Variant-perception thesis · personal research變異觀點投資論點 · 個人研究
Applied Materials (AMAT)
The picks-and-shovels enabler of chip complexity. The reasoning is the model's; every number carries its source and date. Where the first draft ran on memory, live data rewrote the conclusion.晶片複雜度背後「賣鏟子」的設備商。推論由模型產生，但每個數字都標註來源與日期。第一版靠記憶寫成的地方，即時資料改寫了結論。
How this is built. A variant-perception framework — the discipline of ignoring consensus to find what the market may be misunderstanding, underestimating, or ignoring. It walks the same nine steps for any stock: business → bull → bear → the variant view → catalysts → management → moat → probability-weighted odds → what it takes to double. The rule that makes it trustworthy: no number is written from memory — every figure is grounded to a filing or live data with a date, and the share price is recomputed the moment this page is built.
YTD · -24% off 07-01 high $739.67距 07-01 高點 $739.67 -24%✓ IBKR
~46×
Fwd P/E預估本益比live ÷ FY26 EPS現價÷FY26
$28.37B
FY25 rev年營收✓ 10-K
Forward P/E = live $560.00 ÷ FY26 consensus EPS $12.26 (stockanalysis, 39 analysts, 07-15; supersedes the ~$13.0 secondary figure used on 06-30). On FY27 consensus EPS $16.75: ~33×.預估本益比 = 即時 $560.00 ÷ FY26 共識每股盈餘 $12.26（stockanalysis，39 位分析師，07-15；取代 06-30 版所用的次級來源 ~$13.0）。若以 FY27 共識 $16.75 計：約 33 倍。
headlineThe bull re-rating has largely already happened多方期待的「評價重估」其實大半已經發生
The variant bull case is "the market underestimates how long AI-driven chip complexity lifts equipment demand, so a re-rating is coming." AMAT peaked at $739.67 intraday around 07-01, then gave back -24% in three weeks to $560.00 (07-22 close, still +118% YTD) ✓ IBKR · 07-22. Forward P/E is now ~46× on FY26 consensus EPS $12.26 (39 analysts) ✓ stockanalysis · 07-15, ~33× on FY27's $16.75; trailing ~52× TTM ✗→ was ~55×/~76× at 06-30 prices06-30 版為 ~55×/~76×.
The 06-30 finding said the re-rating was largely done and warned: priced for perfection — any wobble compresses the multiple. The wobble came (07-02 HBM-slowdown scare, 07-07 Samsung print), and the multiple compressed — while AMAT's own fundamentals strengthened (FQ3 guide +13% QoQ, TSMC capex raised to $60-64B, customer visibility "to 2030"). The question has moved from "is there edge after a double" to "does FY27's +37% consensus EPS survive the memory-capex cycle" — see the 07-23 update below and the playbook at the end. ◆
Update log — material news since publish更新紀錄 — 發布後的重要消息
Dated additions, newest first. The core thesis above stays as-is unless an entry says otherwise.
逐則標日期、最新在上。除非條目另外註明，上方核心論點維持不變。
✓ 2026-07-23-24% in three weeks: the "any wobble" bear leg fired — on sentiment, not on AMAT's numbers三週 -24%：「一有閃失」的空方句兌現了 — 但閃失來自情緒，不是 AMAT 的數字
The tape. $723.00 all-time closing high 06-30 → $739.67 intraday ~07-01 → -7.4% on 07-02 (report that SK Hynix is slowing part of its HBM4 expansion) → ~-10% on 07-07 (Samsung's record ₩89.4T preliminary OP still missed the "AI bar") → $529.66 trough 07-17 (~-27% off the closing high) → rebound after CEO Gary Dickerson described customer equipment forecasts extending "to 2030" → $560.00 close 07-22✓ IBKR · 07-22 + CNBC/MacroTrends. Michael Burry disclosed a short position at ~$729 on 06-30 ⚠ single outlet單一來源.
