AMAT

We rate Applied Materials Avoid: our central fair value is $290.17 a share, 29% below the $409.29 price (0.71x), and the price would need a cycle with no down year. Applied sells the machines that chipmakers use to build chips, so its sales follow how much chipmakers spend on new equipment, which fell in the downturns of 2009, 2019 and 2023. No single driver reaches the price inside our ranges. The Street's revenue through FY2029 is worth $309-$343 on our margins and exit, and only the Street's FY2030 jump takes it above the price ($434).

How to read the numbers.

• Cases. The central case is our best estimate. The bear case is a severe but plausible bad outcome and the bull case the matching good one. The central case sets the view; the bear and bull show how far the value could move.
• WFE and SEMI. WFE is wafer-fab equipment, the tools chipmakers buy to process wafers, and SEMI is the industry body whose forecasts we use. Applied's Systems revenue is its share of WFE.
• Segments. Semiconductor Systems ("Systems") is the tool business; AGS (Applied Global Services) is the services business; "Other" is display and related.
• The exit price at the end of FY2031 is the mean of two estimates. The long fade lets growth and returns drift toward mature levels over many years ($325). The outside view values FY2031 operating profit (EBIT) at the multiple similar companies have traded at over ten years ($423). NOPAT is operating profit after tax.
• Fiscal years. Applied's fiscal year ends in late October, so FY2026 ends 2026-10-25.
• GAAP and non-GAAP. GAAP is the standard accounting measure; non-GAAP is the company's adjusted measure, which leaves out items such as acquisition charges and restructuring.
• The Hold line is the fair value ($348, 0.85x the price) at which our rule would move the view from Avoid to Hold.
• Tags. [I] marks an assumption of our own with no dated source.

Applied sells the machines that deposit, etch, implant, polish and inspect the layers of a chip, and services the installed base. FY2026 revenue will be about $34.5B: Semiconductor Systems $26.2B, services (AGS) $6.9B and display and other $1.4B. Systems revenue is Applied's share, about 18.8%, of an industry WFE total that SEMI puts at $143.9B in 2026, +21.8% in 2027 and +14.1% in 2028. FY2027 Systems starts from our $8.03B forecast for Q4 FY2026 and grows 2% a quarter, which lifts the FY2027 share to 19.9%; the share then drifts down to 19.3% in FY2030, one point above FY2025 and half of what management targets. The central case takes SEMI's two years, then +9.6% in 2029 (SEMI's 300mm outlook of 2026-10-13), a flat 2030 that is our own undated pause assumption, and +4% in 2031. Systems gross margin is held at today's 55.3%, which keeps the operating margin near 35%, the level the Q4 FY2026 guide supports; management's claim is 40% non-GAAP. FY2031 revenue is $56.4B at a 35.2% operating margin. With 90% of free cash flow paid out, that is worth $290.17 a share today. The value rests mostly on how high the cycle sits in 2029-31, which no order book covers: only $28.75 of it is dated by cash on hand and the FY2027-28 payouts. The Q4 FY2026 release on 2026-11-12 and the Affiliates Rule decision on 2026-11-10 (the US export rule that extends controls to companies 50% or more owned by listed parties) are the next tests.

1. Business economics: tool spending per wafer rises, but the cycle sets the year

Who pays. Chipmakers buy tools when they add capacity or move to a new node. In Q3 FY2026 Systems revenue was 67% foundry and logic, 26% DRAM and 7% flash; the customers are few (TSMC, Samsung, Intel, SK hynix, Micron and Chinese fabs). Tool spend per wafer rises with each node, because gate-all- around transistors, backside power, 3D DRAM and hybrid bonding add process steps, and those steps are mostly deposition, etch and materials modification, where Applied is largest. Management says about 75% of WFE growth to 2030 is in areas where it leads (investor meeting 2026-10-13, management claim).

What limits it. The cycle: WFE fell in 2009, 2019 and 2023, and every fall cut Applied's margin. Company gross margin dropped 17 points in FY2009 and 1.3 points in FY2019's 13% revenue fall (10-Ks, sector note). China policy: management says its access to China's WFE fell from about 90% to about 70% in 2025. Competition in each tool line, and capacity: Applied is more than doubling its manufacturing capacity (Singapore campus) for a demand path that is not ordered yet.

