---
title: "AMAT Trans: 2030 target to add 2ppt WFE share; GPM to 54%."
type: "Topics"
locale: "en"
url: "https://longbridge.com/en/dolphin/post/44318252.md"
description: "Below is Dolphin Research's Trans from AMAT's 2026 SEMICON West investor tech session. Key takeaways: AI is concentrating incremental WFE on the company's leading positions in advanced logic, DRAM, and advanced packaging. By 2030, management targets GPM of 54%, OPM of 40%, and 200bps+ share gains. However, it gave no WFE outlook and flagged that cleanroom and supply-chain constraints could make growth uneven, while betting on EPIC and AIx services to drive yield innovation..."
datetime: "2026-10-14T03:36:24.000Z"
locales:
  - [en](https://longbridge.com/en/dolphin/post/44318252.md)
  - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/44318252.md)
  - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/44318252.md)
author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)"
generator: "portal-rs"
---

# AMAT Trans: 2030 target to add 2ppt WFE share; GPM to 54%.

**Compiled by Dolphin Research:**$Applied Materials(AMAT.US) **SEMICON West 2026 Investor Tech Brief (Trans):**

**Key takeaways**: AI is concentrating WFE upside in the company's lead positions across advanced logic, DRAM and advanced packaging. Targets for 2030 are non-GAAP GPM 54% and OPM 40%, plus >200bps share gain; no WFE forecast given, with cleanroom and supply-chain constraints implying uneven growth. EPIC and AIx services are core bets to drive yield innovation.

**I. Core highlights**

1\. **Shareholder returns**: Committed to returning 80%–100% of FCF to shareholders over the long term. a. Dividends: Since initiation in 2005, DPS grew at a 14% CAGR and doubled over the past four years; next year's dividend policy will follow BOD routine. b. Buybacks: Share count down >50% over the past decade.

2\. **2030 targets (no WFE forecast, company framework only)** a. Share: WFE share to rise by >200bps; model semi equipment to 'outgrow the industry'. b. Margins: Non-GAAP GM 54%, OP 40%; path is +400bps GM (approx. +100bps per year, non-linear) plus 200–300bps opex leverage. c. Other: 'Other' businesses (Display, SENZ waveguides, etc.) to exceed $2bn by 2030 (near term still model at ~$400mn per quarter); tax rate est. 15% in 2030.

3\. **10-year financials and GM drivers** a. Revenue CAGR 12.2% to $34.3bn; non-GAAP GM up 680bps to 50%, OPM up >1,100bps; EPS CAGR 22%, Avg. FCF margin 21%, Avg. ROIC >30%. b. Primary GM driver is a more enabling mix plus value pricing implemented over the past four years, followed by higher mix of semi systems with GM >55%, then manufacturing in Asia (Singapore) and supply-chain diversification efficiencies.

4\. **Investment and cost discipline** a. >50% of operating cash reinvested into the biz.; manufacturing and lab capacity planned to more than double (higher than last disclosed). b. Capex ~8% of revenue this year; no fixed target, ratio to trend down over time. c. Opex ratio cut from 22% ten years ago to 17%; R&D rose from $1bn to $4.2bn, now 72% of opex; SG&A to gain significant efficiency via AI and automation.

**II. Investor Day details**

**2.1 Management remarks**

1\. **Industry and demand (CEO)** a. Data centers have displaced smartphones as the main driver of advanced logic wafer starts, with DRAM likewise led by DC/HPC; Agentic AI compute demand surges in 2026, prompting an upward WFE view, with Physical AI ahead. b. Agentic AI pods consume 2–4x the wafers of GenAI pods, shifting the architecture from GPU-heavy to more CPU; ~85% of wafer starts are in advanced logic, DRAM and advanced packaging, and ~80% of this year's WFE upside is concentrated in these three, where the company ranks No.1 in each. c. Growth constraints: cleanroom space, supply-chain parts, energy and financing; value pricing can lift WFE growth. No specific WFE forecast, only that CAGR will be 'very strong'.

2\. **Semi Systems: share and SAM** a. Company SAM is ~55% of WFE; by 2030, ~75% of WFE growth will come from areas where the company is dominant, with >200bps share gain expected (including M&I outgrowth). b. Advanced logic: SAM rises from $12bn at 3nm to $15bn at 1.4nm, share +300bps; serviceable market share rises to >50%. c. DRAM: 6F² SAM ~$6bn, 4F² adds >40%; 3D DRAM higher, SAM share to rise +500bps, serviceable market share >60%, with ~+1,000bps lifetime DRAM share gain already. d. Tech drivers: from smartphone chips to Agentic AI, logic silicon area up 70x, transistor count up 20x, metal layers tripled; HBM has >2x the routing layers vs. conventional DRAM. e. Transistor contact co-optimization can deliver ~5% perf uplift vs. customer targets of ~15% per node. f. Products: Copper interconnect revenue to reach $3bn in 2026; integrated solutions mix in systems up from 20% to 30%; 13 new products launched in the past 12 months; deposition scale ~2.5x No.2, materials modification >2x.

