AMAT › analyze
AMAT · Analyze
Date: 2026-07-29 (intraday, ~14:19 ET — market open) · Sector: Semiconductors (Semi Equipment & Materials)
Command: /analyze AMAT · Framework: analysis_notes.md §0–§5 + §4 dividend-grower overlay
Trigger: Watchlist flagged [—] $476.46 · −20% in 14 days, entry stale. Refresh via /analyze — this is that refresh.
Live price: $452.11 (−5.11% today, prev close $476.46). 52-wk range $154.47 – $739.67. −38.9% off the all-time high set 2026-06-30 · but still +151% over 52 weeks and ~23% ABOVE its 200-day MA.
Data notes. Yahoo Finance MCP connected ✅. roic.ai connected but on the free plan — returned only FY24/FY25, blocking the framework's 5–8yr depth; deep history was reconstructed from Yahoo + stockanalysis + Macrotrends and FY2016–2020 is single-sourced (flagged inline). Public.com used for live quotes (read-only). Primary sources (Q2 FY26 press release + 10-Q) fetched directly and are the authority for all earnings-quality figures below. Two data traps caught and corrected: (1) roic.ai's FCF field for AMAT is wrong — it returns OCF because
cf_cap_expendituresis null, overstating FY25 FCF by $2.26B (40%). (2) Yahoo misclassifies strategic-investment gains as "Interest Income" — Q2 FY26 shows $771M, which on ~$8.2B of cash would imply a 38% annualized yield. Neither figure is used here.
1. The one-paragraph answer
Applied Materials is a genuinely good business having a genuinely good year, whose reported earnings are flattered, whose returns on capital are compressing, whose competitive share is quietly eroding, and whose stock — even after a 39% crash — still offers no margin of safety. The −39% drawdown is ~100% multiple de-rating and ~0% earnings revision: forward estimates rose throughout the fall, and not one company in the supply chain has announced a spending cut. That makes this a real opportunity in the making. But it is not yet an opportunity at $452, because the price is still roughly 1.3–1.5× our fair-value range, the "record" TTM EPS is ~26% non-operating, and — the detail that matters most — the CEO sold $105M of stock at $590–736 in the two weeks straddling the top, while every analyst on the Street was raising targets into the fall. The framework's §0 test returns "great-ish company, still expensive." Verdict: WATCH, do not buy. Accumulate $300–350.
2. Snapshot
| Metric | Value | Metric | Value |
|---|---|---|---|
| Price | $452.11 | Market cap | $361.2B |
| Enterprise value | $360.3B | Shares (diluted) | 799M |
| P/E (TTM GAAP) | 42.5x | P/E on normalized EPS | 57.7x |
| P/E on FY26 non-GAAP est | 37.4x | P/E fwd (consensus FY27) | 27.0x |
| P/FCF (TTM) | 67.6x | FCF yield | 1.48% |
| P/B | 15.0x | P/S | 12.4x |
| Gross margin (Q2 FY26) | 49.9% (best in 25+ yrs) | Operating margin (TTM) | 28.6% |
| Pretax ROIC (TTM, adj.) | ~33% (from 44%) | ROE | 40% |
| Net debt | −$973M (net CASH) | Debt / Assets | 18.0% |
| Dividend yield | 0.47% | 10yr dividend CAGR | +18.2% |
| Beta | 1.57 | Short float | 2.1% (falling) |
| Consensus | 35 analysts, 0 Sell, mean PT $627.66 | Next catalyst | Q3 earnings Aug 13 |
3. 🚩 The finding that reframes everything: the "record" earnings are not what they look like
This is the most important section of the report, verified against the Q2 FY26 10-Q and press release, not a data vendor.
AMAT reported record GAAP EPS of $3.51 in Q2 FY26 (+33% YoY). Non-GAAP EPS was $2.86 (+20%). The reconciliation line naming the gap reads:
A portfolio-specific passage was removed from the public build.
A portfolio-specific passage was removed from the public build.
What the securities are: AMAT's 9% stake in BE Semiconductor (BESI), bought ~April 2025, making it BESI's largest shareholder. It is a hybrid-bonding/advanced-packaging pure play — i.e. a semiconductor equipment stock. AMAT's reported earnings are therefore partly a leveraged bet on its own sector's share price.
The TTM bridge — where the earnings growth actually came from
| Metric | FY2025 | TTM (thru Apr-26) | Change |
|---|---|---|---|
| Revenue | $28,368M | $29,024M | +2.3% |
| Operating income | $8,289M | $8,299M | +0.1% — FLAT |
| Non-operating income (net interest + investment gains) | $982M | $2,084M | +112% |
| Effective tax rate | 24.5% | 18.1% | −6.4 pts |
| Net income | $6,998M | $8,508M | +21.6% |
| Diluted EPS | $8.66 | $10.65 | +23.0% |
Operating income is dead flat year-over-year. 100% of the TTM earnings growth came from below the operating line. Normalizing — taxing the $8,299M of operating income at FY25's 24.5% rate — gives normalized EPS ≈ $7.84. So roughly $2.80 of the $10.65 TTM EPS (≈26%) is non-operating gains plus a tax benefit.
Corroboration
- H1 FY26 GAAP EPS $6.05 vs non-GAAP $5.24 — GAAP is running 15.5% hot.
- Q1 FY26's headline "+75% EPS growth" is an easy-comp artifact; Q1 FY26 non-GAAP EPS was explicitly flat YoY.
- H1 FY26: net income +45%, operating cash flow +1.4%, free cash flow −22%.
Per §0 — "FCF is harder to fake than earnings" — believe the FCF line.
⭐ The Aug 13 setup this creates — actionable
BESI's ADR has fallen from ~$296 at AMAT's fiscal Q2 close (Apr 26) → ~$258 at fiscal Q3 close (Jul 26) → $216.78 today (−42% off its high). Marks reverse:
| AMAT fiscal quarter | BESI mark | Estimated GAAP impact |
|---|---|---|
| Q2 FY26 (closed Apr 26) | ~$296 | +$0.80/sh gain (reported) |
| Q3 FY26 (closed Jul 26) | ~$258 | ≈ −$0.30 to −$0.40/sh loss |
| Q4 FY26 (closes late Oct) | $216.78 and falling | materially larger loss if it stays here |
So: Q3 non-GAAP EPS is guided to $3.36, but GAAP EPS may print ~$3.00–3.10 — a sequential GAAP decline despite revenue rising from $7.91B to ~$8.95B. A headline reading "AMAT GAAP EPS falls" would be pure mark-to-market noise on a quarter where the operating business grew ~23%.
Watch for this on August 13. If the market sells a GAAP "miss" that is really a BESI mark, that is a better entry than today's price — and it is a mispricing you can see coming. Judge the print on non-GAAP EPS vs $3.36, revenue vs $8.95B, the CY26 >30% guide, and FCF conversion — not on GAAP EPS.
