In fiscal 2025 Applied Materials turned over a record $28.37B (up 4%, its sixth straight up year) and earned $9.42 in non-GAAP EPS[1]. By June 2026, after a roughly 150% one-year run, it was worth about $398B — among the 35 most valuable companies on earth[3][4].

Almost every advanced chip on the planet passes through Applied's machines — it makes the deposition, etch, polishing, ion-implant and inspection tools that physically build transistors, and holds an estimated ~18% of the wafer-fab-equipment (WFE) market, the broadest portfolio of any single vendor[11]. The open questions are not whether it is an established franchise — it is one of only two near-full-line vendors — but whether its bet that materials engineering (not lithography alone) wins the AI era pays off[23], how much a shrinking China market costs it[28], and whether a mid-40s multiple is earned on a business this cyclical[42]. The evidence cuts both ways. This study lays out both cases; the verdict is yours.

The decisive questions

Each links to the section that lays out the evidence on both sides.

Five years of revenue

Total revenue, US$B, fiscal years ending in late October. FY2026 is an estimate implied by reported Q1 ($7.0B) and Q2 ($7.91B), Q3 guidance (~$8.95B) and management's “>30% equipment growth” framing — not a reported figure.

How to read this

Ten sections, each built the same way: a neutral synthesis, a two-sided case-for / case-against ledger, sourced data and charts, and dated facts. Start with the question that interests you, or read in order from the Overview.

Applied makes the materials-engineering tools — deposition, etch, planarization (CMP), ion implant, epitaxy and inspection — that physically build transistors, plus a large services arm and a small display business. It is the broadest single supplier in chip equipment[11], but twice in the last decade its bids to get broader still (Tokyo Electron, Kokusai) were undone by antitrust and Chinese regulators[6][7].

The business in one paragraph

Chipmakers like TSMC, Samsung, Intel, SK Hynix and Micron buy hundreds of process tools to turn blank silicon wafers into finished chips. Applied supplies the steps that add, shape and remove the atomic-scale layers of material on each wafer: deposition (CVD, PVD, ALD, epi), etch, chemical-mechanical planarization (CMP), ion implantation, rapid thermal processing and metrology/inspection[6]. It is one of only two vendors (with Tokyo Electron) offering a near-full line across these domains — the rest specialize. Around this equipment sits Applied Global Services, which keeps the installed base running and is now almost entirely recurring revenue[15].

How it got here

Founding to the current AI up-cycle. Dates and figures are sourced in the relevant sections.

The leadership question

Gary Dickerson has run Applied since 2013 and is the public face of the “materials engineering wins the AI era” strategy[6]. The $4B EPIC Center, opening in 2026 with Samsung and SK Hynix as partners, is the physical embodiment of that bet — co-developing process tech with customers years ahead of production[8]. Critics counter that the two deals that would have widened Applied's moat the most — Tokyo Electron and Kokusai — both collapsed, leaving organic R&D and minority stakes (Besi) as the growth levers instead of consolidation[7][9].

The wafer-fab-equipment market is an oligopoly: five companies (ASML, Applied, Lam, Tokyo Electron, KLA) hold ~70% of it[11]. SEMI forecasts record equipment sales of ~$139B in 2026 and ~$156B in 2027, driven by AI[10] — but the same market is deeply cyclical, and a chunk of it (China) is being walled off by export controls[14].

Where Applied sits

WFE is the equipment used to fabricate chips on silicon wafers. It splits into a few big domains: lithography (patterning — ASML's near-monopoly, including EUV), deposition and etch (adding and removing material — Applied, Lam and Tokyo Electron), process control / inspection (KLA), plus implant, CMP and thermal steps. Applied is the broadest single player, with an estimated ~18% of total WFE — second only to ASML by revenue[11][20].

Shares are third-party estimates and approximate — shown for relative scale, not precision. ASML's revenue-share lead reflects the price of EUV lithography systems; Applied, Lam and TEL compete across the deposition/etch domains[20].

The demand backdrop: AI up, China down

Two forces dominate the 2026 picture. On the upside, AI is pulling forward investment in leading-edge logic and in DRAM/HBM memory: SEMI projects 300mm fab equipment spending up ~18% to ~$133B in 2026 and ~$151B in 2027[12], and Yole sees the broader WFE-plus-services market reaching ~$184B by 2030[13]. On the downside, US export controls are shrinking the China portion of that market — the single biggest swing factor for Applied and its peers[14] (covered in China & Export Controls).

