06 Yield Management Metrology Inspection
The structural escalation of semiconductor process complexity—driven by 2nm GAA transistors, HBM4 stacking, and heterogeneous integration—is forcing a disproportionate increase in process-control intensity (PCI) as a percentage of WFE, from ~7.4% today toward an estimated 9% by 2030. This creates above-market growth for yield management, metrology, inspection, and cleaning equipment providers. The ecosystem is bifurcated between hardware-based defect detection/remediation (KLA, Onto, Camtek, SCREEN) and software/analytics orchestration (PDF Solutions); Nova Ltd. sits at the intersection with hardware-integrated metrology and growing software capabilities. The core investment opportunity lies in companies with wide moats and direct leverage to the most technically demanding node transitions, but elevated valuations across the group demand selectivity.
| Outlook | bull |
|---|
| Sentyment sektora | Bullish |
|---|
| Faza cyklu kapitałowego | mid_cycle |
|---|
| Ekspozycja na AI | Direct and high. Every AI chip (GPU, ASIC, HBM) requires multiple inspection, metrology, and cleaning steps. Companies with HBM- or GAA-specific tool sets (NVMI, ONTO, CAMT, SCREEN) have highest direct leverage; PDF Solutions provides the software layer for yield optimization across all AI nodes. |
|---|
Uzasadnienie outlooku
The metrology/inspection market is projected to grow at 7.8% CAGR to $14.7B by 2033, outpacing historical WFE growth. AI-driven node complexity (GAA, HBM4, advanced packaging) multiplies inspection and cleaning steps per wafer, creating a structural tailwind independent of broader semiconductor capex cycles. The shift from standalone hardware to integrated, AI-driven process control adds recurring software and service revenue layers.
Ranking spółek
- NVMI — Nova Ltd. ()
- ONTO — Onto Innovation Inc. ()
- 7735.T — SCREEN Holdings ()
- PDFS — PDF Solutions ()
- KLAC — KLA Corporation ()
- CAMT — Camtek ()
Kluczowe ryzyka
- {'risk': 'Geopolitical escalation (US-China export controls): Further restrictions on metrology and inspection tools could cut 33-44% of revenues for exposed companies (NVMI, KLAC, CAMT). The Chinese "countering extraterritorial jurisdiction" law creates a legal paradox for dual-use equipment.', 'timeframe': '6-12 months'}
- {'risk': 'Valuation multiple compression: Current trailing P/E ratios above 150x for ONTO and CAMT and 70x for KLAC leave little room for error. Any growth deceleration could trigger 30-50% drawdowns.', 'timeframe': '3-6 months'}
- {'risk': 'Customer concentration and fab pushback: Top 3 customers account for 49-80% of revenue for most names. A single fab pushback by TSMC, Samsung, or Intel can materially impact quarterly deliveries.', 'timeframe': '12-24 months'}
- {'risk': 'Margin pressure from mix shift and competition: Gross margins may decline as product mix shifts toward lower-margin inspection tools or as competition intensifies in HBM packaging (Camtek vs. Onto).', 'timeframe': '6-12 months'}
Katalizatory
- {'catalyst': "Q2 2026 earnings beats (July-August 2026) for NVMI and ONTO. Revenue guidance of $245-255M for NVMI and upward revisions could reinforce growth narrative. KLAC's Q4 FY2026 (late July) revenue guide above $3.575B would support the stock.", 'timeframe': 'July-August 2026'}
- {'catalyst': 'HBM4 volume ramp at major memory makers (SK Hynix, Samsung, Micron), drives demand for inspection, metrology, and cleaning equipment across the ecosystem.', 'timeframe': 'H2 2026'}
- {'catalyst': 'SEMICON West (July 2026) and industry conferences. Announcements of new tool wins (especially GAA "tool of record" designations) or PCI expansion studies could reset expectations.', 'timeframe': 'July 2026'}
- {'catalyst': 'Successful refinancing of NVMI convertible notes without equity dilution. Would remove a balance sheet overhang; failure could pressure the stock.', 'timeframe': 'medium-term'}
- {'catalyst': 'US CHIPS Act disbursements and new fab announcements. Could provide incremental demand for domestic fabs, benefiting all equipment suppliers.', 'timeframe': 'medium-term'}