2026-05-15 — views $AMAT · Applied Materials · Deposition · Etch · CMP · Ion Implant · Thermal · Metrology · WFE — broad portfolio
Applied Materials (AMAT) — the broadest WFE portfolio, leveraged to every chip ever made
AMAT sells almost every category of wafer fab equipment except lithography — deposition, etch, CMP, ion implant, thermal processing, and metrology. The most diversified bet on chip manufacturing growth.
Applied Materials (NASDAQ: AMAT) is the largest wafer fab equipment (WFE) company by revenue, and the broadest by product portfolio. Where Lam owns etch leadership and KLA owns process control, AMAT owns most of the rest — deposition, CMP, ion implant, thermal processing, and metrology. The only major WFE category AMAT doesn’t compete in is lithography (ASML’s near-monopoly). For builders, AMAT is the cleanest leveraged read on “more chips get manufactured” — every chip, every node, every fab.
The six segments
| Segment | What it does | AMAT product family |
|---|---|---|
| Deposition | Lays down thin films of material on the wafer | Endura (PVD), Producer (CVD), Olympia (ALD) |
| Etch | Carves away material to form circuit features | Centura, Sym3 Y, Producer Etch |
| CMP | Chemical-mechanical planarization — flattens the wafer between layers | Reflexion |
| Ion implant | Injects dopant atoms into the silicon | VIISta, Quantum |
| Thermal / RTP | Rapid thermal processing — anneals the wafer | Vantage, Centura |
| Metrology | Measures what was deposited / etched / implanted | VeritySEM (review SEM), various inline metrology |
The breadth is the moat. A new fab tooling order from a hyperscaler-adjacent foundry (TSMC, Samsung, Intel, SMIC) will spend across multiple categories — AMAT participates in nearly every line item. Lam captures the etch portion. KLA captures the inspection portion. ASML captures lithography. AMAT captures everything in the gaps.
Sculpta — the post-EUV pattern-shaping bet
Announced in 2023, Sculpta is AMAT’s attempt to extend chip scaling past EUV lithography’s physical resolution limit. Instead of finer lithography, Sculpta uses directional ion beams to reshape existing patterns after they’ve been printed — extending narrow features, trimming wide ones, and reducing the number of multi-patterning steps required.
For 2nm and 1.4nm nodes, the alternative to Sculpta is High-NA EUV (ASML’s even more expensive next-generation lithography tool, ~$380M each). Foundries that adopt Sculpta-style pattern shaping reduce their dependence on High-NA EUV, which means Sculpta is both an AMAT growth story and a partial bear case for ASML’s High-NA ramp if it works at scale. The next 4-6 quarters will reveal which foundries adopt it for production-volume nodes.
Service business — the recurring revenue engine
Applied Global Services (AGS) is the recurring-revenue side of the business: service contracts on installed tools, parts, software, and yield-improvement consulting. AGS grows roughly in line with the installed base of AMAT equipment, which only gets larger over time. In any given year, AGS is a meaningful share of total revenue and a much higher gross margin than systems sales. For a business that otherwise lives and dies by capex cycles, AGS provides the stabilizing baseline.
Why this matters for builders
For anyone thinking about durability of the AI buildout:
- Compute layer competes hard. NVDA vs AMD vs Cerebras vs custom silicon. Margins compress over time.
- Memory layer competes hard. Samsung vs SK Hynix vs Micron. HBM is a duopoly+1, with capacity expansions driving cyclical pricing.
- WFE layer is structurally less competitive. AMAT + Lam + KLA + ASML + Tokyo Electron own ~80% of WFE between them. Switching costs for fabs are enormous — recipes are tuned to specific tool chambers; replacing a vendor means re-qualifying years of process know-how.
This is the deeper picks-and-shovels layer: compute and memory companies sell to hyperscalers; WFE companies sell to compute and memory companies. Whoever wins among the chip-makers, the WFE bill gets paid.
Practitioner note
For builders evaluating exposure:
- AMAT moves with WFE capex cycles, not directly with AI demand. There’s a 1-3 quarter lag between AI-driven chip demand and the resulting WFE orders. Watch foundry capex guidance (TSMC, Samsung) and memory capex guidance (SK Hynix, Micron, Samsung) — those are the leading indicators.
- Service revenue is the visibility floor. When systems revenue is cyclical, AGS is the baseline that doesn’t drop. Read the segment breakdown in quarterly filings — AGS gross margin tells you about pricing power even when system orders are soft.
- Sculpta adoption is the asymmetric upside. If 2nm and 1.4nm foundries adopt Sculpta at production volume, AMAT captures incremental TAM that would otherwise go to ASML High-NA. If foundries skip Sculpta and go straight to High-NA, the upside narrows. Track foundry tape-out announcements for 2nm production node tooling decisions.
- China export controls are the macro overhang. US export restrictions cap how much advanced-node equipment AMAT can ship to Chinese foundries. The bear case is a tightening that materially cuts addressable revenue. The bull case is that domestic-China substitution can’t actually replace AMAT at advanced nodes, so the cap is just a flat ceiling, not a contraction.
The under-considered angle: AMAT’s display business is a real option. Equipment for OLED panel manufacturing is a meaningful adjacency — same physics (deposition, etch, lithography-adjacent processes), same customer profile (large capex buyers). When the OLED panel cycle is healthy, it provides a second growth lever that doesn’t correlate to semiconductor cycles. The market mostly ignores it, but it’s there.