Why this market is hard to enter — and hard to hold

Decades of process IP, an enormous installed base and the need for customer trust make leading-edge WFE almost impossible to enter from scratch. But the market is not static: it is cyclical (capex booms and busts), concentrated on a handful of buyers, and now contested at the trailing edge by state-backed Chinese entrants gaining share in deposition and etch[19].

Two engines: Semiconductor Systems (~$20.8B in FY2025) sells the process tools and is capex-cyclical; Applied Global Services (AGS) (~$6.4B) services the installed base, with more than two-thirds of revenue from subscriptions and renewals above 90%[1][15]. The services tail smooths — but does not eliminate — the cyclicality of the systems business[17].

The three segments

Applied reports along three lines: Semiconductor Systems (the deposition, etch, CMP, implant, epi and inspection tools — the bulk of revenue and profit), Applied Global Services (spares, upgrades, 200mm equipment, and subscription service agreements that keep customers' fabs running), and a small Display & Adjacent Markets business plus corporate/other[1].

Why the services arm matters

Every tool Applied ships becomes a multi-decade annuity: chipmakers need parts, upgrades and process support for the life of the fab. AGS grew to a record ~$6.4B in FY2025, with >2/3 of revenue from subscriptions and >90% renewal rates[15]. From FY2026, Applied moved its 200mm equipment business out of AGS into Semiconductor Systems so that AGS becomes entirely recurring — sharpening the contrast between the cyclical equipment line and the subscription tail[16].

The economics

Applied runs a high-margin hardware model: FY2025 gross margin ~48.7% and operating margin ~29.2%, generating ~$7.0B of net income and ~$8.0B of operating cash flow[1]. But the cash conversion is cyclical — free cash flow fell ~24% to ~$5.7B in FY2025 as working capital and capex moved with the cycle[1]. And demand outside AI (memory, mature-node) recovered more slowly than hoped, a reminder the systems line still rides a capex curve[17].

Applied's rivals are formidable but largely non-overlapping: ASML owns lithography, KLA owns inspection, and Applied fights Lam and Tokyo Electron in deposition and etch[20]. Its real competitive questions are whether it can defend deposition/etch share against Lam and rising Chinese vendors, and whether it captures enough of the leading-edge value that ASML's monopoly keeps absorbing[19].

Five Forces

Click a force to see the rated pressure and the evidence behind it.

Positioning: breadth vs. leading-edge exposure

Two axes that actually separate this market: portfolio breadth (single-domain specialist → full process line) and leading-edge / AI exposure. Hover or tap a point for the basis.

Qualitative placements drawn from the cited competitive evidence, not precise scores.

The rivals, briefly

• ASML — the lithography monopoly (sole EUV supplier). Not a direct product competitor, but the benchmark for value capture: it earns the most per wafer and carries a ~$633B cap vs Applied's ~$398B[37].
• Lam Research — the closest head-to-head rival in etch and deposition, with a memory tilt (DRAM/HBM, 3D NAND); ~$17.8B revenue and a market cap that overtook Applied's in the AI re-rating[38].
• Tokyo Electron — the only other near-full-line vendor (coat/develop, etch, deposition, test); ~$16B revenue, significant China and mature-node exposure[20].
• KLA — dominates process control / inspection (~half that market); ~$12.2B revenue, an adjacent franchise rather than a direct overlap[39].
• NAURA, AMEC, Piotech — state-backed Chinese challengers gaining mature-node deposition and etch share; collectively ~6.5% of WFE in 2025, up from ~1.2% in 2021[19].

Applied argues that at 2nm and below, performance gains come less from shrinking features (lithography) and more from new materials and structures — gate-all-around transistors, backside power, HBM stacking, hybrid bonding — which lean on deposition, etch, CMP and packaging[23]. If true, the AI roadmap plays to its strengths. The counter: the most leading-edge-levered company by value is still ASML, and much of the optimism is already in Applied's price[27].

The three AI levers Applied is chasing

Applied frames AI chipmaking around three inflections, all materials-intensive:

• Leading-edge logic / GAA. In October 2025 it unveiled tools — Precision Selective Nitride PECVD, Trillium ALD and the Centura Xtera Epi system — to deposit metals and dielectrics at atomic tolerances inside gate-all-around transistors at 2nm and beyond[22].
• High-bandwidth memory (HBM). Applied positions itself as the leading materials-engineering supplier for the HBM inflection — DRAM scaling plus 3D stacking — central to feeding AI accelerators[25].
• Advanced packaging / hybrid bonding. Its ~9% stake in Besi and an extended co-development deal target die-to-wafer hybrid bonding, where chips are joined by direct copper-to-copper bonds for denser, more efficient systems-in-a-package[24].

