Semiconductor Machinery Manufacturing (U.S.) — An Investor's Primer

NAICS 2022 code 333242 · Semiconductor Machinery Manufacturing

1. Overview

Semiconductor machinery manufacturing is the business of building the machines that build chips. It sits one step upstream of the semiconductor industry itself: chipmakers such as TSMC, Samsung, Intel, and Micron buy the tools, and companies in this industry make them. A modern chip factory ("fab") is essentially a room full of these machines — depositing, etching, patterning, and measuring layers on silicon wafers a few atoms at a time. The collective term for these tools is WFE (wafer fab equipment), and the broader category is SME (semiconductor manufacturing equipment).

Why an investor cares: this is the "picks and shovels" layer of the digital economy. Whether the winning chip is a graphics processor for artificial intelligence (AI), a memory chip, or an automotive controller, almost every one is made on tools from a short list of suppliers. That gives the leading firms durable, near-monopoly positions in their niches, gross margins in the 45–60% range, and a growing stream of recurring service revenue [1][16]. The catch is cyclicality: equipment orders swing with chipmakers' capital-spending cycles, so revenues and share prices can move sharply.

Ways in. For public-market investors, this is one of the most directly investable corners of the chip world: the three U.S. leaders — Applied Materials, Lam Research, and KLA — are large, liquid, dividend-paying stocks, alongside a set of smaller specialists [6][7][8]. For private investors, the large incumbents are effectively out of reach except through public shares, but venture capital and private equity are active in the layer beneath them — subsystem, component, materials, and advanced-packaging-tool startups that feed the majors.

2. What it is and how it's structured

Scope. NAICS (North American Industry Classification System) code 333242 covers U.S. establishments that make wafer-processing equipment, semiconductor assembly and packaging equipment, and other chip-making machinery [2]. In plain terms, that includes:

• Deposition tools (laying down ultra-thin films)
• Etch tools (carving patterns into those films)
• Lithography tools (printing the circuit pattern using light)
• Ion implantation, chemical-mechanical planarization (CMP), and thermal/anneal tools
• Metrology and inspection tools (measuring and finding defects — also called process control)
• Assembly and packaging machinery (bonders that connect finished chips into packages)
• Surface-mount machinery for populating printed circuit boards [2]

What it excludes (name-checking the adjacent codes so the boundaries are clear):

• The chips themselves are NAICS 334413, Semiconductor and Related Device Manufacturing — these are the customers, not this industry.
• Printed-circuit-board-making machinery is NAICS 333248, All Other Industrial Machinery Manufacturing [2].
• Semiconductor test equipment — automated test systems (ATE) from firms like Teradyne and Advantest — is NAICS 334515, Instrument Manufacturing for Measuring and Testing Electricity [2]. So chip testers sit just outside this code even though they live in the same fabs. This boundary matters: the commonly quoted "semiconductor equipment market" normally includes test equipment that is outside 333242.

Operating model. Manufacturing an equipment system is less vertically integrated than the word "machinery" suggests. Applied Materials describes its own factories as primarily assembling, integrating, and testing proprietary and purchased parts, components, and subassemblies; it also uses contract manufacturers and a geographically distributed supply chain [6]. Critical inputs include precision optics, lasers and electron-beam sources, vacuum pumps, RF power systems, valves, robotics, fluid-delivery systems, computers, sensors, rare-earth-containing components, and extensive control software. Some key components have only one qualified supplier or a small qualified group [6].

The commercial process begins years before a production order. Vendors collaborate with chipmakers while the customer is developing a new process, place development tools in laboratories, and attempt to become the development and eventually production "tool of record." Winning depends on yield, throughput, uptime, process repeatability, and cost per good wafer — not simply the machine's purchase price. Qualification is lengthy because changing a process tool can alter a customer's recipe, yield, and product reliability [7].

Ownership mix. This is overwhelmingly a public-company industry. A handful of large, publicly traded corporations account for the bulk of activity, trailed by a tail of smaller private and specialty machine shops and subsystem suppliers. The federal count is just 144 firms across 170 U.S. establishments [1] — a small, concentrated population by manufacturing standards.

3. How big it is

Our ground-truth federal figures for U.S.-located establishments in NAICS 333242:

The concentration numbers tell the core story: the top four firms book six of every ten dollars, and the top twenty book more than nine [1].