Services. AGS sells spares, service contracts, upgrades and AIx analytics on 37,000 connected chambers, with contracts longer than two and a half years and above 90% renewal (management). It grew from $2.11B in FY2014 to $6.39B in FY2025 and never fell in a year. Its gross margin (35.6% in Q3 FY2026) is well below Systems (55.3%): it is steadier, not richer.

Cash. Applied pays out 80-100% of free cash flow. Net financial claims are small: $14.5B of cash and investments and a $1.2B CHIPS investment tax credit expected in cash, against $6.6B of debt at principal and $5.0B of leases, long-term tax payable, deferred tax on investment gains, pensions and operating cash (+$5.65B, $7.06 a share, with Q4 cash). Stock pay runs about $0.7B a year and is in the cost lines, GAAP and Applied's non-GAAP alike; buybacks offset dilution, which the 800M share count holds flat.

2. Results: Applied has never missed its revenue guide in eight quarters

Q3 FY2026 (reported 2026-08-13): revenue $9.115B against a $8.95B guide midpoint (+1.8%), GAAP gross margin 50.3% (+1.5 points year on year, the 13th straight rise), operating margin 33.7%, free cash flow $2.33B. The Q4 guide is $10.25B ± $0.5B (+51%), non-GAAP gross margin about 50.4% and non-GAAP EPS $4.02 ± $0.20.

The guidance record over eight quarters: actual revenue landed 0.0% to 3.4% above the midpoint, median +1.6%, never below. We forecast Q4 at $10.41B (range $10.15-10.70B), Systems $8.03B, AGS $1.86B, non-GAAP gross margin 50.5% and EPS $4.10; these are in assumptions.md, committed before the release.

Management's calendar 2030 model (2026-10-13): AGS above $12B, other above $2B, 54% non-GAAP gross margin, 40% non-GAAP operating margin, 15% tax and more than two points of WFE share gain from 2025. It is "not a WFE model"; Systems is to grow faster than the industry.

3. Our forecast gives $290.17, below management's targets

The model is model.py; its run is output.txt; inputs and reasons are in assumptions.md.

FY2027 has 53 weeks (it ends 2027-10-31; FY2026 ends 2026-10-25), and its first quarter has 14. The 53rd week adds about $0.8B of revenue in FY2027 and falls out again in FY2028. FY2027 starts from the Q4 FY2026 forecast: Systems $8.03B and R&D plus SG&A $1.58B a quarter, then Systems grows 2% a quarter and R&D plus SG&A rise 17 cents for each dollar of revenue above the Q4 run rate.

Against management. FY2027 Systems grows 31%: the Q4 FY2026 forecast annualised is already 23% above FY2026, and 2% a quarter on top fits "another strong growth year". Our FY2030 is below the 2030 model on every line: gross margin 50.9% (GAAP) against 54% (non-GAAP), operating margin 35.3% against 40%, AGS $10.4B against more than $12B, Other $1.8B against more than $2B. GAAP and non-GAAP margins differ by only 0.1 to 0.2 point in an ordinary quarter (Q2 FY2026 31.9% against 32.1%; Q3 FY2025 30.6% against 30.7%), so the target is about 39.8% in GAAP terms, not 39%: the 1.0-point gap in FY2025 (29.2% against 30.2%) holds a $181M restructuring charge in Q4 FY2025, and Q1 FY2026's 3.9-point gap holds a $253M legal settlement (releases 2025-11-13 and 2026-02-12).

Margin: held at what the guide supports. The central operating margin is held at the level the latest dated guide supports and stays flat from there. The Q4 FY2026 guide (revenue $10.25B, gross margin about 50.4%) with Q4 R&D plus SG&A of $1.58B gives 35.0% at the guide midpoint and 35.3% on our Q4 forecast; Q3's actual was 33.7%. Systems gross margin is held at Q3's and the guide's 55.3%, which gives 35.5% in FY2027 falling to 35.2% in FY2031 as AGS and Other margins rise a little. Management's 40% non-GAAP (39.8% GAAP) is a claim: the CFO's bridge on 2026-10-13 (value pricing, mix toward Systems above 55% of revenue, operational efficiency; a 54% gross margin by 2030 at about 100 basis points a year; 200-300 basis points of operating leverage from AI automation) names no dated pricing or mix action with a number, so the rule applies and the claim is disclosed. The remaining gap is 4.6 points (35.2% against 39.8%); closing it is worth +$35.30 a share and the Q3 actual of 33.7% is -$11.10. Before this revision the central margin rose to 36.9% on a Systems gross margin that gained 0.4 point a year; holding it flat is -$11.91.