Numbers refreshed. Price $713.00 → $560.00 (-21.5% vs the 06-30 build; IBKR 07-22); YTD +177.8% → +118.2%; market cap ~$560B → $444.6B (793,959,430 shares, 10-Q 04-26 ✓ SEC); forward P/E ~55× → ~46× — and the FY26 consensus EPS denominator is corrected to $12.26 (stockanalysis, 39 analysts, 07-15) ✗→ replaces the ~$13.0 secondary figure取代 ~$13.0 次級數字; FY27 consensus EPS $16.75 (+37%) → ~33× at the live price. CEO Dickerson sold 78,321 shares for ~$55.5M at ~$700-736 on 06-29/30 — at the top, filed 07-01 ✓ SEC Form 4.
How it bears on the thesis. The 06-30 warning — priced for perfection, any wobble compresses — executed to the letter, except the wobble was the customers' capex sentiment, not AMAT's own numbers. The de-rating the finding said was needed has begun (~55×→~46×). The live question is now FY27: consensus embeds +37% EPS growth that assumes the 2027 memory-capex wave lands intact — the same wave whose durability is the open question across this desk's memory coverage. This update also adds the playbook section below (the page predates the decision-layer format). ◆ framing = judgment定位＝判斷
✓ 2026-06-30Numbers refreshed — the re-rating kept going, now on rising estimates數字已更新 — 評價重估還在續，且這次是靠獲利上修
Re-pulled live: $713.00, +177.8% YTD, a fresh 52-week high✓ IBKR · 06-30 — up another ~21% from the $589 build price. Yet the forward P/E fell to ~55× (vs ~57× at build), because consensus FY26 EPS was revised up to ~$13.0 (from ~$10.4) and the FY26 revenue forecast lifted to ~$33.3B (from ~$31.6B) ✓ Yahoo / Simply Wall St · ~06-26. Driver: a high-conviction product showcase plus a wave of Wall-Street target upgrades.
How it changes the finding. The headline above said the re-rating was largely done and a further double would need earnings growth (engine 1), not multiple expansion. That is now visibly happening: the multiple compressed while the price rose, because the E got marked up — engine 1 is firing, partly validating the variant bull. The flip side hardens too: at +178% YTD and a fresh high (~76× trailing on FY25 EPS; TTM since risen), the "priced-for-perfection, any-wobble-compresses" bear is sharper, and "edge after a double" now means going from ~$560B to ~$1.1T+. Net: not more bullish, not more bearish — the core tension is resolving in real time, with estimates doing the work the thesis said they would have to. ◆ framing = judgment定位＝判斷
✓ 2026-06-16AMAT × EssilorLuxottica — a new AR-optics growth vectorAMAT × EssilorLuxottica — 一條新的 AR 光學成長軸
On 2026-06-16, EssilorLuxottica (the Ray-Ban / Oakley owner) and Applied Materials signed a long-term joint development agreement for next-generation intelligent optical systems in AR / AI smart glasses. It pairs EssilorLuxottica's lenses-frames-eyewear scale with AMAT's materials engineering and waveguide expertise; scope covers waveguides, adaptive (auto-dimming) lenses, materials innovation and advanced lens packaging, with a dedicated joint lab on AMAT's Silicon Valley campus. ✓ GlobeNewswire · 06-16
2026 年 6/16，EssilorLuxottica（Ray-Ban／Oakley 母公司）與應用材料簽署長期共同開發合約，鎖定 AR／AI 智慧眼鏡的次世代智慧光學系統。結合 EssilorLuxottica 的鏡片－鏡框－眼鏡規模，與 AMAT 的材料工程與導波路（waveguide）技術；範圍涵蓋導波路、適應型調光鏡片、材料創新、先進鏡片封裝，並在 AMAT 矽谷園區設專用聯合實驗室。✓ GlobeNewswire · 06-16
How it bears on the thesis. This is optionality, not a near-term number. It supports the variant view that AMAT's materials-engineering moat is broader than core semi WFE — extending into a large new AR-optics TAM. But it is an early-stage R&D agreement (a joint lab, no disclosed revenue or timeline), so it does not change the headline finding above: at ~57× forward P/E the re-rating is largely done, and edge-after-a-double still depends on earnings growth (engine 1), not this. Treat it as a multi-year call option on AMAT's TAM, not a 2026 catalyst. ◆ framing = judgment定位＝判斷
對論點的意義。這是選擇權式的想像空間（optionality），不是近期數字。它支持「AMAT 的材料工程護城河不只在核心半導體 WFE、能延伸到一塊很大的新 AR 光學 TAM」這個變異觀點。但它是早期 R&D 合約（只設聯合實驗室，未揭露營收或時程），所以不改變上方的結論：約 57 倍預估本益比下，評價重估大半已發生，翻倍後還有沒有超額報酬，仍取決於獲利成長（引擎一），而非此案。把它當成 AMAT 市場規模（TAM）的「多年期買權」，而不是 2026 年的催化劑。◆ framing = judgment定位＝判斷
Jargon decoder — read this first術語解碼器 — 先看這個
AMAT sells the steps that turn a blank silicon wafer into a chip. To judge the thesis you have to know what those steps are. Here is the wafer's journey, then a plain-language glossary.