⚠️ Estimate, not a forecast: derived from BESI's ADR price and a 9% stake against BESI's ~79.2M share count. AMAT holds other strategic investments (a $5,142M long-term book) whose composition is not fully disclosed, so the true mark will differ.
4. Fundamentals — §1 General Health
4.1 Free cash flow — peaked three years ago
| FY | OCF ($M) | Capex ($M) | FCF ($M) | FCF margin | FCF/share |
|---|---|---|---|---|---|
| 2016–2020 | — ⚠️ | — ⚠️ | 2,313 → 3,382 ⚠️ | — | — |
| 2021 | 5,442 | −668 | 4,774 | 20.7% | $5.19 |
| 2022 | 5,399 | −787 | 4,612 | 17.9% | $5.26 |
| 2023 | 8,700 | −1,106 | 7,594 ← peak | 28.6% | $8.99 ← peak |
| 2024 | 8,677 | −1,190 | 7,487 | 27.5% | $8.98 |
| 2025 | 7,958 | −2,260 | 5,698 | 20.1% | $7.05 |
| TTM | 7,993 | −2,650 | 5,343 | 18.4% | $6.69 |
TTM independently verified two ways: quarterly sum (2,050+2,043+1,040+210) = 5,343, matching stockanalysis exactly.
| FCF CAGR | Period | Rate |
|---|---|---|
| 5yr | FY20→FY25 | +11.0% |
| 8yr | FY17→FY25 | +6.5% ⚠️ |
| Peak→now | FY23→TTM | −29.6% 🚩 |
🚩 FCF peaked in FY2023 and has fallen 30% since, while revenue rose every single year. The 8yr FCF CAGR of 6.5% lags the 8.6% revenue CAGR — conversion is deteriorating. Two structural causes: capex more than doubled ($1.1B → $2.65B, the EPIC R&D center + capacity build, with no stated normalization date), and working-capital absorption during the ramp.
✅ One flag I investigated and am dropping: the $1.4B quarterly receivables build looked like channel-stuffing, but DSO actually improved year-over-year (73 days in Q2 FY26 vs 78 in Q2 FY25). The jump is a rebound off an abnormally low January quarter. Timing, not quality. Inventory at 146 days is high but normal for AMAT's lead times.
4.2 Capital allocation — historically excellent, currently signalling
| FY | OCF | Capex (%OCF) | Buybacks (%OCF) | Dividends (%OCF) | M&A | Total (%OCF) |
|---|---|---|---|---|---|---|
| 2022 | 5,399 | 14.6% | 118% | 16.2% | 441 | 148% |
| 2023 | 8,700 | 12.7% | 27.2% | 11.2% | 25 | 51% |
| 2024 | 8,677 | 13.7% | 47.4% | 13.7% | 0 | 75% |
| 2025 | 7,958 | 28.4% | 64.6% | 17.4% | 29 | 110% |
| TTM | 7,993 | 33.2% | 38.0% | 18.3% | 175 | 89% |
- ✅ FY2022 was a masterclass — spent 118% of OCF on buybacks at the cycle low.
- ✅ Negligible M&A (~$670M over 5 years). No empire-building. Clean.
- ✅ Buyback execution is genuinely outstanding: FY21–25 deployed $21.9B and retired 130M net shares — an effective ~$169/share (est. gross avg ~$125 adjusting for SBC issuance) now worth $452. Real, large value creation.
- 🚩 FY2025 over-distributed at 110% of OCF — funded from cash and net new debt, while the stock was near record highs. Poor timing vs FY2022.
- ⭐ Buybacks have been cut ~70%. Q1+Q2 FY26 totalled $737M, annualizing to ~$1.5B vs $5.1B in FY25. Management is not buying at $452. From the same team that bought aggressively at the FY22 lows, that is a credible valuation signal — and it corroborates the insider selling in §6.
4.3 Balance sheet — the unambiguous strength
| Metric | FY2023 | FY2024 | FY2025 | Q2 FY26 |
|---|---|---|---|---|
| Debt / Assets | 19.5% | 19.2% | 19.4% | 18.0% ✅ |
| Debt / Equity | 36.7% | 34.8% | 34.5% | 30.4% ✅ |
| Net debt | −130 | −1,410 | −1,523 | −973 (net CASH) ✅ |
| Current ratio | — | — | 2.61 | 2.51 ✅ |
| Interest coverage | — | — | ~30x | ~30x ✅ |
No balance-sheet risk whatsoever. Net cash, 18% debt/assets, ~30x coverage. Interest income comfortably exceeds interest expense. This is a fortress and it is what makes AMAT genuinely ownable at the right price.
4.4 Shares outstanding
1,084M (FY17) → 794M (Q2 FY26) = −26.8% over 8 years (−3.8%/yr) ✅ — a consistent ~3.5–4pt/yr tailwind to every per-share metric. ⚠️ But the reduction stalled in FY26 (793.0 → 794.0, slightly up) as buybacks were cut and SBC issuance continued. The per-share tailwind has paused.
4.5 Growth and per-share — the owner's-eye view
| Metric | 5yr CAGR | 8yr CAGR |
|---|---|---|
| Revenue | +10.5% | +8.6% |
| Net income | +14.1% | +9.0% |
| FCF | +11.0% | +6.5% ⚠️ |
| Revenue/share | +13.9% | +12.6% ✅ |
| FCF/share | +14.0% | +10.5% |
| Dividend/share | +14.6% | +20.0% ✅ |
✅ Revenue/share compounding 12.6% vs revenue's 8.6% — buybacks adding a genuine ~4pts/yr. 🚩 But FCF/share peaked at $8.99 (FY23) and is now $6.69 — three consecutive down years in the metric that best represents what an owner actually receives. Buybacks can no longer mask the FCF erosion.
4.6 Margins — expanding, but check why
| FY | Gross | Operating | Net |
|---|---|---|---|
| 2021 | 47.3% | 29.9% | 25.5% |
| 2022 | 46.5% | 30.2% | 25.3% |
| 2023 | 46.7% | 28.9% | 25.9% |
| 2024 | 47.5% | 28.9% | 26.4% |
| 2025 | 48.7% | 29.2% | 24.7% |
| TTM | 49.0% | 28.6% | 29.3% ← inflated |
| Q2 FY26 | 49.9% | 31.9% | 35.5% ← inflated |
⚠️ FY2016–2020 gross/operating margins unavailable (roic.ai free-tier block) — a genuine gap in the 8yr trend.
- ✅ Gross margin genuinely expanding: 46.5% → 49.0%, hitting 49.9% in Q2 FY26 (best in 25+ years).