Share: half the target, from FY2025. Management's target is more than two points of WFE share from 2025 to 2030. FY2025 was 18.3%; we take one point, so FY2030 is 19.3% (the target is 20.3%). The Q4 exit rate lifts the FY2027 share to 19.9%; the drift after that is the input that puts FY2030 at FY2025 plus one point, and it is negative (-0.9 point by FY2031) because the Q4 level partly reflects the FY2026 cycle. Before this revision the share kept rising, to 20.5% in FY2030 (FY2025 plus 2.1 points, above the target we said we were halving): the correction is -$17.57. FY2030 at 20.3% (the whole target) is +$15.05; back at 18.3% is -$13.94.

Against history. Revenue grows 10.4% a year FY2026-31, against the 12.2% a year management cites for the last ten years. The mature return, 48%, is Applied's own pooled figure for FY2016-25. From FY2027 R&D and SG&A rise 17 cents for each dollar of revenue added, in every case. 17 cents is an endpoint ratio: opex rose by 17.0% of the revenue added from FY2016 to FY2025 (12 cents in FY2016-22 while revenue grew, 44 cents in FY2022-25 while it was flat). The least-squares slope of opex on revenue over the same ten years is 15.8 cents, which is worth +$2.55. FY2027 starts from the Q4 FY2026 exit rate (R&D plus SG&A of 15.2% of revenue) and that anchor is worth +$11.67; we keep it, because the Q4 run rate is the nearest dated cost base. Opex is 15.6% of FY2031 revenue against a ten-year low of 16.1%; at 44 cents the value is $54.48 lower.

Value. The exit is the mean of the long fade ($325) and the outside view ($423, 16.7x EBIT, the median of LRCX, KLAC and ASML's ten-year median multiples). The outside view sits 30% above the fade because the peers' history did not include a fade to 3% growth. At WACC 10.22% and cost of equity 10.48%, fair value is $290.17. The cost of equity includes our standing calibration of -1.00 point (set 2026-10-16 so that the median fair value of our current batch of reports sits near 0.85x the Street target; it has no direct evidence); without it the value is $267.96.

Where we differ from the sector note. We now take its WFE path: our AI semis note proposed +9.6% in 2029 and flat in 2030, which matches SEMI's 300mm outlook of 2026-10-13 ($149B in 2026 to $227B in 2029, a 1.52x rise, as our 1.52x from $143.9B). The blind model had +5% and -3% and no dated reason for either. The flat 2030 is our own assumption [I], a pause after three double-digit years. It is undated, and it moves value toward the price: flat instead of the blind -3% is +$6.85 on the first revision's central (+$6.03 on this one), and no dip at all adds $8.04. If SEMI's growth stops after 2028 (flat to 2031), the value is $265.91 (-$24.26). We lean less on management's share target than its "more than two points": we take one.

Sensitivity. WFE in 2031 is the largest: -$35.8 at $190B and +$30.9 at $265B (+$36.7 at $272B, the highest six-year window in the sector note's series). The operating margin follows: 33.7% or 39.8% in FY2031 is -$11.1 or +$35.3. Opex added per revenue dollar at 12 or 44 cents is +$9.5 or -$54.5. FY2027 Systems growth after Q4, 0% or 4% a quarter against our 2%, is -$11.9 or +$12.2. FY2030 share at FY2025 or two points above it is -$13.9 or +$15.1. FY2031 NOPAT ±10% in the fade alone is ±$9.8. AGS growth ±2 points a year is about ±$2.5. Growth after FY2031 at 4% or 6% is -$4.8 or +$5.0. The mature return barely matters (under $2.1 between 30% and 50%), because returns on new capital are already high. Applied's own beta (1.48 against the peers' 1.47) is -$1.