AMAT 賣的是把一片空白矽晶圓變成晶片的那些製程步驟。要判斷這個論點，得先知道這些步驟是什麼。先看晶圓的旅程，再看白話術語表。
The whole bull case in one picture: AI chips need more layers, so the loop runs more times, so AMAT sells more tools. AMAT leads every step here except lithography (step 2 — that's ASML's EUV).一張圖看懂多方論點：AI 晶片需要更多層 → 循環跑更多次 → AMAT 賣出更多機台。圖中每一步幾乎都是 AMAT 主場，只有微影（第 2 步）是 ASML 的 EUV。
And here is what those steps physically do to the silicon — building one metal wire, in cross-section (the real copper "damascene" process):
This loop builds just one wiring layer. An AI chip stacks 10–15+ of them — so the loop runs 10–15+ times, and AMAT sells a tool for nearly every pass. More complexity = more passes = more AMAT.這套循環只做出一層線路。一顆 AI 晶片要疊 10–15 層以上 — 循環就跑 10–15 次以上，幾乎每一輪 AMAT 都賣得到機台。越複雜 = 跑越多次 = AMAT 賣越多。
Step 1 (deposit) isn't one thing — it's a ladder of tools for ever-harder 3D structures: PVD → CVD → ALD. The more vertical the chip, the more you need the precise (slower, pricier) end. This is AMAT's backyard.
第 1 步（沉積）不是單一技術，而是一排對應越來越難 3D 結構的工具：PVD → CVD → ALD。晶片越立體，越需要精準（但更慢、更貴）的那一端。這正是 AMAT 的主場。
PVDphysical · sputter物理氣相沉積
Blast a solid target into atoms that fly straight down. Fast, high throughput — great for metal layers, barriers, liners. Struggles to reach the bottom of deep 3D structures.把固態靶材打成原子，直線飛濺到晶圓上。快、吞吐量高 — 金屬層、阻障層（barrier）、襯墊層（liner）的主力。但難顧到深 3D 結構的洞底與側壁。
CVDchemical · grows化學氣相沉積
Gaseous precursors react on the surface — the film "grows." Better coverage and uniformity than PVD; the production workhorse. Hits its precision limit as structures get more 3D.氣態前驅物在表面反應，薄膜是「長出來」的。覆蓋性與均勻性優於 PVD，量產主力。結構越 3D，會碰到精度極限。
ALDatom by atom原子層沉積
Self-limiting — each cycle grows exactly one atomic layer, almost regardless of shape. Indispensable for FinFET, GAA, high-aspect-ratio. Slow and pricey, so used only where nothing else works. Key for GAA.自限制反應 — 每循環只長一層原子厚度，幾乎不受結構形狀影響。FinFET、GAA、高深寬比必備。慢又貴，只用在非用不可處。GAA 關鍵。
AMAT is the deposition leader — #1 in PVD and a top player in CVD and ALD. (EPI — growing a single crystal for the transistor channel — is a fourth AMAT strength.) Illustrations are conceptual.AMAT 是沉積龍頭 — PVD 市占第一，CVD 與 ALD 也是要角。（另有磊晶 EPI，長單晶當電晶體通道，也是 AMAT 強項。）圖為示意。
Whatever the method, deposition is judged on three things — and 3D chips make all three harder:
不論用哪種方法，沉積的好壞看三件事 — 而 3D 晶片讓這三件事都更難：
Every method chases the same three: even thickness everywhere (uniformity), full coverage down into 3D trenches (conformality), a smooth top (surface smoothness). The deeper and narrower the structure, the harder all three get — which is why the precise tools (ALD) keep gaining share. Recreated from a vocus.cc explainer.每種方法追求的都是同三件事：到處厚度一致（均勻性）、連 3D 深溝底部都完整包覆（覆蓋性）、表面平整（表面平整度）。結構越深越窄，這三件事越難 — 這正是精準工具（ALD）市占持續成長的原因。圖表重製自 vocus.cc 解說。
The process steps + 2 key terms製程步驟 ＋ 2 個關鍵詞
Deposition薄膜沉積（Deposition）
Lay an ultra-thin film onto the wafer. Four tools, by rising precision: PVD (sputter a solid, fast), CVD (gas grows a film, the workhorse), ALD (one atom-layer at a time, for 3D), EPI (grow a single crystal). See the diagrams above.
