shops report AMBA's 2026 survey calling shipments, backlog, and quoting broadly stagnant, with automotive stalled for some. Both readings are survey sentiment; BLS real output for dies fell in 2023 to near its 2020 trough while unit labor cost rose. The honest reading is soft-to-flat with improving quotation activity — and, either way, a private-market consolidation story rather than a public one, where value accrues to operators who automate quoting and machining, add recurring maintenance and production work, and serve IP- and speed-sensitive niches (medical, defense, aerospace). [3][5]

The unifying takeaway for a general investor: metalworking machinery is a small, cyclical, skilled-labor-constrained, reshoring-favored, overwhelmingly private corner of U.S. manufacturing that has been shrinking in plant count for two decades even as its strategic importance has risen. Public investors should treat it as a set of leading indicators — three of them, on three different clocks — and as a source of indirect allocations (consumables leader Kennametal, mill-roll and adjacent small caps, foreign OEMs, controls and diversified industrials) rather than a standalone theme. Private investors will find the largest and most direct opportunity, concentrated in the succession-driven roll-up of thousands of independent tool, mold, and machine shops.

---

Sources

1. U.S. Census Bureau. County Business Patterns 2023 — NAICS 33351 and children 333511/333514/333515/333517/333519 (establishments, employment, annual and Q1 payroll). https://www.census.gov/programs-surveys/cbp.html
2. U.S. Census Bureau. 2022 Economic Census, Concentration by Largest Firms — NAICS 33351 and children (receipts, firms, CR4/CR8/CR20/CR50, HHI; rollup CR4 9.4%, HHI 37.2; child CR4s 9.1%/16.6%/17.6%/19.3%/28.6% and HHIs 50.4/99.9/130.2/190.5/314). https://www.census.gov/programs-surveys/economic-census.html
3. Histometrics child primer, Special Die and Tool, Die Set, Jig, and Fixture Manufacturing (NAICS 333514) — drawing on CBP 2023 and 2022 Economic Census (2,074 firms; CR4 16.6%, HHI 99.9); BLS Industry Productivity via FRED (sectoral output $9.85B in 2022 and $9.15B in 2023; real output index 88.9 → 78.9; unit labor cost 132.2; all-persons employment 53,900 in 2022 → 49,000 in 2025); BLS Occupational Outlook Handbook (median wage $63,180 in 2024; tool-and-die employment −11% 2024–2034, 55,200 → 49,300); AMBA Business Forecast Survey 2025 (quoting +20%; utilization ~71%→73%→77%; tool-steel and aluminum cost inflation); Grand View Research metal-stamping data (automotive ~60% of U.S. stamping, >40% of global tooling demand); JEELIX workforce and EV-tooling estimates (~75% over 45, ~40% retirement-eligible, ~2% under 35; ~30% lower tooling spend per EV; EV ~28% of automotive tooling demand by 2030; ~40% offshore price gap); Modern Machine Shop / CADDi job-shop benchmarks (30–60% gross on direct cost; ~13.5% vs ~8% net; >75% utilization); Exco Technologies (TSX: XTC) FY2024 (C$637.8M); MISUMI (TYO: 9962), Mayville Engineering (NYSE: MEC), Cleveland-Cliffs (NYSE: CLF); Tooling Tech Group and Tooling Systems Group roll-ups; OSHA manufacturing-amputation NEP (CPL 03-00-027, NAICS 333514 named); White House Section 232 action (25%→50%, June 2025); IMARC broad U.S. tooling market (~$49–51B).
4. Histometrics child primer, Machine Tool Manufacturing (NAICS 333517) — drawing on CBP 2023 and 2022 Economic Census (1,195 establishments; 862 firms; $9.26B receipts; CR4 28.6%, HHI 314); ITIF "Mapping Industrial Strength: US Machine Tool Production and Consumption" (U.S. production ~$5.9B and consumption ~$10.5B in 2022, 5th globally; 2024 imports ~$1.6B vs exports ~$550M); AMT/USMTO order reports (2025 orders $5.74B, +22.5%, record December $814M; Jan–May 2026 $2.77B, +~32%; aerospace +45.1% in 2025 and +233% in the first two months of 2026); Hurco (Nasdaq: HURC) FY2025 Form 10-K (sales $178.6M; gross margin 18%; operating margin −6%; service revenue $33.9M / 19%; inventory $142.9M; ~45 days backlog); BLS Current Employment Statistics (333517 employment 39,600 in March 2025, 42,300 in April 2024); Haas, Yamazaki Mazak, DN Solutions (Heller acquisition, early 2026), Trumpf, Hardinge/Centre Lane (2024), DMG Mori, Okuma, Georg Fischer; BIS ECCN 2B001 five-axis controls; Section 232 actions including HTS 8457.10.00 and the robots/machinery investigation; OSHA 29 CFR 1910.212.