4. Questions the sector note asked, and our answers

Our semiconductor-equipment sector note (note) set questions for each company. Our answers for Applied:

Which WFE total and what share? Management cites WFE without naming a source and gives no WFE forecast. On SEMI's total mapped to Applied's fiscal year, Systems was 19.1% in FY2024 (old segment basis), 18.3% in FY2025 (recast, with 200mm now in Systems) and 18.8% in FY2026 with the Q4 guide. The FY2024-25 fall is the China access loss management describes.

What is the revenue mix over the last eight quarters? (Systems revenue split; China share of total revenue; releases and 10-Qs):

FY2026 quarters are on the recast basis. Applied does not split leading edge from mature nodes, or HBM from other DRAM.

What happens in a year with WFE down 15-20% and no China fill? From the central FY2027, WFE -18%, China -20% and AGS +3% (its weakest history): revenue falls from $44.1B to $36.3B (-18%), EBIT from $15.6B to $10.3B, a 28.3% margin. The bear case spreads such a fall over FY2028-29.

How much of China is by customer type? Not disclosed. The one number is Applied's own Affiliates Rule estimate: about $0.6B of FY2026 revenue, about 6% of China revenue (8-K 2025-10-02). The rest of the split is a range we cannot set from filings. The bear case charges $0.6B a year, scaled with WFE.

Where did the gross margin gain since FY2022 come from? Company GAAP gross margin rose from 46.5% (FY2022) to 48.7% (FY2025) and 50.0% in FY2026 with the Q4 forecast. Management attributes it to value pricing and the portfolio; Applied does not split price, mix and cost. China fell from 37% to 28% of revenue over the same time and the margin still rose, so China mix was not the whole gain (inference). Our central holds Systems gross margin at today's 55.3%, so company gross margin edges up only with mix; the bear takes 2.5 points off in the down years.

What do backlog and terms show? The 10-Q gives no backlog figure. Contract liabilities rose from $2.57B (2025-10-26) to $3.27B (2026-07-26), +27%, while revenue was guided up: no sign of the sector note's test E8 (order books are real: contract liabilities falling while revenue is guided up). Cancellation terms are not disclosed beyond the 10-K's risk language.

What are capex and headcount for the peak? Capex rose from $1.19B (FY2024) to $2.26B (FY2025) and about $2.0B in the first three quarters of FY2026; the $500M Singapore campus doubles clean-room capacity there. Headcount was 36,500 at FY2025 end (35,700 a year earlier). The model keeps capex at 6.5% of revenue in FY2027 (a path of 8%, 7%, 6%, 5.5% and 5% is -$3.72 a share) and charges depreciation at 10% of plant; in the bear case the extra plant costs about $0.1B a year in depreciation (inference), small against gross profit. The larger trough cost is people, which the bear case charges by holding R&D and SG&A at their dollars when revenue falls (the 17-cent rule adds cost with revenue and never removes it).

Questions specific to Applied.

• AGS durability. Strong evidence: it never fell in a year, contracts run two and a half years with over 90% renewal, and 37,000 chambers are connected (90,000 planned by 2030). The contracted share is not disclosed; the subscription rate is "low 40%" rising to near 70% by 2030 (management). We take +15% in FY2027 fading to 8%, below the mid-teens model, and keep the margin at 35.8-37.4%, below Systems.
• DRAM and HBM packaging separately; Chinese DRAM in FY2024. Not disclosed. DRAM was 22-34% of Systems in the last eight quarters; Applied does not split Chinese DRAM customers.
• Display ("Other"). A drag for now: Q3 FY2026 other revenue $0.294B with a $0.118B operating loss (ramp costs). The CEO said on 2026-10-13 the two businesses run at "around $400 million per quarter" and that he would not model anything near term beyond that; the Q4 guide of about $0.51B is one quarter above it. We take $1.6B to $1.9B at a 28% gross margin; four times the Q4 guide ($2.08B) in FY2027-28 adds $0.52 a share.
• The suspended BIS denial order. Applied paid $253M in Q2 FY2026; the denial order is suspended and waived after three years if the audits are done (10-Q). A failed audit could revive the order, which would bar exports; that is a legal risk we cannot size, and it is not in the cases.
• Doubling output capacity by 2028. Capacity is for the path, not proof of demand. Our central FY2028 Systems revenue is $38.5B, 47% above FY2026; the bull reaches $42.6B. If demand stops at the FY2026 level, the fixed cost of the extra plant is about $0.1-0.2B a year of depreciation; the bear case covers it.