More AI = more complexity stacked per chip = more process steps AMAT sells into.AI 越多 = 每顆晶片堆疊的複雜度越高 = AMAT 能賣進去的製程步驟越多。
AMAT is the world's broadest supplier of materials-engineering equipment for making chips — the deposition, etch, CMP, implant and metrology tools from the decoder above. As chips move to GAA transistors, backside power and HBM stacking, each wafer needs more steps and higher-value tools. AMAT sells the steps.
AMAT 是全球產品線最廣的材料工程設備供應商 — 也就是上面解碼器裡的沉積、蝕刻、CMP、佈植、量測機台。當晶片走向 GAA 電晶體、晶背供電、HBM 堆疊，每片晶圓需要的步驟更多、機台單價更高。AMAT 賣的正是這些步驟。
How it makes money怎麼賺錢 FY2025 ✓ 10-K
Semiconductor Systems — $20.8B (73%). Capital tools to TSMC, Samsung, SK Hynix, Intel.
AGS — $6.4B (23%). High-margin recurring service on the installed base.
Rising chip complexity = more materials-engineering steps per wafer. AMAT is the diversified toll-taker on that complexity, not a bet on one node or customer.
GAA transistors, backside power and HBM/advanced packaging each add deposition, etch, CMP and metrology steps — AMAT's strongest areas. AI capex could stay high for years.
GAA 電晶體、晶背供電、HBM／先進封裝，每一項都增加沉積、蝕刻、CMP、量測步驟 — 都是 AMAT 最強的領域。AI 資本支出可能維持高檔好幾年。
Mix upgrade, not just a headwind不只是逆風，是品質升級
China fell from 37.2% of revenue in FY24 to 30.1% in FY25 ✓ 10-K. The bull read: mix shifts toward higher-ASP leading-edge tools for TSMC/Samsung/SK Hynix — better margins, steadier demand.
$6.4B of recurring, high-margin service grows with the installed base and smooths the WFE cycle — a layer the market underweights vs. lumpy systems sales.
Leadership in EPI for GAA and CMP for new materials; possible expansion into power semis and process-control software; CHIPS-Act onshoring as a 2026-27 tailwind. R&D $3.57B/yr funds the edge ✓ 10-K.
What the market may be missing:市場可能漏看的：the duration and breadth of AI-complexity intensity.AI 複雜度需求的持續時間與廣度。butsee the headline — much of this is now in the price.但見開頭結論 — 這大半已反映在股價裡。
Bear case — what could go wrong空方論點 — 哪裡可能錯
Valuation is the bear case now估值本身就是空方理由
At ~46× forward / ~52× trailing even after July's -24% pullback (was ~55×/~76× at the 06-30 high), any growth wobble = multiple compression — the July selloff was this card executing. The stock still prices in durable, low-cyclicality growth. live · 07-22
If hyperscalers digest AI capacity, equipment orders can drop sharply — history says so. A "super-cycle" is itself an assumption.