Why the “co-optimization” framing matters

Applied's pitch is that no single step solves GAA or backside power; instead deposition, etch and CMP must be co-optimized together — and being the one vendor that owns all of them is the advantage[21]. Independent analysts describe its newest angstrom-class tools as squarely aimed at arming chipmakers for the GAA era[26]. The $4B EPIC Center is built to do exactly this co-development with customers ahead of high-volume production[8].

China was 30% of Applied's FY2025 revenue (28% of systems + services) and once neared 40%[1][30]. US controls already cut ~$400M from FY2025 and Applied flags a further ~$600M FY2026 hit — and says it can no longer supply China's memory or mature-node markets at all[28]. The open question is whether AI-driven demand elsewhere more than fills the hole.

The trajectory

China surged as a customer through 2024 — much of it mature-node capacity — then began retreating as successive US rules narrowed Applied's addressable market. The arc, as a share of total revenue:

FY2024's ~37% reflects a China mature-node spending surge; FY2025 is 30% of total revenue (28% of systems + services), with Q4 already down to ~25%[1][30].

What the controls actually do

A series of US export rules — tightened repeatedly through 2025 — restrict the sale of advanced chipmaking equipment to Chinese customers. Applied's CEO has said multiple trade-rule changes “reduced the size of the company's addressable market in China,” and that Applied can no longer supply China's memory-chip sector or its mature-node market[28]. The company quantified the FY2026 revenue impact at roughly $600M and expects China WFE spending to fall in 2026[28].

Sizing the hole: the offset arithmetic

A back-of-envelope walk, derived from the cited figures (illustrative, not company numbers): China at 30% of FY2025's $28.37B was roughly an ~$8.5B business[1]. The quantified control losses — ~$400M in FY2025[30] plus ~$600M in FY2026[28] — total about ~$1.0B, or ~3.5% of FY2025 revenue. Set against that: the two reported FY2026 quarters ($7.0B and $7.91B) plus the ~$8.95B Q3 guide put the year on a path toward roughly $33B[34][32] — an implied ~$4.6B revenue lift over FY2025. On those inputs, AI-driven demand elsewhere is generating roughly $7–8 of new revenue for every $1 of quantified China headwind this year. The arithmetic says the hole is real but currently overwhelmed — and that the offset only works for as long as the AI capex cycle does.

The double squeeze: controls plus domestic rivals

The controls don't just remove sales — they accelerate Beijing's push to build a domestic equipment industry. China's home-grown toolmakers (NAURA, AMEC, Piotech) lifted China's share of global WFE from ~1.2% in 2021 toward ~6.5% in 2025, concentrated in deposition and etch[19], and China's domestic-equipment adoption beat its 2025 localization target at ~35%[31]. So the part of the China market Applied can still address is also the part its new competitors are targeting.

FY2025 was a record: $28.37B revenue (+4%), $9.42 non-GAAP EPS (+9%), ~48.7% gross margin, ~$7.0B net income[1]. FY2026 is accelerating — Q2 was a record $7.91B (+11%) at a 50% non-GAAP gross margin, and management guides to >30% equipment growth for the year[32]. The caveat: free cash flow fell ~24% in FY2025, and the model is capex-cyclical[35].

FY2026 is inflecting upward

Reported quarterly revenue, US$B. Q3 is the company's guidance midpoint ($8.95B ±$0.5B), not a reported figure.

The cross-cycle picture

Across the cycle, Applied compounded steadily: $23.1B (FY2021) → $25.8B (FY2022) → $26.5B (FY2023) → $27.2B (FY2024) → $28.4B (FY2025), with gross margin climbing to ~48.7% and operating margin to ~29.2%[33]. Q1 FY2026 came in at $7.0B (+ AGS record $1.56B), topping guidance, before Q2 set a fresh record[34][32].

Margins, cash and capital return

Applied is highly profitable and returns most of its cash. In FY2025 it generated ~$8.0B of operating cash flow and ~$5.7B of free cash flow (down ~24% year over year), and returned ~$4.9B via buybacks plus ~$1.4B in dividends[1]. The FCF dip and the post-Q4-FY2025 stock wobble on the China warning are the reminders that this is a cyclical, not a smooth compounder[35].