5. The bear case, the bull case, and why we sit in between

Bear ($154.92). FY2027 Systems stay at the Q4 FY2026 level (no sequential growth; WFE +12% in 2027), then WFE -15% and -5% (digestion after a three-year build, as in 2019 and 2023), reaching only $146B in 2031. Applied loses 1.5 points of share to China rules and Chinese toolmakers and $0.6B a year of China revenue. This is a share-loss bear, so the Q4 share step (the FY2027 level above FY2026) fades to zero by FY2031: FY2031 share is 17.3%, 1.1 points below FY2025 (-$20.32 against holding the step). Systems margin falls to 52%. R&D and SG&A follow the same 17-cent rule as the other cases, which keeps their dollars when revenue falls (+$5.32 against the growth lists the bear had). FY2031 operating margin 28.8%, NOPAT $8.3B. The mature return is the central's 48%: this world's lost share, price and volume are charged in its revenue, share and margins, and a 25% return on top charged them a second time (+$3.03). The bear's exit still averages two views that sit 41% apart, the fade at $152 and the outside view at $214. They differ because the fade values the bear's own FY2031 NOPAT, $8.3B growing 4% a year at a 48% mature return, 13.5x NOPAT, while the outside view applies the peers' ten-year median of 16.7x EBIT (about 19.6x NOPAT), a multiple earned in years when those companies grew; for a world of lost share and 4% growth the fade is the better guide, and the method averages both rather than choosing. This is a coherent world: Lam's revenue fell 39% in two quarters in 2023, and China was then filling the gap, which it may not do again.

Why the bull could be right ($399.42). SEMI and every toolmaker forecast growth to 2028-29, AI capacity is supply-limited, and Applied says WFE grows faster to 2030 than in the last decade. If WFE grows every year to $253B and Applied meets its share and margin targets, FY2031 NOPAT is $23.8B at a 38.4% operating margin (38.1% in FY2030). The bull's R&D and SG&A follow the same 17-cent rule, and the bull reaches the $409.29 price at about 14.7 cents, so it is $10 short of the price at 17. Its FY2030 share is 21.9%, FY2025 plus 3.6 points, more than management's target; at the target (20.3%) the bull is $371.48 (-$27.94). Its 2029 WFE moved from +8% to +10% in the first revision (+$7.92 on the bull, a change after the blind commit), and its Q4 run-rate step of 4% a quarter is part of the exit-rate change (+$15.03). Management has beaten its revenue guide eight quarters running, and margin has risen 13 quarters in a row.

Why we sit in between. Every past build ended in a down year, and its date was never known three years ahead. Applied lost share in 2024-25 and its margin target needs the pace of the best decade. We take SEMI's two dated years and its 300mm outlook for 2029 in full, and half of everything management targets beyond them.

6. Catalysts and what would prove us wrong

7. The price: what it requires, the Street through our model, and the market pass

The blind commit 788fbee fixed the first model (central $298.65). This section was added after the price was read (the market pass, a separate step after the analyst's report), and sections 1 to 6 were revised twice the same day: after a model audit and an opposite-side review, then after the model auditor's rerun. Section 8 lists every finding, what changed and the value before and after. The scripts for this section are market.py and street_path.py, with outputs market-output.txt and street_path.txt; revision.py prints the step from the blind model to this one.

Price and reaction. The last regular-session close, 2026-10-16, is $409.29 (market data SIP daily bar; last trade at 19:59:59 UTC was $409.38, 0.02% apart; the vendor's quote agrees). The stock fell 17.7% in the four sessions from the 2026-10-12 close of $497.53, including 6.7% on 2026-10-13 (close $464.24), the day of the investor meeting. It is 23.4% below $534.54, its close on 2026-08-13, the day of the Q3 report (after the close; the next day closed at $507.18, -5.1%). All closes are market data SIP daily bars. We read no news for any cause and do not infer one.

What the price requires. The central case is worth $290.17, 29% below the price; the bull, $399.42, is 2% below it. Alone, the price needs an exit price of $570 (22.6x FY2031 EBIT) or 2031 WFE of $373B, 21.3% a year from 2025. The highest six-year window in the sector note's own series reaches $272B (15.1% a year, the opposite-side advocate's reading of SEMI's history, not re-derived here); that is the top of our WFE range in both the base and the bull world, and the central at $272B is worth $326.90. In our simulation P is 8.0%: under 1% in the base world, under 1% in the bear world, 32% in the bull world, which carries a quarter of the weight.