3\. **Advanced packaging and ICAPS** a. Advanced packaging revenue to exceed $5bn within a few years; HBM packaging SAM share >60%, with 2026 revenue >$1bn, covering 3/4 of process steps. b. Hybrid bonding (with Besi) covers ~80% of front- and back-end steps and will double HBM SAM vs. micro-bump+TSV; panel-level packaging footprint completed via Tango, NEXX M&A. c. In ICAPS, Power and Photonics revenue exceeds $1bn this year; 28nm migration drives 2026 growth, with Physical AI as the longer-term driver.

4\. **Metrology & Inspection (M&I)** a. Optical inspection grows ~50% ('25→'26), currently addressing ~$2bn of a ~$9bn market, with the remaining ~$7bn TAM to be entered via new products over the next few years. b. eBeam also grows ~50%; cold field emission drives revenue to ~2x No.2 and ~200x optical resolution; eBeam metrology to reach ~$2bn revenue, and M&I systems+services ~$3bn in 2026. c. New markets: online, non-destructive film thickness/composition metrology using cold field emission, sized at hundreds of millions of dollars and not yet in reported numbers; advanced packaging optical inspection market expanding from hundreds of millions to well over $1bn, with a new platform launched.

5\. **Services (AGS)** a. 2026 services revenue to grow >20%, with long-term CAGR raised to mid-teens; AIx-connected chambers to rise from 37k to >90k by 2030, with ~80% remote. b. Subscription mix to climb from low-40% to near 70% by 2030; contract terms >2.5 years and renewal rate >90%; per-tool service revenue up ~20% in 2025–2030. c. Cases: AIx accelerates fault handling by ~30%; CrossMatch boosts match rate by ~50%; ESC failures predicted 30 days in advance with >90% accuracy; HDP CVD maintenance iterations cut from 6 to 2, ~65% time saved.

6\. **EPIC and internal AI** a. EPIC centers bring together leading chipmakers and hyperscalers to co-develop in parallel on N+1 to N+4 nodes, halving the cycle from R&D to commercialization, enabling earlier design-ins. b. Internal AI: hardware development pilots accelerate >4x and scale to 20 products; 100% of software developers use AI, with Avg. efficiency +44%; >80% of employees use it weekly, with tokens consumed up 45x YoY.

**2.2 Q&A**

**Q: How do you view China's push into 4F² DRAM, and if successful, what would it mean for overseas equipment vendors' share?** A: **No direct comment on China's 4F² progress; management emphasized the top three memory makers will remain in the lead.** Innovation across memory-logic co-optimization is moving very fast, with deep collaborations between top fabs and AI server architects requiring quicker iterations, underpinning high confidence in DRAM growth and device, packaging and system-level innovation.

**Q: How real is 'de-spec' memory in AI workloads, and will it compress storage TAM?** A: **DRAM is likely to stay tight for a while; de-spec is load reallocation, not cannibalization.** System architects are tuning memory by workload—training, inference, Agentic—to maximize compute output. Training still pushes for the highest HBM stacks; newer workloads may not fully utilize top HBM stacks, but they are incremental demand, with memory placed elsewhere in the system—partitioning, not substitution—and total demand still rising.

**Q: How are you positioned in optical networks, from pluggables to CPO, waveguides, hybrid bonding and new materials?** A: **Only qualitative: Photonics and Power are high-growth, no specific optical networking blueprint disclosed.** Hybrid bonding delivers better performance, power and thermal; in some architectures, yield is the key, and the company is advancing with leading customers, with yield progress determining the inflection timing.

**Q: A 54% GM by 2030 implies ~+400bps in four years. Structure alone can contribute 100–200bps—are the targets conservative?** A: **54% is the right target, with ~+100bps per year but non-linear.** Tool GM is already >55%. The company must absorb new plant and logistics infrastructure, plus headwinds from fuel and rare earths; meeting the target requires continued rollout of more enabling tools, with price increases offsetting part of cost inflation.

**Q: Share fell in 2024–2025; is the 2030 share gain just regaining lost ground?** A: **Targets already factor in growth from restricted China customers; excluding controls, addressable share is rising.** The company cannot serve advanced logic, DRAM and NAND in China, yet models still assume growth in these segments with trade rules unchanged; in the serviceable market, share gains would be larger. Quantitatively, the company's reachable share of China WFE was ~100% in early 2020s, fell to ~90%, then ~70% in 2025, with no easing thereafter. Excluding this factor, share is rising, and performance is strong even at 28nm in China.