- ⚠️ But operating margin is flat-to-down (30.2% → 28.6%) — R&D ($3.57B, 12.6% of revenue) and SG&A are growing faster than revenue. Gross-margin gains are being spent, not banked.
- 🚩 Ignore TTM/Q2 net margins — artifacts of investment gains and the 18% tax rate.
⭐ Is the margin expansion pricing power or mix? The Moat analyst flagged this as the highest-value cross-check, and the Q2 10-Q segment data settles it: Semiconductor Systems operating margin 35.1% vs AGS 29.2%. So the gain is not a services-mix shift — it is a leading-edge mix effect (Taiwan 27% + Korea 20% of revenue) during an AI capex boom, amplified by exiting low-margin China mature-node business. That is mix, not pricing power — and a chunk of it reverses when the cycle mix normalizes. It materially weakens the "structurally better business, so pay a higher multiple" argument.
4.7 Health scorecard
| # | §1 Dimension | Grade | Rationale |
|---|---|---|---|
| 1 | Free cash flow | C+ | 8yr CAGR 6.5% lags revenue; peaked FY23, −30% since; margin 28.6%→18.4%. Absolute generation still substantial. |
| 2 | Capital allocation | A− | Superb long-run record; buybacks at ~$125–169 now worth $452; near-zero M&A. Marked down for FY25's 110%-of-OCF payout at highs. |
| 3 | Debt / balance sheet | A | Net cash, 18% D/A, ~30x coverage. Non-issue. |
| 4 | Shares outstanding | A− | −26.8% over 8yr, consistent. Marked down because it stalled in FY26. |
| 5 | Top & bottom line | B− | 8yr rev +8.6%, NI +9.0% — profitable, undiluted. But TTM operating income is flat and reported growth is entirely non-operating. |
| 6 | Per-share metrics | B | Revenue/share +12.6% ✅. But FCF/share −26% from peak, three straight down years. |
| OVERALL | B / B− | A fortress-balance-sheet, well-run, high-ROIC business whose cash economics are quietly deteriorating while reported earnings look their best in a decade. |
4.8 Dividend overlay (§4) — nascent grower, not an income name
| Metric | Value |
|---|---|
| Quarterly dividend | $0.53 (raised from $0.46 in May 2026) |
| Annualized | $2.12 → yield 0.47% (5yr avg 0.79%) |
| 10yr CAGR | +18.2% ✅ |
| 8yr CAGR | +20.0% ✅ |
| Earnings payout | 17% |
| FCF payout | 17.6% (FY21) → 12.8% (FY23) → 27.4% (TTM) |
| Consecutive raises | 9 years |
✅ The dividend is very safe — 27% of FCF, 17% of earnings, net-cash balance sheet, 9 straight raises at ~20%/yr. ⚠️ But coverage deteriorated from 12.8% → 27.4% in two years — and note the cause: the dividend didn't get reckless, FCF fell. The denominator shrank 30%. ⚠️ Historical caution: AMAT froze its dividend at $0.10/quarter for four years (2014–2018) through the last real downturn. It didn't cut — but it froze. That is the cyclicality tell. Verdict: correctly a nascent grower for the 📈 Income–Yield Tomorrow thesis, not an income holding. At 0.47% the yield contributes essentially nothing to total return today. DYT actively reads "expensive" — yield is near its 5yr low — a weak but confirming negative signal.
4.9 Cyclicality — where are we?
| Cycle event | Move |
|---|---|
| FY2018 → FY2019 (memory downturn) | −12.6% |
| FY2019 trough → FY2025 | +94%, six consecutive up years |
| FY2023 FCF → TTM FCF | −29.6% |
The critical observation: revenue and FCF have decoupled. Revenue hasn't had a down year since FY2019 — an unusually long uninterrupted expansion for a WFE company — but FCF already peaked three years ago. The cash cycle turned in FY2023 while the revenue cycle kept climbing.
Classic late-cycle markers now present: capex doubling into the upturn; inventory at 146 days; FY25 payout of 110% of OCF; guidance accelerating into a demand surge.
Read: mid-to-late expansion, accelerating. Q3 FY26 guides +22.6% YoY and management guides semi-equipment +30% in CY2026. The cycle is a tailwind to the next 4–6 quarters of the income statement and a risk to the multiple thereafter.
4.10 China / export controls
| Period | China % of revenue |
|---|---|
| CY2024 peak | ~45% |
| FY2025 | 28% |
| Q2 FY2026 | 27% ($2,087M) |
Q2 FY26 mix: China 27% · Taiwan 27% · Korea 20% · US 12% · Japan 8% · Europe 4%. 89% of revenue from outside the US.
- ✅ The legal overhang is fully resolved. On 2026-02-11 AMAT settled with BIS for $252.5M (ion implanters routed via Korea to SMIC, 2021–22 — second-largest BIS penalty ever), and the DOJ and SEC both closed their inquiries with no enforcement action. Multi-year criminal exposure is gone. This was the −$265M "unusual item" in Q1 FY26.
- ✅ China concentration nearly halved (45% → 27%) and the business still grew revenue every year — Taiwan and Korea absorbed it. A genuine demonstration of demand durability.
- 🚩 27% remains the largest single-country exposure, and it is subject to political rather than economic risk. AMAT is now barred from China's memory and mature-node markets, with a guided ~$600–710M FY26 revenue hit. The MATCH Act (H.R. 8170) would extend the ban to DUV/advanced etch — still in committee, a live binary tail risk.
5. Moat — §2
5.1 Rating: WIDE — but the weakest Wide among the Big Five, trending toward Narrow
The structural fact that matters most: AMAT is a "broad portfolio, few monopolies" company.
| Segment | AMAT position | Main rival |
|---|---|---|
| PVD | #1, dominant — a real chokepoint | Ulvac, Naura (low end) |
| Epitaxy | #1, dominant — critical for GAA channel | ASM International |
| CVD | #1 | Lam, TEL, Kokusai |
| CMP | #1 / co-#1 | Ebara (arguably co-leader) |
| ALD | Contested #2/#3 — weak spot | ASM International (leader) |
| Etch | Distant #2 | Lam >50% of dry etch |
| Ion implant | #1 but eroded: >75% (2011) → <60% (2017) | Axcelis (5% → ~30%) |
| Process control | Distant #2 — 9.8% (2025), DOWN from 11.1% | KLA 73.8% |
| Lithography / track | Does not compete | ASML monopoly; TEL+SCREEN duopoly |
⚠️ Third-party WFE share estimates for AMAT are mutually inconsistent (19%, 13.2%, 8.4% all appear in sources). The defensible statement is ~15–20% of WFE, #2 behind ASML. Segment rankings above are directionally sourced.