The undercount caveat — read this carefully. These federal figures measure production shipped from U.S.-located establishments. They are not a measure of how big the U.S. firms in this business are. The domestic-shipment figure of about $13.75 billion (2022) is dwarfed by the global revenue of the three American champions alone — Applied Materials booked $28.4 billion in fiscal 2025, and Lam Research and KLA together well over $30 billion more [6][7][8]. The gap exists because these companies manufacture and sell a large share of their systems through overseas establishments (Singapore, Malaysia, and elsewhere), and because the global equipment market — about $135 billion in worldwide billings in 2025 [3] — is served partly by imported tools (ASML, Tokyo Electron) that never touch this NAICS code. So the federal number understates the true economic weight of U.S. firms in this sector; treat it as "domestic factory output," not "industry size." Similarly, the domestic HHI of 1,153 [1] looks only "moderately concentrated" because it counts many small U.S. specialty shops; at the technological leading edge, individual process segments are effectively single-supplier monopolies (see §8).

4. The investable universe

Unusually for an upstream industrial niche, most of the meaningful players are publicly listed — this is a stock-picker's industry, not a private one.

U.S.-domiciled public companies (the core):

Foreign-domiciled but part of the same universe (they sit outside NAICS 333242 but dominate the tools U.S. fabs buy, and several trade in the U.S.):

• ASML (Netherlands) — the sole maker of EUV (extreme ultraviolet) lithography systems, with essentially the entire EUV market and the only High-NA (high numerical aperture) tools [17]; U.S.-listed as an ADR. ASML reported €32.7 billion of 2025 sales, a 52.8% gross margin, and €9.6 billion of net income [23].
• Tokyo Electron (Japan) — a top-five deposition/etch/coater supplier.
• ASM International (Netherlands) — deposition (notably atomic-layer deposition).
• Nikon and Canon (Japan) — older-generation DUV (deep-ultraviolet) lithography.
• Nova (NVMI, Israel) — metrology, ~$0.67B (2024) [16]; Camtek (CAMT, Israel) — inspection and advanced-packaging tools, ~$0.43B (2024) [13].

Private and "other" owners. There are few U.S. private toolmakers of scale — the pure-play equipment layer consolidated into public companies long ago. Private and PE-owned activity concentrates one level down: subsystem, component, precision-parts, and materials suppliers that sell into the majors, plus a venture-funded fringe building next-generation metrology, advanced-packaging, and specialty-process tools. Attractive private niches include vacuum and abatement systems, precision motion, RF power, valves and fluid delivery, subsystems, advanced-packaging tools, and legacy-tool service and refurbishment. If there are few or no direct private plays here, that is the honest answer: this is an industry you mostly access through public equity.

5. How the money works

This is a capital-equipment manufacturing business, and its economics follow that template rather than any recurring-utility or asset-based model.

The two revenue engines. (1) Systems — large, episodic tool sales. A single leading-edge machine is a multimillion-dollar purchase; ASML's advanced EUV systems reportedly run well over $100 million each, with its newest High-NA systems around $380 million apiece [17]. (2) Service and installed base — spares, consumables, upgrades, and service contracts on the tens of thousands of tools already running in fabs worldwide. This aftermarket is annuity-like: it grows every time a system is sold and keeps paying through downturns. Lam reported fiscal 2025 system revenue of $11.49 billion and customer-support and other revenue of $6.94 billion [7]. For ASML, recurring service is roughly a quarter of revenue [16]. The installed base is both an annuity-like revenue source and a source of operating data that can improve future tools.

Margins and reinvestment. Gross margins sit around 45–60% — Applied Materials at 48.7%, Lam Research around 48.7%, ASML at 52.8%, and KLA in the low 60s [6][7][8][23] — reflecting extreme precision, limited competition, and pricing power at the leading edge. Against that, research and development (R&D) intensity is very high: each transition to a smaller transistor node forces a new generation of tools, so sustained double-digit-percent-of-sales R&D is the price of staying in the game. Applied spent $3.57 billion on R&D in fiscal 2025 (about 12.6% of revenue), and Lam spent $2.10 billion (11.4% of revenue) [6][7].

The metrics investors actually watch:

• WFE spending — the total addressable pool; chipmakers' collective capital expenditure ("capex") is the demand signal.
• Bookings, backlog, and book-to-bill — order flow relative to shipments; a book-to-bill above 1.0 signals a rising cycle.
• Fab utilization at customers — full fabs order more tools.
• Gross margin and service mix — the share coming from installed-base revenue smooths the cycle.
• China revenue mix — a swing factor because of export controls (see §7). Lam derived 34% of fiscal 2025 revenue from customers' facilities in China [7].
• Memory vs. logic split — memory capex (DRAM and NAND) is the most volatile driver.