**Q: Advanced packaging optical inspection is competitive. What differentiates the new product set launching in 1H27?** A: **No specifics disclosed; the new platform is architecturally differentiated, focused on highest sensitivity and throughput.** The company spent ~1.5 years evaluating existing solutions and roadmaps, and would not enter without architectural differentiation. Historically, cost sensitivity constrained product investment; now advanced packaging yield is highly valuable, and customers are willing to accept higher ASPs.

**Q: With connected chambers CAGR near 20% and per-tool revenue +20%, why are services only mid-teens?** A: **Metric is connected chambers, not tools; new tools transition post-warranty and not all shift to AIx services.** After installation, tools enter a 12–18 month warranty; thereafter, based on customer needs and AIx ROI, they move into tiers of service (parts, predictive parts mgmt, full-kit mgmt to AIx), not a 1:1 conversion. The +20% per-tool uplift comes from new AIx capabilities; the ~20% services growth this year reflects utilization recovery and AIx penetration. Installed base is near 60k tools; mid-teens growth combines 5%–10% installed-base growth per year plus contributions from new service offerings.

**Q: Agentic AI may imply unlimited compute demand. What truly constrains equipment demand today?** A: **Cleanroom capacity is the primary constraint, setting the growth cadence.** The company builds medium-term forecasts from large customers' rolling eight-quarter plans (by fab, monthly capacity, nodes and tools), with cleanroom expansion speed as the key variable. The company tracks >100 fab projects globally, with >10 added per quarter in the past two quarters; the pool of usable fabs and cleanroom space is expanding.

**Q: Is the plan to double capacity an industrywide, synchronized move, and will it create bottlenecks?** A: **Peers are likely investing similarly; there are chokepoints to manage, but no fundamental barriers.** The company shares forecasts and parts needs with >2,000 suppliers, and the supply chain is lifting capacity by broadly similar magnitudes. Data centers are also competing for resources; smartphones and PCs face capacity limits from memory shortages, so resource competition will impose some constraints.

**Q: How do you reach 40% OPM, and is there an FCF target?** A: **Another +400bps GM, plus 200–300bps opex leverage.** Leverage comes from AI and automation, selling more into the same customer set without scaling corporate functions at revenue pace. No explicit FCF target; historical Avg. >21%, and while ongoing investment is needed, management expects to run slightly above that level.

**Q: Are customers shifting non-critical process steps to competitors, and how do you respond?** A: **Gains or losses in non-critical steps matter far less than design-in; no direct confirmation of share loss.** ~80% of customers' WFE additions are in advanced logic, DRAM and advanced packaging, with focus on performance and yield; cost is rarely the primary discussion. The company continues to innovate across the full portfolio, but that is not customers' top priority today.

**Q: Can the supply chain resolve bottlenecks early in 4–5 years, and will the company fund supplier expansions?** A: **No structural barriers in the medium term, but growth will be uneven; no comment on funding suppliers.** Suppliers agree on the AI demand signal without needing repeated validation and are investing, hiring and expanding. Chokepoints include hiring, tooling and test capacity that require lead time; the company raised supply-chain expectations every quarter this year, asking them to move at full speed.

**Q: Any changes to pricing in the current environment?** A: **Mechanisms and processes are unchanged; what's changed is solution value.** For more enabling solutions, the company prices to the value in yield and output, and continues to enforce pricing discipline; rising industry value is also a factor in pricing. 

**Q: Semi revenue was ~$1tn last year, with WFE at ~$115–120bn, or ~12% intensity. Will intensity rise or fall, and which categories—etch, dep, litho—grow faster?** A: **Internal forecasts do not use intensity; near-term intensity is driven mainly by memory and GPU pricing.** More process and packaging steps push intensity higher, but the company forecasts off fab and end-market growth. WFE per wafer is rising materially in each advanced logic and DRAM generation; 3D architectures and multi-chip interconnects depend more on materials innovation, lifting both SAM and share. SAM was raised in parts of the portfolio: process complexity is adding steps like treatment and anneal—many are customers' non-public IP—and thermal interface materials in packaging are adding new steps.

**Risk disclosure and statement:**[**Dolphin Research Disclaimer & General Disclosure**](https://support.longbridge.global/topics/misc/dolphin-disclaimer)

### Related Stocks

- [AMAT.US](https://longbridge.com/en/quote/AMAT.US.md)
- [WF.US](https://longbridge.com/en/quote/WF.US.md)
- [AMA.US](https://longbridge.com/en/quote/AMA.US.md)

---
> **Disclaimer: This article is for reference only and does not constitute any investment advice.**