The world where the price is right. Put every driver at the bull world's mode at once: 2031 WFE $255B, Systems growth 3.5% a quarter in FY2027, opex 15 cents per added revenue dollar, an 18x exit multiple, share drift +2.5 points, Systems gross margin 59% and AGS growth 14.4% a year. The central model gives $411.23 for that, just above the price. Added one at a time from the central $290.17: WFE $312.78, then $322.75, $328.62, $340.00, $375.97, $405.04 and $411.23 (revision-output.txt). Each driver is inside its range; the price needs all seven together, which is what the 8.0% prices (a quarter of the scenarios are this world, and a third of those reach the price).

The Street, run through our model. Fair value is 0.44x the average target ($659.83). The Street's market data annual estimates, fiscal years ending October (data/street-estimates.json), against ours:

Street EBIT is implied by the Street's net income (market data netIncomeAvg), over one less the tax rate (11%, 13%, 15%), less $0.3B of interest and other income. The diluted shares that net income and EPS imply are 778M for FY2027 and 750M for FY2028 (808M for FY2029-30, a different and thinner set of analysts), not 800M; EBIT built from EPS times 800M had overstated the Street's FY2027 and FY2028 by 3% and 7%. The vendor's ebitAvg field is not used: it gives $13.44B for FY2026 against Applied's own GAAP operating income of about $11.0B. Street revenue goes in through Applied's FY2027 Systems growth and then through WFE, keeping our segment margins and our opex rule. With R&D and SG&A held at our dollars the same steps give $342, $392, $387 and $525; solving Systems gross margin to the Street's implied EBIT as well gives $309, $344, $343 and $503 (Systems gross margin 62% in FY2030-31). Step by step from our $290: FY2027 +$19, FY2028 +$34, FY2029 -$3, FY2030 +$94, then $226 remain to the $660 target.

• FY2027 is +$19, and all of it is a higher base carried forward. Street FY2027 revenue is 6% above ours and needs Systems growth of 5.1% a quarter against our 2%. As a one-year effect (FY2027 payouts alone, nothing carried into FY2028) it is worth +$0.26.
• FY2028-29 add +$31 together. The Street's revenue is 17% above ours in both years; FY2029 gives back $3 because the Street's FY2029 sits below our path rolled off its higher FY2028 base.
• FY2030 is the largest single step (+$94). The Street's FY2030 is +28% on FY2029 with a 42.3% margin, from 12 analysts (7 for EPS); its low estimate ($72.6B) is above the FY2029 high ($68.3B). At a share of 20.9% (our drift from the level the Street's FY2027 implies), $80.4B needs calendar 2030 WFE of $342B, +39% on the Street's own 2029 level of $247B, against our $219B. The analysts behind that year may differ from those before it; the Street's summary does not say. It is the biggest part of the target and rests on thin coverage.
• The price sits between the Street's FY2029 and FY2030. The Street's FY2027-29 revenue on our margins and exit is worth $309-$343, 16-24% below the price; with its margins too, $343 through FY2029. It takes a part of the FY2030 step, which is 12 analysts and a $342B WFE.
• The Street's margin is above ours after FY2027. Its implied EBIT margin is 0.3 point below ours in FY2027, then 1.8, 4.6 and 7.0 points above. Applied's non-GAAP margin leaves out acquisition charges, restructuring and legal settlements but not stock pay (its earnings releases), so the gap is not stock pay: it is the Street's operating leverage against our 17 cents of opex per added revenue dollar and flat Systems gross margin. The margin adds $0 through FY2027, $3 through FY2029 and $69 through FY2030.

Range. Our 10th to 90th range, $164-$403, is 65% as wide as the Street's target spread ($530-$900), so it is not narrow against it. The price sits near our 92nd percentile (P 8.0%). We do not widen it toward the price.