Applied is #2 by revenue (~$28.4B) behind ASML (~$35B) and ahead of Lam, Tokyo Electron and KLA[41]. But by market cap, ASML (~$633B) and Lam (~$421B) both trade above Applied (~$398B)[37][38] — investors pay up for ASML's monopoly and Lam's memory leverage. Breadth is Applied's edge; value capture is its question.

Revenue (most recent fiscal year, US$B)

Fiscal years differ: ASML Dec-2025 (€32.7B ≈ $35B), Applied Oct-2025, Lam Jun-2025, Tokyo Electron Mar-2025 (¥2,431.5B ≈ $16B), KLA Jun-2025. Currency conversions are approximate.

Market capitalization (June 2026, US$B)

Side by side

Revenues are reported fiscal-year figures (currency-converted where noted); market caps are point-in-time (~June 2026) and approximate. Sources are listed in full on the Sources page.

The skeptics' case rests on four pillars: cyclicality (AMAT draws down harder than the market in shocks)[43], concentration on a few buyers, China/policy risk[44], and valuation — a mid-40s trailing P/E that prices in an uninterrupted AI super-cycle[42]. The bull rebuttal: a diversified toolmaker “wins no matter which chip wins”[45].

SWOT

Applied even-handedly; each item is sourced in the relevant section.

Parenthetical (s#) refer to source ids on the Sources page.

The valuation debate

After a ~150%+ one-year rally to a ~$398B market cap, Applied trades at a trailing P/E in the mid-40s; some analyses put the forward multiple far higher and conclude the stock is richly valued[42][5]. Commentators question whether the upside is exhausted given cyclicality, concentration and China limits[44].

That is the crux: bulls argue Applied is the “picks-and-shovels” supplier that benefits from rising chip complexity regardless of which company or architecture wins[45]. Bears reply that the same logic applied to the last build-out's infrastructure darlings — and that a cyclical business at a mid-40s multiple has little margin for an air-pocket[43].

The weighing

Four questions organize this study. Here is where the evidence leans, at what confidence, and the dated signals that would change each call. None of this is a buy or sell view — it is a scorecard of the arguments.

On whether materials engineering — not just lithography — wins the AI era: the evidence leans partly toward Applied (medium confidence). The controlling evidence is that leading-edge foundry logic, DRAM and advanced packaging account for more than 80% of year-over-year WFE growth in 2026 — Applied's deposition, etch, CMP and packaging turf[32] — and that Applied has shipped concrete angstrom-class tools (Precision Selective Nitride PECVD, Trillium ALD, Centura Xtera Epi) aimed squarely at 2nm GAA[22]. That outweighs the pure-narrative objection because the WFE-growth mix is third-party-corroborated, not just Applied's own framing. The strongest surviving counter-argument: lithography stays indispensable and ASML still captures the most value per wafer — its ~$633B market cap sits ~60% above Applied's ~$398B precisely because the EUV monopoly absorbs the most leading-edge value[37]. What would flip this reading: Applied's Semiconductor Systems growth lagging total WFE growth for two consecutive quarters in the FY2027 earnings releases; or ASML's lithography revenue growing materially faster than Applied's systems revenue across 2026. Pre-mortem: if this looks wrong in two years, the most likely reason is that the GAA/HBM inflection enriched the litho and process-control layers more than the deposition/etch layer — or, on the other side, that “co-optimization” across owned process steps proved a wider moat than skeptics priced.

On how much the China squeeze really costs: the evidence leans toward “real but currently overwhelmed” (medium confidence). The controlling evidence is that the quantified control losses — ~$400M in FY2025[30] plus a flagged ~$600M in FY2026[28] — total ~$1.0B against a China business that was ~30% of FY2025's $28.37B, or roughly $8.5B[1], while the two reported FY2026 quarters ($7.0B and $7.91B) plus the ~$8.95B Q3 guide put the year on a path toward ~$33B[32][34] — a far larger lift than the headwind. That outweighs the bear read for now because the offset is arithmetic, not hope. The strongest surviving counter-argument: whole China segments (memory, mature node) are now permanently off-limits, and home-grown rivals NAURA and AMEC lifted China's WFE share from ~1.2% (2021) to ~6.5% (2025) in Applied's core deposition/etch domains, with domestic adoption already at ~35%[19][31] — so the offset only holds while AI capex does. What would flip this reading: China falling below ~20% of systems+AGS revenue while total revenue also declines in any FY2027 quarter (offset failing); or a further export-rule round adding a quantified headwind above ~$1B in a single fiscal year. Pre-mortem: if this looks wrong in two years, the most likely reason is that Chinese domestic tools displaced Applied faster than AI demand could refill — or, on the other side, that the China line proved a small, already-de-risked tail (it was ~24% by Q2 FY2026[32]) the market over-weighted.