Customer concentration. A few chipmakers drive the majority of orders. Applied reported that two customers represented approximately 19% and 15% of fiscal 2025 revenue [6]. KLA identified TSMC and Samsung as customers exceeding its disclosure threshold and stated that 89% of fiscal 2025 revenue was international [8]. A delayed fab, changed process architecture, or capex cut by one large foundry or memory manufacturer can therefore move a vendor's quarterly results.

Unit economics in one line: a single new leading-edge fab represents billions of dollars of equipment demand, and every technology node adds more process steps — more deposition, etch, and inspection passes per wafer — so tool intensity per wafer rises over time even when wafer volumes are flat.

6. What drives demand

Demand is derived from chip demand, filtered through chipmakers' willingness to spend:

• End markets for chips — AI data centers (graphics processors and HBM, high-bandwidth memory), smartphones, PCs, automotive electronics, and industrial/IoT devices.
• Customer capex cycles — a concentrated set of buyers (TSMC, Samsung, Intel, Micron, SK Hynix). Industry-wide chip capital spending is projected near $200 billion in 2026, with TSMC alone more than a quarter of it [22].
• Technology transitions — each new node (5nm → 3nm → 2nm), the shift to gate-all-around transistors, backside power delivery, new interconnect materials, and the rise of advanced packaging all require new and additional tools. High-NA EUV adoption is a fresh demand layer [17].
• The memory cycle — DRAM and NAND pricing drives memory makers' capex, historically the most boom-and-bust segment [19].
• AI as a potential cycle-extender — the current up-cycle has been powered largely by AI-driven memory (especially HBM) and advanced-logic investment. SEMI reported that in 2025 wafer-processing equipment sales rose 12%, test-equipment billings rose 55%, and assembly and packaging equipment sales rose 21% [3]. Whether AI structurally lengthens the cycle or merely front-loads it is the central forward-looking debate in the sector [19].
• Regionalization — governments are subsidizing fabs for supply security, creating duplicate or geographically dispersed capacity that may be less economically efficient but still requires equipment. As of January 2025, the U.S. Commerce Department said semiconductor and electronics companies had announced nearly $450 billion of U.S. private investment, catalyzed in part by public incentives [24]. Announcements are not completed spending, however, and construction schedules, award conditions, and customer economics determine when equipment orders actually occur.

7. Regulation

Two policy forces dominate this industry, and both are unusually consequential for revenue.

Export controls (the headwind). Because chip-making tools are dual-use, the U.S. Bureau of Industry and Security (BIS) has progressively restricted sales of advanced equipment to China through rules in October 2022, October 2023, April 2024, and December 2024 [18]. In August 2025, BIS removed license-free treatment previously available to certain foreign-owned fabs in China, requiring licenses for affected exports [20]. The most advanced lithography — EUV — has never been sold to China (the Netherlands blocked it in 2019), and controls now extend to advanced DUV, etch, and deposition tools [18]. Allied governments have followed: the Netherlands added certain metrology and inspection systems to its control list in January 2025, and the European Union incorporated Dutch national controls in November 2025 [18]. This matters because China has been one of the largest equipment markets — China, Taiwan, and Korea together made up about 79% of global equipment spending in 2025 [3] — so lost China revenue is a real and recurring risk to the U.S. and allied suppliers. Controls can also accelerate Chinese substitution and divert revenue to vendors outside U.S. jurisdiction.

Industrial policy (the tailwind). The U.S. CHIPS and Science Act (2022) appropriated $52.7 billion to rebuild domestic chip manufacturing, with about $33.7 billion in direct-funding awards finalized by early 2025 [21]. Marquee projects — TSMC's roughly $165 billion Arizona complex, Intel's Ohio and Arizona fabs, Samsung in Texas, Micron memory fabs — plus a 25% federal investment tax credit for fab equipment translate directly into domestic tool demand [21]. The offset: several of these builds have slipped (Intel's Ohio fab is now targeted around 2030) [21], so the timing of equipment orders is lumpier than the headline dollars suggest.

8. Competitive dynamics and consolidation

This is a segmented oligopoly: within each process step, one or two firms dominate globally.