The load-bearing bear fact: per Castellano (2026), AMAT and TEL both lost share within their served markets in 2025, while Lam, KLA and ASML all gained — in a market that grew 13%. (Underlying tables paywalled; direction verified, magnitudes not.) Wide moats do not lose relative ground in boom years.
And the precedent is not hypothetical: ion implant went from >75% share to <60% over six years while Axcelis went 5% → ~30%. Recipe lock-in did not prevent it. That is a completed erosion, in public, while the company kept reporting good numbers.
5.2 The real moat is AGS — and it's a lagging indicator
| AGS metric | Value |
|---|---|
| FY25 revenue | ~$6.5B (23% of total) |
| Operating margin | 28.1% |
| Backlog | $7.1B — equal to ALL of Semiconductor Systems |
| Parts/service under subscription | 63%, >90% renewal, 2.8yr avg term |
| Chamber installed base | >2x nearest competitor |
| Q2 FY26 growth | +17% YoY ($1.67B) |
An installed chamber is a 15–20 year annuity; the fab cannot buy qualified spares or service elsewhere without voiding process qualification. This is the one place AMAT's broad portfolio converts into durable structural advantage — breadth that dilutes it in any single product market becomes strength in aggregate service scale.
⭐ But the sharpest point in the whole analysis: AGS is a derivative of historical systems share. Every tool AMAT fails to sell today is 15 years of AGS revenue it doesn't book starting ~2029. AGS's current strength is evidence about AMAT's competitive position five years ago, not today — so it cannot be used to refute today's share-loss evidence. Bulls citing AGS growth as proof the moat is intact have the causality backwards.
⚠️ Comparability broken: the 200mm business moved out of AGS into Semiconductor Systems in Q1 FY26 (making AGS fully recurring), and Display ceased to be a reportable segment as of Oct 2025. FY26 AGS growth is inflated by mix cleanup.
5.3 Adversarial stress-test (§2.2) — four attacks
| Attack | Works? | Severity | Timeline |
|---|---|---|---|
| Chinese state-backed substitution | YES — already working | HIGH | In progress, 5–10yr |
| Lam concentrated attack at node inflections | Partially | Moderate | Ongoing, per-node |
| TEL bundling off the track duopoly | Mostly no | Low–Moderate | — |
| Customer buyer power (34% of revenue = 2 customers) | Yes, as a permanent margin ceiling | Moderate, chronic | Permanent |
Attack 1 — the one that works. "I don't beat AMAT on technology. I get Beijing to mandate domestic content, take the China market by decree, start at mature nodes, and use guaranteed subsidized volume to fund the climb up the node ladder. My cost of capital is zero. I don't have to win — I just have to exist until AMAT is gone."
| Metric | 2021 | 2024 | 2025 |
|---|---|---|---|
| Chinese vendors' share of global WFE | 1.2% | 5.6% | 6.5% |
| Domestic share within China | — | 25% | 35% (target was 30%) |
| Etch & deposition substitution in China | — | — | >40% |
Naura is now #5 globally (from #8 in 2022, passing KLA) at ~$6.7–7.5B revenue; three Chinese toolmakers are in the global top 20; Hwatsing shipped its 1,000th CMP tool and is expanding into ion implant. Q1 2026 growth: Piotech +57%, ACMR +33%, Naura +26%.
The asymmetry that indicts AMAT specifically: the segments where China is advancing fastest (etch, deposition, CMP, cleaning, ion implant) are exactly AMAT's. The segments where China is stuck (litho, high-end process control) are ASML's and KLA's.
AMAT's rebuttal is fair — lost China revenue was lower-margin mature-node business replaced by higher-margin leading-edge demand, and Chinese vendors are still only 6.5% of global WFE. The counter-rebuttal is stronger: 6.5% was 1.2% four years ago — the bull extrapolates from a level, the bear from a rate. And export controls destroyed the natural check: AMAT can no longer compete on merit in China, so Washington has guaranteed the domestics a captive market with no Western competition in it.
Crucially, this threat is unmanageable. Every other risk has a management response. This one has none — AMAT's own 10-K concedes the regulations "provide an advantage to our international competitors." Unmanageable risks deserve heavier weighting than manageable ones.
5.4 Disruption — the GAA bull case, evaluated hard
The bull argument is real and it is the strongest genuine positive in the file. Litho intensity is plateauing while GAA and backside power delivery are deposition-, etch- and materials-intensive — shifting WFE dollars from ASML's pocket toward AMAT's. GAA nanosheet channel formation is an epitaxy problem and AMAT owns epi; backside power needs wafer thinning, deep vias and new metallization. Management claims >50% of incremental GAA spend and ~50% of backside power spend (company claim, unverified). Leading-edge logic + DRAM + advanced packaging = >80% of WFE growth in 2026–27.
Three problems the bull case skips: 1. A rising tide that lifts Lam more. Lam's GAA + advanced-packaging shipments went >$1B (2024) → >$3B (2025). If AMAT's served market grows 20% and Lam's grows 40%, AMAT's absolute revenue rises while its relative position weakens — which is exactly what the 2025 share data shows. Rising revenue is being used as evidence of a rising moat. It is not the same thing. 2. New materials favor the challenger. Every new material is a fresh bake-off with no incumbent recipe to defend — Lam already won ALD molybdenum (sole production supplier in foundry/logic) for sub-2nm. The very inflection that expands AMAT's market is the mechanism by which AMAT loses share. Same story, not two stories. 3. Concentration into more powerful buyers. Only TSMC, Samsung and Intel run leading-edge GAA. A bigger leading-edge market means more dependence on the two customers already at 34% of revenue.
Verdict: genuine tailwind to revenue and served market — overstated as moat evidence. It expands the pie without strengthening AMAT's claim on its slice.
Advanced packaging: $1.7B → >$2B in 2026 (+>50%). Real, but AMAT leads the process steps and does not own the bonder — the actual chokepoint. BESI does. Read the 9% BESI stake correctly: it is a defensive option to stop Lam getting the chokepoint, not an offensive position. As of March 2026 BESI is fielding takeover interest from both Lam and AMAT. ⚠️ If Lam acquires BESI, AMAT is locked out of the highest-value step in the fastest-growing part of its market. Binary, live, monitor.
AI concentration: the chain is short and brittle — hyperscaler capex → NVDA/AVGO → TSMC/Samsung/Hynix → AMAT. The ICAPS hedge that was meant to smooth this is being legislated away by export controls, leaving AMAT more AI-concentrated than it chose to be.