On whether this is a durable growth story or a well-run cyclical: the evidence is contested (contested confidence). What deadlocks it: the same numbers support both reads. Applied posted six straight record years — $23.06B (FY2021) rising to $28.37B (FY2025) — with gross margin climbing to ~48.7%[33], and FY2026 is guided to >30% equipment growth[32]; yet free cash flow fell ~24% in FY2025 and the stock wobbled on the Q4 China warning[35], and across 15 systemic shocks AMAT drew down ~−23% versus ~−16% for the S&P 500[43]. Record revenue and a steeper-than-market beta are both true; the question is purely about the next turn of the capex cycle, which neither side can yet observe. What would flip it toward “cyclical”: two consecutive quarters of year-over-year systems-revenue decline, or SEMI revising its ~$139B 2026 equipment forecast[10] downward in a subsequent update; toward “durable”: FCF re-expanding above the FY2024 level alongside continued AGS recurring growth through FY2027. Pre-mortem: if this looks wrong in two years, the most likely reason is that an AI-capex air-pocket exposed the old cyclicality — or, on the other side, that a structurally higher recurring-services base genuinely smoothed the trough.

On whether the price is earned after a ~150% run: the evidence is contested (contested confidence). What deadlocks it: at a ~$398B market cap, AMAT trades at a trailing P/E in the mid-40s with forward multiples cited far higher[42], which prices in an uninterrupted super-cycle and leaves little room for a cyclical air-pocket given its sharper-than-market drawdowns[43] — yet the bull frame that a diversified toolmaker “wins no matter which chip wins”[45] is corroborated by the >30% FY2026 equipment guide[32], so a rich multiple on rising estimates is not self-evidently wrong. The valuation case and the growth case rest on the same forward estimates pulling in opposite directions. What would flip toward “not earned”: any quarter where guidance is cut and the multiple does not compress to absorb it; toward “earned”: the forward P/E falling into the 20s on delivered earnings without a price decline through FY2027. Pre-mortem: if this looks wrong in two years, the most likely reason is that the mid-40s entry multiple compressed in a capex downturn faster than earnings grew — or, on the other side, that AI WFE demand compounded long enough to grow into the multiple.

How the research was done

We ran fan-out web search and fetched the underlying pages: Applied's FY2025 results and FY2026 quarterly releases and earnings-call transcripts[1][32], company newsroom and product announcements[22][24], third-party market data (SEMI, Yole) for industry sizing[10][13], reputable trade and financial press, and peer disclosures (ASML, Lam, TEL, KLA)[36][39]. Every URL cited was opened during the run; the bibliography is machine-checked for dead links.

Frameworks used

• Pyramid Principle — answer-first synthesis (the balanced state of the debate), then evidence.
• Porter's Five Forces — industry structure of wafer-fab equipment.
• 2×2 positioning — breadth vs. leading-edge exposure across the Big Five plus China.
• Peer comparables — revenue and market cap vs. ASML, Lam, Tokyo Electron, KLA.
• SWOT — applied even-handedly, with weaknesses and threats given equal weight.

Disclosed vs. estimated

Applied's revenue, EPS, margins, segment splits and capital return are disclosed figures from its results[1]. Treat as estimates: WFE market size and the ~18% market-share / mix chart (third-party estimates that vary by source)[11]; the FY2026 full-year revenue (~$33B, our estimate from reported quarters and guidance, not a company number); peer market caps (point-in-time, ~June 2026); and currency conversions for ASML and Tokyo Electron. Valuation multiples change daily.

Neutrality

Each section presents both supporting and countervailing evidence and a two-sided ledger. The source base is tagged by stance — roughly a third supporting, a third critical, a third neutral — to keep the compilation balanced rather than advocating. Where we interpret, we say so and show the basis. The judgment is left to you.

Cross-checked at build time by an automated link checker. A few primary filings (SEC EDGAR) and some news sites bot-wall automated fetchers; the equivalent figures here are taken from Applied's own newsroom / GlobeNewswire releases and reputable press, which were fetched and read. See Methodology & Limits.