• Lithography — ASML holds essentially the entire EUV market and the only High-NA tools; Nikon and Canon retain older DUV [17].
• Etch — Lam Research leads with roughly 45% share [17].
• Deposition — Applied Materials and Tokyo Electron are the leaders, with Lam and ASM also large [17].
• Process control (inspection and metrology) — KLA holds around 52% [17].
• Broadest portfolio — Applied Materials spans deposition, etch, ion implant, CMP, and inspection.

The top five suppliers — Applied Materials, ASML, Lam Research, Tokyo Electron, and KLA — together command roughly 65% of the market [17]. These positions are defended by enormous R&D, deep co-development relationships with a handful of customers, and switching costs (a qualified tool is designed into a customer's process and hard to replace).

Consolidation happens mostly at the edges. The defining recent U.S. deal is Axcelis Technologies' announced ~$4.4 billion acquisition of Veeco Instruments (October 2025), which would create the fourth-largest U.S. WFE supplier, with combined pro-forma 2024 revenue near $1.7 billion — still a fraction of the big three [11]. Earlier roll-ups produced today's players (Onto Innovation from the Nanometrics–Rudolph merger; MKS's acquisitions in subsystems and photonics). Expect scale-building among the mid-caps rather than mergers of the giants, which would draw antitrust scrutiny.

9. Risks

• Cyclicality. Equipment demand swings with customer capex on 3–4-year cycles; the 2022–2023 downturn saw memory revenues fall sharply and memory makers slash tool orders. Lam's revenue fell from $17.43 billion in fiscal 2023 to $14.91 billion in fiscal 2024, principally because of lower non-volatile-memory spending, before recovering to $18.44 billion in fiscal 2025 [7][19]. A "digestion" phase after any boom hits this group first and hardest.
• Customer concentration. A few chipmakers drive the majority of orders; a single customer's delay (as with Intel's fab slippage) moves the numbers [21].
• China / export-control exposure. Tighter rules can strip out a chunk of the addressable market, and China is building domestic toolmakers that could eventually compete at trailing edges [18][20].
• Technology risk. Missing a node transition (a failed High-NA ramp, a bet on the wrong process) can shift share for a decade.
• Geopolitical concentration of demand. With ~79% of spending in China, Taiwan, and Korea, any disruption in East Asia is an outsized shock [3].
• Supply chain. Critical inputs include sole-source optics, precision stages, pumps, power systems, electronics, and proprietary supplier IP that may take substantial time to redesign and requalify. Applied and KLA both identify Chinese restrictions on certain rare-earth minerals, supplier concentration, inflation, and long qualification periods as risks. Equipment also contains chemicals or components potentially affected by restrictions on PFAS, hydrofluorocarbons, and other regulated substances [6][8].
• Labor. An SIA/Oxford Economics study projected that 67,000, or 58%, of expected new U.S. semiconductor-industry jobs could go unfilled by 2030 at then-current degree-completion rates [25]. Equipment makers compete for the same engineers and technicians.
• Valuation and timing. These are volatile, cyclical equities; buying at a cycle peak has historically meant large drawdowns.

10. How to invest and the outlook

Public-market routes (where tickers and valuation belong). The purest large-cap plays are the U.S. big three — Applied Materials (AMAT), Lam Research (LRCX), and KLA (KLAC) — all liquid, profitable, dividend-paying, and highly cyclical [6][7][8]. Smaller specialists give more focused (and more volatile) exposure: Axcelis (ACLS), Onto Innovation (ONTO), MKS Instruments (MKSI), and Kulicke & Soffa (KLIC), plus foreign-listed leaders such as ASML (via ADR) and Tokyo Electron. Note there is no widely held U.S.-listed pure equipment fund; broad semiconductor ETFs — the iShares Semiconductor ETF (SOXX) and VanEck Semiconductor ETF (SMH) — hold the equipment names heavily but blend them with chipmakers. Because the group is cyclical, entry point and valuation multiple matter as much as company quality.

Private-market routes. Direct private ownership of a major toolmaker is not on the table. Private capital instead flows into the supply chain beneath them: subsystem, component, precision-parts, and materials suppliers, and venture-stage startups in metrology, advanced packaging, and specialty processes — often with the majors themselves as strategic acquirers. Diligence should focus on whether the company owns qualification-critical IP, is designed into a leading vendor's platform, has sole- or dual-source status, and earns recurring spares or service revenue. The counter-risk is extreme customer concentration and the possibility that the system OEM internalizes the component or forces a second source. Supplying into that chain, or backing it through PE/VC, is the realistic private way in.