若雲端大廠消化既有 AI 產能，設備訂單可能急跌 — 歷史有前例。「超級循環」本身就是一個假設。
China / geopolitics中國 / 地緣政治
30.1% of FY25 revenue, already down to 27% in FQ2 FY26✓ 10-Q · 05-14. AMAT guided a $600M FY2026 hit from expanded U.S. export curbs ✓ Reuters/TF; domestic substitution (NAURA) and Taiwan tail-risk add more.
Must hold for the thesis to fail:論點要被推翻，需同時成立：sustained high AI capex from a concentrated customer set, on-time node transitions, no new export shocks, continued pricing/mix gains. Break one and the multiple unwinds.集中客戶群的 AI 資本支出持續高檔、節點轉換準時、沒有新的出口管制衝擊、定價/組合持續改善。任一條斷掉，本益比就會回吐。
Variant perception變異觀點— the reason to read— 值得讀的地方
Consensus → variant → what the tape already reflects. The variant thesis was right enough that the market adopted it; the +118% YTD (after a -24% give-back from the 07-01 high) is the receipt — drawdown included.共識 → 變異觀點 → 股價已反映的。變異論點夠對，所以市場已採納；+118% YTD（自 07-01 高點回吐 -24% 之後）就是收據 — 連同回檔一起。
Management & insiders經營層與內部人
Track record經營紀錄
CEO Gary Dickerson✓ public — long tenure, navigated multiple cycles, positioned AMAT for advanced nodes and AI. High-quality capital allocation: heavy R&D ($3.57B), buybacks, growing dividend.
執行長 Gary Dickerson✓ 公開 — 任期長、歷經多次循環、把 AMAT 卡位在先進節點與 AI。資本配置品質高：重研發（$3.57B）、買回庫藏股、配息成長。
Insider signal — net selling內部人訊號 — 淨賣出
Insiders own just 0.3% (~$1.1B) ✓ SWS. CEO Gary Dickerson sold 78,321 shares for ~$55.5M at ~$700–736 on 2026-06-29/30 — the all-time-high days — filed 07-01; 1,599,843 shares held after ✓ SEC Form 4. President Prabu Raja sold 50,000 shares on 2026-06-04 for $25,264,197 ($503.99–507.23, via living trust) ✓ SEC Form 4. Sales at highs are common; the cluster of them AT the top, with zero buying, is a data point the bear column gets to keep.
內部人僅持股 0.3%（約 $1.1B）✓ SWS。執行長 Gary Dickerson 於 2026-06-29/30 — 正是歷史高點那兩天 — 以約 $700–736 賣出 78,321 股、約 $55.5M，07-01 申報，餘 1,599,843 股 ✓ SEC Form 4。總裁 Prabu Raja 於 2026-06-04 賣出 5 萬股、$25,264,197（$503.99–507.23，透過生前信託）✓ SEC Form 4。高檔賣股常見；但成群出現在頂點、且無人買進，是空方欄位可以留著的一個數據點。
Probability-weighted outcomes機率加權情境
judgmentThese percentages are a subjective call, not a statistic. The June note here — that a fresh high argued for higher bear odds — partially played out in July's -24% pullback. The split is kept as originally framed; the live tension has moved to whether FY27's +37% consensus EPS survives the memory-capex cycle (see the 07-23 update).這三個百分比是主觀判斷，不是統計數字。6 月在此加註的「創新高應調高空方機率」，已在 7 月 -24% 回檔中部分兌現。機率分布維持原判；現在的張力移到 FY27 共識 EPS +37% 撐不撐得過記憶體資本支出週期（見 07-23 更新）。
To double from $560 means a ~$890B market cap. A rising stock has two engines: (1) the company earning more, and (2) the market paying a higher P/E (re-rating). Engine 2 has now compressed twice — ~57× at the June build, ~55× at 06-30, ~46× today — while the price round-tripped through $739.67. So a double still rests on engine 1. The concrete math: at an unchanged ~46×, FY27 consensus EPS $16.75 landing in full puts the stock near $770 (+38%) — a double needs either years of compounding beyond FY27, or the re-rating this page has argued is spent. ◆
Added at the 2026-07-23 refresh — this page predates the decision-layer format. Every row is this analysis's judgment ◆, derived from the refreshed numbers above. The page keeps its original no-target-price qualitative frame; a full spine upgrade would be a rebuild.