5.5 Evergreen verdict: evergreen industry, evergreen-adjacent company
| Claim | True? |
|---|---|
| "Semiconductors are forever" | TRUE, high confidence |
| "Wafer fab equipment is forever" | TRUE — complexity raises equipment intensity per wafer |
| "AMAT specifically keeps its share and margins" | UNCERTAIN — and it's the only question that matters |
High confidence AMAT exists in 20 years and is profitable. Not confident it holds ~19% WFE share at 49% gross margins. Base case: slow, grinding share erosion of 20–50bps/yr, masked for a decade by a growing pie and by AGS's lagging annuity. Perfectly investable — but hold with monitoring, not hold and forget. The ion implant history is the template.
Falsifiable downgrade trigger: if the 2025 served-market share loss repeats in 2026 and 2027, move this rating to Narrow.
5.6 Moat defense — what management is doing
| Weakness | Response | Rating |
|---|---|---|
| Share loss at node inflections | EPIC Center (~$5B, >$25B R&D/10yr; Samsung, TSMC, Broadcom committed; opens spring 2026, unveiling Oct 12) | HIGH reward / MODERATE risk — best-conceived initiative. Converts a per-node bake-off into continuous co-development, the only real answer to the share-reset problem. |
| AGS erosion risk | Subscription conversion (63%, >90% renewal); 200mm moved out to make AGS fully recurring | HIGH reward / LOW risk — best-executed initiative. |
| Not owning the packaging chokepoint | 9% BESI stake; ASMPT NEXX acquisition | HIGH reward / HIGH risk — BINARY. 9% is influence, not control. |
| China revenue loss | ICAPS strategy | LOW reward. Largely obsolete — export controls barred AMAT from the exact markets ICAPS was meant to defend. Not management's fault; also not a working defense. |
| Process control (9.8% vs KLA 73.8%) | Integrated eBeam metrology; AIx (>35,000 connected chambers) | LOW / LOW — effectively conceded. Losing to KLA even in eBeam. Means AMAT is structurally absent from the industry's best business. |
Management has correctly diagnosed every weakness and is spending real money against each. But the two largest threats — Chinese state substitution and buyer concentration — are not addressable by management action. Don't credit effort against threats immune to effort.
6. Sentiment — §3 of the brief
6.1 What actually caused the −39%
~100% multiple de-rating, ~0% earnings revision.
| At the high $739.67 | Today $452.11 | Change | |
|---|---|---|---|
| Forward EPS (unchanged) | $16.90 | $16.90 | 0% |
| Forward P/E | 43.8x | 26.7x | −39% |
Forward estimates did not fall during the drawdown — they rose. The entire move is the market paying less for the same (rising) earnings stream.
6.2 ⭐ Analysts vs insiders — the sharpest signal in the file
Analysts: 24 consecutive positive actions in 12 months. Zero downgrades. Targets roughly doubled in six weeks chasing price, then kept rising into the crash:
| Date | Firm | PT: old → new |
|---|---|---|
| 2026-06-30 (the exact top) | Susquehanna | $575 → $900 |
| 2026-06-29 | Cantor Fitzgerald | $650 → $850 |
| 2026-06-26 | B. Riley / Jefferies | $550 → $790 / $510 → $770 |
| 2026-07-06 (mid-crash) | Morgan Stanley (Equal-Weight) | $502 → $647 |
| 2026-07-10 | Needham | $530 → $740 |
| 2026-07-15 (mid-crash) | UBS | $570 → $705 |
Consensus: 35 analysts, 4 Strong Buy / 28 Buy / 7 Hold, 0 Sell, mean PT $627.66.
Insiders over the same window: ~$157M sold, $0 bought.
| Date | Insider | Shares | Price | Value |
|---|---|---|---|---|
| 2026-06-30 | Dickerson (CEO) | 78,321 | $700–736 | $55.6M |
| 2026-06-18 | Raja (Pres. Semi Products) | 10,000 | $633.53 | $6.3M |
| 2026-06-16 | Dickerson (CEO) | 83,000 | $590–599 | $49.2M |
| 2026-06-16 | Nalamasu (CTO) | 35,000 | $590–597 | $20.8M |
| 2026-06-16 | Iannotti (Director) | 9,250 | $599.77 | $5.5M |
| 2026-06-15 | Deane (SVP AGS) | 8,621 | $590.76 | $5.1M |
| 2026-06-04 | Raja | 50,000 | $505–507 | $25.3M |
| 2026-06-03 | Hill (CFO) | 2,500 | $498.86 | $1.2M |
6-month net: −246,272 shares (−9.2% of insider holdings).
Why this is a higher-quality signal than routine diversification: on 2026-04-03 Dickerson bought 50,000 shares at $137.29–137.70 (~$6.9M) on the open market, before a ~5x move. He bought the bottom and sold the top. That is a demonstrated timing record.
Insiders sold precisely what analysts were upgrading.
⚖️ Fair counterweights: the June sales cluster around RSU vesting from the 2025-12-11 grants, so mechanical diversification explains some of it. Short interest FELL from 21.9M to 16.7M shares (only 2.1% of float) — shorts are covering, not building. Institutional ownership 84.9%, sticky index money, no wholesale exit.
6.3 The Burry short — real, but weak evidence
Disclosed 2026-06-30 on his paid Substack ("Cassandra Unchained"), entry $729.40. Basket: NVDA, AMAT, SOXX, Tesla, Caterpillar; Micron added ~July 3. AMAT and NVDA are short stock, not puts.
Four reasons to discount it heavily: 1. There is no filing. Burry deregistered Scion as an SEC investment adviser effective 2025-11-10. No 13F exists to verify the position, its size, or whether he still holds it. All coverage is syndicated repackaging of one monetized post. 2. No size disclosed. Conviction is unquantified. 3. It is not the accounting thesis. His GPU-depreciation argument (~$176B understatement 2026–28) was November 2025 and aimed at NVDA/PLTR — not AMAT. The AMAT leg is a pure valuation/blow-off-top technical call: "The proximate cause of today's rally is big spending announced out of Korea. Well, I see that as the beginning of the end." Note this is the inverse of how that spending actually flows to AMAT's revenue. 4. Chronically early. His NVDA puts were still underwater in early July 2026.
Assessment: a narrative accelerant, not evidence. Well-timed articulation of a repricing already underway. Do not treat as an independent fundamental red flag.
6.4 The WFE cycle — every peer guided UP. There is no capex cut.
| Company | Date | Action |
|---|---|---|
| ASML | 07-15 | RAISED 2026 guide to €43–45B; "close to receiving all orders needed for 2027"; low-NA EUV capacity +30% in 2027 |
| KLA | 07-28 | Record $3.66B, beat high end, RAISED guide to $4.0B; momentum "accelerating through 2027" |
| Lam | 04-22 | RAISED WFE outlook to $140B "with a bias to the upside" |
| Tokyo Electron | ~07 | RAISED FY net income +¥44B; WFE record in 2026 |
| TSMC | 07 | RAISED 2026 capex to $60–64B (from $52–56B) |
| Micron | 03-19 | RAISED FY26 capex to >$25B; ~$40B run-rate by FY27 |
| Samsung | 03-19 | 2026 chip capex/R&D ~₩110T (+22%) |
SEMI: WFE +9.0% in 2026, +7.3% in 2027 to $135.2B. No credible forecaster projects a 2027 WFE contraction.