Common analytical errors. The most common is mixing three incompatible datasets: U.S. establishment output under NAICS 333242, global revenue of U.S.-headquartered vendors, and SEMI's worldwide equipment billings including test. They measure different things. A second error is treating the industry as a homogeneous machinery market; in practice, it is a collection of narrow process oligopolies with materially different margins and competitive positions. A third is assuming that smaller nodes automatically mean fewer machines: increasingly three-dimensional devices, new materials, process-control intensity, and advanced packaging can increase equipment content even when wafer volumes grow slowly.

Near-term drivers (forward-looking). SEMI projects total equipment sales rising to a record ~$139 billion in 2026 and ~$156 billion in 2027, up from ~$135 billion in 2025 [3][4][5]. The bullish case rests on AI-driven demand for advanced logic and HBM memory, continuing node transitions, growth in advanced packaging, the start of High-NA EUV adoption, and CHIPS-funded U.S. fab construction [17][21][19]. The bearish case is the industry's oldest one: cyclicality, customer concentration, and the risk that AI simply pulls demand forward. Overlaying both is China — a swing factor that export policy can enlarge or shrink at the stroke of a rule [18][20]. Reported facts point to a record 2026–2027; the judgment call is whether AI has genuinely tamed the cycle or merely postponed the next digestion.