5. Histometrics child primer, Industrial Mold Manufacturing (NAICS 333511) — drawing on CBP 2023 and 2022 Economic Census (1,252 establishments; 1,095 firms; $6.26B receipts; $1.92B payroll; CR4 9.1%, CR50 42.8%, HHI 50.4); 2002 Economic Census baseline (2,468 establishments; 46,722 employees; $5.528B shipments); AMBA 2026 Business Forecast Report (84 manufacturers; 54% under $5M revenue; stagnant shipments, backlog, and quoting; automotive stalled) and 2025 wage survey; AMBA Section 301 filing (custom injection molds averaging >$100,000, range $15,000–$8M); American Mold Builder (median shop 36→21 shop employees and 6→4 design employees, 2015–2024); GPGI / Husky Technologies Form 10-K (~4,000 customers, ~140 countries, ~13,500 installed systems, ~65% aftermarket revenue); Hillenbrand (NYSE: HI), DME and Mold-Masters, and Milacron (Bain Capital ~51%, March 2025; Hillenbrand ~49%; FY24 ~$526M); Proto Labs (NYSE: PRLB) 2025 Form 10-K ($533M revenue; injection molding $191.5M / 35.9%); Xometry (NASDAQ: XMTR) 2025 Form 10-K; Core Molding (NYSE American: CMT) FY2024 ($302.4M; tooling ~$30–40M); Platinum Equity/Husky; SyBridge Technologies (15 acquired companies); FRED producer price index for 333511 (145.076, April 2026); Moldminds China cost gap (40–70%); USTR Section 301 actions and the late-2025 partial rollback; BLS median wage $63,180; OSHA and EPA machining rules.
6. Histometrics child primer, Cutting Tool and Machine Tool Accessory Manufacturing (NAICS 333515) — drawing on CBP 2023 and 2022 Economic Census (1,060 establishments; 1,050 firms; $5.55B receipts; $1.63B payroll; CR4 17.6%, CR50 60.4%, HHI 130.2); AMT–USCTI Cutting Tool Market Report (2025 U.S. shipments $2.56B, up 2.5%; "cautiously optimistic" on 2026); Kennametal (NYSE: KMT) FY2025 Form 10-K and results release (sales ~$1.97B; Metal Cutting $1.22B at 7.1% operating margin vs 10.4%; company operating margin 7.3% vs 8.3%; Infrastructure $747M; 60% of sales outside the U.S.; $0.20 quarterly dividend); Sandvik AB 2024 annual report (Machining SEK 48.6B, ~40% of group); Berkshire Hathaway 2025 Form 10-K (IMC/ISCAR ~$4.1B revenue, +3.9%); Mitsubishi Materials (Metalworking Solutions ¥148.8B; carbide tool use per EV 25–30% lower); Market Research Future U.S. consumption estimate (~$15.1B, 2024); Harvey Performance / Berkshire Partners and the 2024 ATA acquisition; Privet Fund / Hardinge (~$245M take-private, 2018); MSC Industrial (MSM) and Grainger (GWW) as distributor proxies; USGS tungsten net import reliance (50%); OSHA cobalt PEL 0.1 mg/m³; BLS occupational projections (~34,200 annual openings, employment −2%, 2024–2034); NIST additive-manufacturing post-processing.
7. Histometrics child primer, Rolling Mill and Other Metalworking Machinery Manufacturing (NAICS 333519) — drawing on CBP 2023 and 2022 Economic Census (289 establishments; 280 firms; 10,371 employees; $823M payroll; $2.97B receipts; CR4 19.3%, CR8 31%, CR20 51.3%, CR50 75.6%, HHI 190.5); Ampco-Pittsburgh (NYSE: AP) FY2024 results and annual report (net sales $418.3M; Forged & Cast ~67% of sales; mill-roll sales $273.0M; segment operating income $10.5M / 3.7%; backlog $250.5M; segment operating income +38%); Park-Ohio (NYSE: PKOH) 2025 Form 10-K (aftermarket ~49% of induction/melting revenue; tariff, input-cost, and skilled-labor risks); SMS group 2024 annual report (sales €4.033B; order intake €3.620B; 13,500+ employees; metallurgical plant-engineering order intake €2.311B, −37.2%); Danieli (Milan: DAN) FY2024/25 (Plantmaking €3.116B; 12.3% EBITDA margin, 10.3% operating margin); Andritz (Vienna: ANDR) FY2025 (group ~€7.9B; Metals €1.694B; 6.7% EBITDA margin; capital equipment 73% / service 27%); Primetals via Mitsubishi Heavy Industries (Tokyo: 7011); Fives Group silicon-steel processing lines; White House Section 232 fact sheet (25%→50% effective June 4, 2025; >$10B announced new U.S. mill investment); Fastmarkets / S&P Global Commodity Insights (EAF ~69% of U.S. output in 2024; Nucor ~$3–4B capex 2024–2026); World Steel Association Short Range Outlook (U.S. demand +1.7% in 2026, +2.0% in 2027); Persistence Market Research / market.us (metalworking-machinery orders ~$2.52B through H1 2025, +13.7%; automotive ~39% of 2025 demand; aerospace/defense fastest-growing); BLS injury and illness rates (333519 TRC 4.2, DART 2.1, 2022); OSHA machine-guarding standards; EPA Metal Products and Machinery effluent guidelines; FENN, International Rolling Mills, Butech Bliss as U.S. private specialists.
8. U.S. Small Business Administration. Table of Small Business Size Standards Matched to NAICS Codes — NAICS 33351 group, 500-employee standard. 2023. https://www.sba.gov/document/support-table-size-standards