◆Don't chase the rebound. Even -24% off the high, ~46× FY26E / ~33× FY27E still prices FY27's +37% EPS landing intact — the discount is on sentiment, not on the estimate risk.不追反彈。即使距高點 -24%，~46× FY26E／~33× FY27E 仍把 FY27 EPS +37% 的完整落地計入價內 — 折價折的是情緒，不是估計風險。
If holding已持有
◆Hold with 08-13 (FQ3) as the tripwire: revenue below the $8.45B guide floor, or an FQ4 guide implying QoQ deceleration, → trim. Selling into a -24% hole while customer capex signals strengthen (TSMC $60-64B, Micron ~$27B) is panic, not process.續抱，以 08-13（FQ3）為絆線：營收低於指引下緣 $8.45B，或 FQ4 指引隱含 QoQ 減速 → 減碼。在客戶資本支出訊號轉強（台積電 $60-64B、美光 ~$27B）之際往 -24% 的坑裡砍單，是恐慌不是紀律。
Entry zone進場區
◆< ~$490 (≈40× FY26E $12.26, ≈29× FY27E $16.75) and the 08-13 print confirms the guide with no customer-capex-cut evidence. Price alone or evidence alone is half a trigger. (July's trough: $529.66 on 07-17.)< 約 $490（≈40× FY26E $12.26、≈29× FY27E $16.75）且 08-13 財報確認指引、無客戶砍資本支出證據。價格與證據缺一不可。（7 月低點：07-17 的 $529.66。）
Why not more aggressive為何不更激進
◆Not shorting: WFE demand signals are accelerating (TSMC/Micron/SK hynix capex), FQ3 revisions run 25-up/0-down, and the CEO claims visibility "to 2030" — shorting that into a -24% hole is poor asymmetry. Not adding: FY27's $16.75 already embeds the 2027 memory-capex wave landing intact — the exact assumption the memory cycle could break.不做空：WFE 需求訊號在加速（台積電／美光／SK海力士 capex）、FQ3 修正 25 上 0 下、CEO 稱能見度「到 2030」— 對著 -24% 的坑放空，不對稱性很差。不加碼：FY27 的 $16.75 已把 2027 記憶體資本支出波完整落地當前提 — 而那正是記憶體週期最可能斷掉的假設。
Invalidation失效條件
◆(1) FQ3 (08-13) misses the guide floor or FQ4 guide decelerates → trim / view review. (2) Any of Samsung/SK hynix/Micron/TSMC cuts capex guidance ≥15% → FY27 consensus breaks → bear. (Upside) (3) FQ4 guide >$10B or China curbs ease → the entry zone revises up.(1) FQ3（08-13）低於指引下緣或 FQ4 指引減速 → 減碼／觀點重審。(2) 三星／SK海力士／美光／台積電任一將 capex 指引下修 ≥15% → FY27 共識破功 → 空方情境。（上行）(3) FQ4 指引 >$10B 或中國管制放寬 → 進場區上修。
Catalysts — threshold → reaction催化劑 — 門檻 → 反應
When時間
Event事件
Threshold → reaction門檻 → 反應
Status狀態
2026-08-13
FQ3 FY26 resultsFQ3 FY26 財報
Rev ≥ $8.95B mid & FQ4 guide accelerating → base confirmed; below $8.45B floor → trim per playbook營收 ≥ 中值 $8.95B 且 FQ4 指引續增 → 基準確認；低於下緣 $8.45B → 依框架減碼
scheduled ⚠ date via aggregators已排定 ⚠ 日期為彙整源
2026-10
TSMC Q3 call台積電 Q3 法說
First 2027 capex language ≥ 2026's $60-64B → FY27E supported; "digestion" language → FY27E at risk2027 capex 初步語言 ≥ 2026 的 $60-64B → FY27E 有撐；出現「消化」字眼 → FY27E 有險