---

Sources

1. U.S. Census Bureau, County Business Patterns (2023) and Economic Census (2022); U.S. Small Business Administration, Table of Small Business Size Standards (2023) — establishments, firms, employment, payroll, receipts, concentration ratios, HHI, and size standard for NAICS 333242 (Histometrics ground-truth compilation). https://www.census.gov/programs-surveys/cbp.html
2. U.S. Census Bureau / NAICS Association, 2022 NAICS Definition — 333242 Semiconductor Machinery Manufacturing (2022). https://naics.askkodiak.com/naics/2022/333242
3. SEMI, Global Semiconductor Equipment Billings Reached $135 Billion in 2025, Up 15% Year-on-Year (2026). https://www.semi.org/en/SEMI-Reports-Global-Semiconductor-Equipment-Billings-Reached-135-Billion-in-2025
4. SEMI, Global Total Semiconductor Equipment Sales Forecast to Reach a Record of $139 Billion in 2026 (2025). https://www.semi.org/en/semi-press-release/global-total-semiconductor-equipment-sales-forecast-to-reach-a-record-of-dollar-139-billion-in-2026-semi-reports
5. SEMI, Global Semiconductor Equipment Sales Projected to Reach a Record of $156 Billion in 2027 (2025). https://www.semi.org/en/semi-press-release/global-semiconductor-equipment-sales-projected-to-reach-a-record-of-156-billion-dollars-in-2027-semi-reports
6. Applied Materials, 2025 Form 10-K — FY2025 revenue $28.37 billion, gross margin 48.7%, R&D $3.57 billion, Semiconductor Systems segment 35.5% operating margin, customer concentration, supply chain disclosure. https://www.sec.gov/Archives/edgar/data/6951/000162828025056742/amat-20251026.htm
7. Lam Research, 2025 Form 10-K — FY2025 revenue $18.44 billion, gross margin 48.7%, R&D $2.10 billion, system revenue $11.49 billion, customer support $6.94 billion, China 34% of revenue. https://www.sec.gov/Archives/edgar/data/707549/000070754925000075/lrcx-20250629.htm
8. KLA Corporation, 2025 Annual Report — FY2025 revenue $12.16 billion, gross margin 60.9%, 89% international revenue. https://www.sec.gov/Archives/edgar/data/319201/000119312525213412/d912971dars.pdf
9. MKS Instruments, Reports Fourth Quarter and Full-Year 2024 Financial Results (2025) — 2024 revenue $3.586 billion. https://www.globenewswire.com/news-release/2025/02/12/3025457/16396/en/MKS-Instruments-Reports-Fourth-Quarter-and-Full-Year-2024-Financial-Results.html
10. Onto Innovation, Fourth Quarter and Full Year 2024 Results (Form 8-K) (2025) — 2024 revenue $987 million. https://www.sec.gov/Archives/edgar/data/704532/000095017025015383/onto-ex99_1.htm
11. Axcelis Technologies, Axcelis Technologies and Veeco Instruments to Combine (2025) — 2024 revenue $1.02 billion; ~$4.4 billion deal; combined pro-forma 2024 revenue ~$1.7 billion. https://investor.axcelis.com/news-releases/news-release-details/axcelis-technologies-and-veeco-instruments-combine-creating
12. Kulicke & Soffa, 2024 Annual Report (Form 10-K) (2024) — FY2024 revenue $706.2 million. https://www.sec.gov/Archives/edgar/data/56978/000005697825000018/kulickesoffa_2024xar.pdf
13. Camtek, Record Results for the Fourth Quarter & Full Year of 2024 (2025) — 2024 revenue $429 million. https://www.prnewswire.com/news-releases/camtek-announces-record-results-for-the-fourth-quarter--full-year-of-2024-302374668.html
14. (Reserved.)
15. (Reserved.)
16. Nova, Record Fourth Quarter and Full Year 2024 Results (2025), 2024 revenue $672.4 million; gross-margin and installed-base/service-revenue profile from Umbrex company profiles and ASML disclosures (2025). https://www.prnewswire.com/news-releases/nova-reports-record-fourth-quarter-and-full-year-2024-results-302375856.html; https://umbrex.com/resources/company-profiles/asm-international-nv/
17. Intel Market Research, Semiconductor Front-End Equipment Market Outlook 2025–2032, and MarketsandMarkets, EUV Lithography Market (2025) — segment shares (lithography ~29%, ASML EUV dominance, Lam etch ~45%, KLA process control ~52%, top-five ~65%) and High-NA system pricing. https://www.intelmarketresearch.com/semiconductor-front-end-equipment-market-3922
18. Bureau of Industry and Security (U.S. Dept. of Commerce), Commerce Strengthens Export Controls (2024), and CSIS, The True Impact of Allied Export Controls on the U.S. and Chinese Semiconductor Manufacturing Equipment Industries (2024) — BIS 2022–2024 rules, EUV/DUV restrictions, Dutch and EU controls. https://www.bis.gov/press-release/commerce-strengthens-export-controls-restrict-chinas-capability-produce-advanced-semiconductors-military; https://www.csis.org/analysis/true-impact-allied-export-controls-us-and-chinese-semiconductor-manufacturing-equipment
19. (Reserved — Dutch/EU control detail covered under [18].)
20. Bureau of Industry and Security (U.S. Dept. of Commerce), Department of Commerce Closes Export Controls Loophole for Foreign-Owned Semiconductor Fabs in China (August 2025). https://www.bis.gov/press-release/department-commerce-closes-export-controls-loophole-foreign-owned-semiconductor-fabs-china
21. Semiconductor Industry Association, Chip Supply Chain Investments (2025), and TSMC, TSMC Arizona and U.S. Dept. of Commerce Announce up to $6.6 Billion in Proposed CHIPS Act Funding (2024) — CHIPS Act $52.7B / ~$33.7B awards, TSMC Arizona ~$165B, Intel Ohio timeline. https://www.semiconductors.org/chip-supply-chain-investments/; https://pr.tsmc.com/english/news/3122
22. FinancialContent / Token Ring, The Silicon Renaissance: US CHIPS Act Enters Production Era (2026) — industry chip capex ~$200 billion in 2026, TSMC >25% share. https://markets.financialcontent.com/wral/article/tokenring-2026-1-1-the-silicon-renaissance-us-chips-act-enters-production-era-as-intel-tsmc-and-samsung-hit-critical-milestones
23. ASML, Q4 2025 Financial Results (2026) — 2025 sales €32.7 billion, gross margin 52.8%, net income €9.6 billion. https://www.asml.com/en/news/press-releases/2026/q4-2025-financial-results
24. U.S. Department of Commerce, CHIPS Incentives Awards — Corning (January 2025) — semiconductor and electronics companies announced nearly $450 billion of U.S. private investment. https://www.commerce.gov/news/press-releases/2025/01/us-department-commerce-announces-chips-incentives-awards-corning
25. Semiconductor Industry Association / Oxford Economics, Chipping Away: Assessing and Addressing the Labor Market Gap Facing the U.S. Semiconductor Industry (2023) — 67,000 jobs (58% of expected new positions) could go unfilled by 2030. https://www.semiconductors.org/chipping-away-assessing-and-addressing-the-labor-market-gap-facing-the-u-s-semiconductor-industry/
26. Franklin Templeton (ClearBridge Investments), Can AI capex extend the semiconductor cycle? (2026), and NomadSemi, State of the Semiconductor Cycle (2025) — cyclicality, 2022–2023 memory downturn, AI-driven recovery. https://www.franklintempleton.com/articles/2026/clearbridge-investments/can-ai-capex-extend-the-semiconductor-cycle