AMAT is supply-constrained, not demand-constrained. CFO Brice Hill: "the demand outlook has strengthened across almost every leading indicator… the clearest and longest visibility we've ever had… we expect 2027 will be another strong record year." Tracking 100+ fab projects, working from an 8-quarter customer forecast. When UBS pushed that the math implies >40% not >30%, CEO Dickerson: "It's really supply chain… our operations can scale significantly beyond where we're at right now."
⚠️ Correcting two errors circulating in our own files: (1) there was no Morgan Stanley downgrade — MS maintained Equal-Weight and raised its PT to $647; (2) SK Hynix did not cut capex — it reallocated HBM4 lines to DDR5 (near-90% margins) inside a rising capex envelope. For AMAT, DDR5 capacity is still DRAM WFE — roughly revenue-neutral.
Two caveats not glossed over: ASML stopped disclosing quarterly net bookings in Q1 2026, removing the cleanest forward indicator exactly when scrutiny peaked. And mature-node/ICAPS is digesting — Hill: "won't grow a lot year-over-year until we digest the capacity."
6.5 Macro — a genuine hawkish shock doing real work
New Fed Chair Kevin Warsh stripped easing-bias language at his June debut and said "prices are too high" on July 1, spiking yields. 2yr 4.37%, 10yr 4.71% (highest since Jan 2025), PCE ~4.1%, 9 of 18 FOMC participants projecting a 2026 hike (~72% odds). A rising-rate regime is a direct multiple-compression mechanism for long-duration assets — AMAT at 43.8x forward was exactly that. This alone justifies a large share of the de-rate independent of anything semi-specific.
AI-capex digestion is the live narrative: Meta raised FY26 capex to $125–145B and fell 9.25% on it; Alphabet's quarterly FCF reportedly turned negative for the first time. Note the irony of the "sell semis, buy hyperscalers" rotation: it assumes hyperscaler capex keeps rising — bullish AMAT's revenue, bearish only its multiple.
The tell: on 07-28 KLA beat and raised — and the stock fell 8.1% anyway. When good news stops working, you are in a de-rating, not a fundamental repricing.
6.6 Sentiment ratings
- Market sentiment: VERY BEARISH (capitulatory, momentum-driven, indiscriminate)
- Evidence-based assessment: NEUTRAL, leaning mildly bullish on fundamentals / bearish on positioning
Not more bullish because: the CEO — a demonstrably skilled timer — sold $105M at the top and nobody bought the dip; Chinese substitution is real and accelerating in AMAT's core categories; and the stock is still +151% over 52 weeks and 23% above its 200-DMA. This is a de-rating from euphoria, not a distressed valuation.
7. Valuation — §3
Applied conditionally by type. AMAT is a cyclical, asset-light, high-ROIC name with a token 0.47% dividend → Graham gets low weight, DDM/DYT are effectively N/A, and normalized-earnings multiples + Bogle + P/FCF-vs-own-history carry the weight.
The valuation must not be run off reported EPS. Per §3, three earnings bases: TTM GAAP $10.65 (26% non-operating) · Normalized $7.84 (operating-only) · FY26 non-GAAP est $12.09 (Q1 ~$2.32 + Q2 $2.86 + Q3 guide $3.36 + Q4e ~$3.55).
7.1 Where the multiple sits vs AMAT's OWN history
| Value | |
|---|---|
| AMAT historical P/E band (roic.ai) | 14.2x – 29.5x, averaging ~20–22x |
| FY2025 actual P/E | 26.8x (at $233.10); FY25 average 20.1x |
| FY2025 P/FCF | 23.7x; FY25 average 17.8x |
| Today | 42.5x TTM GAAP · 57.7x normalized · 37.4x FY26 non-GAAP · 67.6x FCF |
⭐ Even after a 39% crash, AMAT trades at roughly twice its own historical average earnings multiple, and ~3x its average P/FCF. The drawdown took the stock from absurd to merely expensive. Context: from FY25 year-end ($233.10) the stock is +94% while TTM EPS rose 23% — the multiple did most of the work.
7.2 Multiple grids
On FY26 non-GAAP EPS $12.09 (the most defensible forward base)
| 18x | 20x | 22x | 24x | 26x | 28x | 30x |
|---|---|---|---|---|---|---|
| $218 | $242 | $266 | $290 | $314 | $339 | $363 |
On normalized EPS $7.84
| 18x | 20x | 22x | 24x | 26x | 28x | 30x |
|---|---|---|---|---|---|---|
| $141 | $157 | $172 | $188 | $204 | $220 | $235 |
On FCF/share, vs AMAT's own 17.8–23.7x P/FCF history
| FCF base | FCF/sh | Today's multiple | At 18x | At 24x |
|---|---|---|---|---|
| TTM actual $5,343M | $6.69 | 67.6x | $120 | $160 |
| Normalized (capex $2.0B) $5,993M | $7.50 | 60.3x | $135 | $180 |
| Optimistic FY27 $8,000M | $10.01 | 45.2x | $180 | $240 |
7.3 Reverse DCF — the honest result, and its limits
Discount rate 11.5% (CAPM: 4.71% 10yr + 1.57 beta × 4.5% ERP), 10yr fade to 3% terminal.
Off a normalized $7.0B FCF base, a 10yr FCF CAGR of 8%/12%/16% yields $131 / $152 / $178 per share. Solving in reverse: at $452 the market requires a starting FCF base of ~$15–21B — roughly 2.8–3.9× current TTM FCF — before any growth.
⚠️ I am deliberately down-weighting this rather than leading with it. A 11.5% discount with 3% terminal growth implies an 11.8x terminal FCF multiple, and AMAT has never traded there — its realized history is 17.8–23.7x. A DCF this conservative will always say a quality compounder is 3x overvalued, which makes it a poor primary tool here. The directional signal — that current FCF cannot support $452 under any plausible growth path — is real and worth stating. The precise output is not.
7.4 Bogle expected return — the most useful output
5-year, on FY26 non-GAAP $12.09. Entry P/E 37.4x, dividend yield 0.47%.
| EPS growth ↓ / Exit P/E → | 20x | 24x | 28x |
|---|---|---|---|
| 8% | −3.3%/yr | 0.0%/yr | 2.8%/yr |
| 11% | −0.3%/yr | 3.0%/yr | 5.8%/yr |
| 14% | 2.7%/yr | 6.0%/yr | 8.8%/yr |
| 17% | 5.7%/yr | 9.0%/yr | 11.8%/yr |
At $452 the base case returns ~6%/yr over five years — inadequate for a beta-1.57 cyclical where the hurdle should be 11–12%. You only clear that with 17% sustained EPS growth and a 28x exit multiple, i.e. everything going right.
Solving for the entry price that delivers ~11%/yr:
| Scenario | Entry price |
|---|---|
| Conservative (11% growth, exit 22x) | $276 |
| Base (13% growth, exit 24x) | $332 |
| Optimistic (15% growth, exit 26x) | $396 |
7.5 Graham and DYT — reported for completeness, low weight
Graham IV = √(22.5 × EPS × BVPS), BVPS $30.11: - On reported EPS $10.65 → $84.94 - On normalized EPS $7.84 → $72.88
Weight: very low, and I won't pretend otherwise. Graham systematically lowballs asset-light, high-ROIC businesses — a company earning 40% ROE on a small book will always look absurd on a book-value-anchored formula. $85 is not a fair value estimate for AMAT; it is a reminder that there is essentially no asset backing under this price. Useful as a floor-of-last-resort intuition only.
DYT: yield 0.47% vs 5yr average 0.79% → signals expensive. Directionally consistent with everything else, but a weak confirming signal only — the payout is too small and too young to carry valuation weight. DDM: N/A.
7.6 Peer relative value — ⭐ this changes the recommendation
Live prices, 2026-07-29 intraday:
| Price | Today | vs our 7/15 accumulate zone | Status | |
|---|---|---|---|---|
| KLAC | $175.26 | −8.1% | $165–190 (7/28 revision) | ✅ IN ZONE |
| LRCX | $259.59 | −3.7% | $250–290 | ✅ IN ZONE |
| ASML | $1,569.87 | −0.8% | $1,300–1,500 | ❌ above zone |
| AMAT | $452.11 | −5.1% | $300–350 (set below) | ❌ ~30–50% above zone |
The whole complex fell into the zones the 7/15 compare set — except AMAT. And the 7/15 work named KLAC the preferred name on business quality (61% gross margin, 41% operating margin, ROE 95%, 52 consecutive quarters of services growth, majority share of process control).
This report's moat section independently confirms that call from the opposite direction: KLA holds 73.8% of process control while AMAT holds 9.8% and is losing it. So the two analyses agree from different angles.
If the user wants semi-cap equipment exposure today, the answer is KLAC at $175, not AMAT at $452. AMAT is the worst relative entry in its own peer group right now.
7.7 Fair value and scenarios
Weighting: FY26/FY27 earnings multiple 50%, Bogle 30%, P/FCF-vs-own-history 20%. Graham/DYT confirming only.
Fair value range: $290 – $370
Accumulate: $300 – $350 · Strong buy: below $270 · Trim/avoid: above $420
| Scenario | Prob. | Assumptions | 12–18mo price |
|---|---|---|---|
| Bull | 25% | CY26 +30% delivered, CY27 another record; FY27 EPS $16.90 achieved; multiple holds 28x; BESI acquired outright | $470 – $550 |
| Base | 45% | Strong FY26, FY27 EPS $15–16 (short of consensus); multiple settles at 22–24x as rate regime bites | $330 – $385 |
| Bear | 30% | Cycle peaks in FY27; EPS flat ~$12–13; multiple compresses to a cyclical 16–18x; China substitution accelerates; BESI lost to Lam | $190 – $235 |
Probability-weighted expected value ≈ $352 — roughly 22% below the current $452. Negative expected value at today's price.
The honest bull framing, stated fairly: if you believe consensus FY27 EPS of $16.90 and grant a 26x multiple (above AMAT's 20–22x historical average), you get $439 — approximately today's price. So AMAT is not insane at $452; it is priced for consensus to be delivered at a premium multiple, with no margin of safety. That is precisely what §3's conservative bias exists to refuse.
8. Conflict resolution — §5
Three genuine tensions. None is papered over.
Tension 1 — Fundamentals says ROIC collapsed 44% → 27%; Moat says Wide. Resolution: both are right, and they corroborate rather than contradict. First, a fair adjustment: stripping the $5,142M of long-term strategic investments (non-operating assets sitting in the denominator) lifts TTM pretax ROIC from 27.3% to ~33%. So the real compression is 44% → ~33% — material but less severe than the raw figure. Second, what remains is real: invested capital +72% against operating income +6.6%. And it matches the Moat analyst's independent finding of served-market share loss in 2025. Two analysts starting from different data reached the same conclusion: AMAT is spending progressively more to stand still. That mutual confirmation is the single most important analytical result in this report. Weighting: for a cyclical industrial at a cycle peak, I weight Fundamentals + Valuation over Moat's absolute rating — a 33% ROIC is still excellent, but the direction is what gets priced.
Tension 2 — Sentiment says every peer guided up and AMAT is supply-constrained; Fundamentals says operating income is flat; Moat says share is being lost. Resolution: all three are simultaneously true, and together they tell one coherent story. AMAT is participating in an accelerating market while losing relative share — revenue rises, ROIC falls, operating income lags. This is exactly the Moat analyst's "rising tide lifts Lam more" mechanism, and it is confirmed by Lam's GAA shipments tripling while AMAT's served-market share fell. The near-term income statement is genuinely strong (Q3 guides +22.6%); the long-term competitive trajectory is genuinely negative. Both belong in the verdict — which is why this is a WATCH and not an AVOID.
Tension 3 — 35 analysts, 0 sells, $628 mean target vs the CEO selling $105M at the top. Resolution: they are answering different questions. The analysts are right about the business trajectory (numbers went up, and they were right to raise). The insiders were right about the price. Both can hold: the business is very good and $736 was absurd. But note the asymmetry in signal quality — sell-side targets doubled in six weeks chasing price and are a lagging indicator, while Dickerson bought $6.9M at $137 and sold $105M at $590–736. On the specific question "is this stock cheap?", I weight the insider heavily and the consensus target near zero. The $628 mean target is not information; it is an echo of a price that no longer exists.
9. Verdict
🟡 WATCH — do not buy at $452
Conviction: 6.5 / 10 · Business quality 8.0 / 10 · Entry quality 2.5 / 10
Fair value $290–370 · Accumulate $300–350 · Strong buy <$270
Why not a buy. The framework's §0 test is unambiguous: "It's not just how good the company is — it's how well the market has priced it." AMAT is a good business at 42.5x trailing GAAP earnings that are 26% non-operating, 37.4x forward, and 67.6x free cash flow — roughly twice its own historical average multiple after a 39% crash. Base-case Bogle return is ~6%/yr against an 11–12% hurdle for a beta-1.57 cyclical. Probability-weighted fair value is ~22% below spot. There is no margin of safety, and §3 says rather miss an opportunity than overpay.
Why not an avoid, either. This is a genuinely high-quality franchise: net-cash balance sheet, ~30x interest coverage, 49.9% gross margin, an outstanding 8-year capital-allocation record, a real wide-moat services annuity in AGS, and a demand environment where management is constrained by its own supply chain rather than by orders — with every single peer guiding up and no forecaster projecting a 2027 WFE contraction. The −39% is a de-rating, not a broken thesis. This is a "great-ish company, wrong price" — the quadrant where you wait, not the quadrant where you walk away.
Key risks, named
- 🚩 Chinese state-backed substitution — the biggest threat and the only unmanageable one. 1.2% → 6.5% of global WFE in four years; >40% substitution in etch/deposition — precisely AMAT's segments, while ASML's litho and KLA's process control are untouched. Export controls destroyed AMAT's ability to compete on merit there and handed the domestics a captive market.
- 🚩 Peak-cycle multiple on flattered, mix-boosted earnings. Six straight up revenue years; FCF already −30% from peak; capex doubled; the record 50% gross margin is leading-edge mix, not pricing power, and mix reverses.
- 🚩 Served-market share loss in 2025 while Lam, KLA and ASML all gained — with the ion implant precedent (>75% → <60% over six years) proving quiet erosion is not hypothetical.
- ⚠️ Rate-regime multiple compression. Warsh Fed, 10yr at 4.71%, ~72% odds of a 2026 hike. The mechanism that caused the de-rate is still running.
- ⚠️ Customer concentration — 34% of revenue from two customers; a permanent margin ceiling.
- ⚠️ BESI is binary. Lam is also bidding. Losing it locks AMAT out of the packaging chokepoint — and it simultaneously drives GAAP earnings volatility (§3).
- ⚠️ MATCH Act (H.R. 8170) — would ban DUV/advanced etch to China. In committee. Binary tail risk.
What would change my mind — specific and falsifiable
Toward buying: - Price into $300–350 (or <$270 for a full position) - Insider buying on a Form 4 — Dickerson has bought a bottom before; that is the single signal I most want to see - Aug 13: non-GAAP EPS ≥ $3.36, revenue ≥ $8.95B, CY26 >30% guide reaffirmed or raised, and FCF conversion recovering - Capex guided to normalize (FCF/share turning back up toward $9+) - AMAT regaining served-market share in 2026/27 - AMAT winning BESI outright
Toward avoiding: - A third consecutive year of served-market share loss → moat downgrade to Narrow - FCF/share declining a fourth straight year with no capex normalization date - Lam acquires BESI - MATCH Act advances out of committee - CY26 equipment growth guide cut below +20%
A portfolio-specific passage was removed from the public build.
10. Catalyst calendar
| Date | Event | Why it matters |
|---|---|---|
| 2026-07-29 (today) | FOMC decision + Lam Research Q4 FY26 (post-close) | Rate path is the de-rating engine. Lam's WFE commentary is the closest read-through before 8/13. |
| ⭐ 2026-08-13 | AMAT FY26 Q3 earnings | The referendum. Guide ~$8.95B ±$0.5B, non-GAAP EPS $3.36 (cons. $3.38). Watch: CY26 >30% guide, 2027 language, China %, FCF conversion — and expect a BESI mark-to-market drag on GAAP that is not an operating miss. |
| 2026-08-20 | Ex-dividend $0.53 | Minor (0.47% yield) |
| Late Aug 2026 | NVIDIA Q2 FY27 | Sets AI-capex sentiment for the complex |
| Sept 2026 | FOMC | Hike vs hold resolves the compression driver |
| Late Sept 2026 | Micron FY26 Q4 + FY27 capex | Direct memory-WFE read (>$25B → ~$40B trajectory) |
| 2026-10-12 | EPIC Center unveiling, Sunnyvale | The moat-defense narrative event |
| ⭐ 2026-10-13 | Investor breakfast, SEMICON West (Dickerson + Hill) | Most likely venue for formal CY2027 guidance |
| Mid-Oct 2026 | ASML Q3, TSMC Q3 + 2027 capex | ASML is the lead indicator |
| Mid-Nov 2026 | AMAT FY26 Q4 + FY27 outlook | First full-year 2027 guide; also the quarter carrying the largest BESI mark |
| Ongoing | MATCH Act (H.R. 8170); BESI takeover resolution; Form 4 insider filings | Three binary watch items |
Appendix — resolved data discrepancies
| Item | Conflict | Resolution |
|---|---|---|
| TTM FCF | Sentiment cited $3.04B; Fundamentals computed $5.343B | $5,343M — verified two independent ways (quarterly sum matches stockanalysis exactly) |
| roic.ai FCF field | Returns $7,958M for FY25 = identical to OCF | Unusable for AMAT — cf_cap_expenditures is null so capex is never subtracted. Overstates by $2.26B (40%) |
| Yahoo "Interest Income" | $771M in Q2 FY26 on ~$8.2B cash = 38% implied yield | Misclassification of the $670–685M strategic-investment gain. Backing it out leaves ~$86M (~4.2% annualized) — sensible |
| DSO / receivables | Manager's initial read: 67d → 80d (red flag) | Corrected by Fundamentals: 78d → 73d, an improvement. Manager's calc used period-end AR against TTM revenue, which overstates for a company with rising revenue. Flag dropped |
| ROIC | 27.3% pretax raw | ~33% adjusted for the $5,142M non-operating strategic-investment book. Trend direction (44% → 33%) holds |
| Investment gain size | 10-Q says $670M; Q2 press release reconciliation implies ~$685M | Immaterial; the company-disclosed $(0.80)/share is used as authoritative |
| "Morgan Stanley downgrade" | Circulating in coverage | False. MS maintained Equal-Weight and raised PT $502 → $647 on 07-06 |
| "SK Hynix cut capex" | Circulating in coverage | False. Reallocated HBM4 lines to DDR5 within a rising capex envelope. Roughly revenue-neutral for AMAT |
| FY2016–2020 statements | Single-sourced (Macrotrends) | Flagged inline. FY2021+ confirmed by two independent sources |
| AMAT WFE share | Sources give 19%, 13.2%, 8.4% | Not asserted. Defensible statement: ~15–20% of WFE, #2 behind ASML |
| Segment comparability | 200mm moved AGS → Semi Systems (Q1 FY26); Display de-segmented (Oct 2025) | FY26 YoY segment comparisons are not like-for-like. AGS growth is inflated by mix cleanup |
Agents: Fundamentals · Sentiment & Intelligence · Moat (qualitative + quantitative) · Valuation (run by the manager after the dispatched agent stalled) · Portfolio Strategist (context).