Alpha's view (dated 2026-10-17; alpha ticker analyze AMAT, saved in alpha-ticker.json). Core score 30.4 of 100, Below Average. No ordinal rank: the rank cache is stale and data health is limited. By category, valuation C (31st percentile), expectations C (31st), quality C (36th), growth C (62nd), momentum C (60th). The momentum lens reads breakdown_avoid (score 26): the close is below the 50-day ($482) and 200-day ($438) averages, 3-month return -27.4%, RSI 29.6, against a 12-month return of +80.7%. Valuation: P/E 35.1 against its 5-year median 21.4, richer than its own history on all four multiples (vendor history stale). Why we differ. We do not, in direction: both say avoid. Alpha avoids on a broken trend and a rich multiple; we avoid because our cash flows to FY2031 fall short of the price. The reasons are independent, and neither is a veto on the other. Were the price to fall toward our $290, Alpha's momentum read and our value would part the other way.

Technical read (technical.txt; timing beside the view, never an input). Grade D. Stage topping (close below the rising 30-week average of $479), structure damaged (the $411.51 swing low of 2026-09-16 broken), not extended (-3.3 ATR from the 21-day average), relative strength 29% below its 63-session high against QQQ, market and group (SMH) both hostile. It rises on a weekly close above $479. Earnings 2026-11-12 is 19 sessions away.

Variant view and what settles it. The price discounts more than the Street's FY2027-29 revenue on our margins. Our differences are a cycle that pauses in 2030, an operating margin held at the 35% the Q4 guide supports where management claims 40% non-GAAP, and one point of share above FY2025. The Q4 FY2026 release and the Q1 FY2027 guide on 2026-11-12 (after the close; verified in earnings calendar) settle FY2027: our Q1 guide midpoint forecast is $11.20B for a 14-week quarter, $10.91B in the bear case and $11.45B in the bull case (each our Q1 revenue less the median guide beat of 1.6%). A midpoint above $11.45B is the bull path; below $10.91B is the bear path.

The Affiliates Rule date. The 8-K of 2025-10-02 has the suspension ending 2026-11-10, which is the catalyst date above. The opposite-side review found a 2026-09-25 report that it was extended to 2027-01-10; it is a search snippet, not confirmed in the Federal Register, and we did not use it. If it holds, the catalyst moves to 2027-01-10 and the risk stays about -$4 a share if the rule is permanent (inference).

8. Review findings and answers

Two rounds of review read this report with the market known. The model auditor is an AI reviewer that checks the model's mechanics and numbers; the opposite-side advocate is an AI reviewer asked to argue against the Avoid. Round 1: a model audit (14 findings) and an opposite-side review (15 findings), with decisions by our research lead on four of them. Round 2: the model auditor reran every number of round 1, reproduced them, agreed that the Avoid survives, and left ten fixes. Every answer below is fixed, disclosed or rejected. The step from the blind model to this one is revision.py (revision-output.txt): it rebuilds the blind state and adds one fix at a time, so each dollar effect depends on the order shown. A change was made only on dated evidence or on a rule that does not depend on the price; the price moved no assumption.

Round 2: the bridge from $319.65 to $290.17.

The odds change is separate from the table: the WFE top of 330 becomes 272 in the base and the bull world, the share driver is redefined around the new central, and Systems gross margin centres on 55.3%; P is 8.0%, with a 10th, 50th and 90th percentile of $164, $273 and $403.

Four concerns raised by our automated report check.

• The net move from the blind $298.65 to $290.17 (-$8.48, -2.8%). The largest single change is the WFE path, +$23.95, which rests on SEMI's dated 300mm outlook of 2026-10-13, not on a management claim or a Street figure. It is offset by the opex rule (-$17.92), the share path (-$17.57) and the margin hold (-$11.91), each on Applied's own filings or guide.
• The probability that the price is justified, 8.0%. Answered by number in section 7: the price needs all seven drivers at the bull world's mode at once ($411.23), and the bull world's P is 32%, a quarter of the scenarios.
• The Street. Answered in section 7: the Street's revenue by year through our model, step by step, $19, $34, -$3 and $94 from our $290.
• The range. Answered in section 7: 65% as wide as the Street's spread; the price sits near our 92nd percentile; not widened toward the price.

Round 2: the model auditor's ten fixes

1. Share: fixed. FY2030 share is FY2025 (18.3%) + 1.0 point, 19.3%, not 20.5%. The input is the drift after the Q4-set FY2027 level, now -0.92 point by FY2031 (share_gain -0.0092). -$17.57 (the auditor: -$17.59). The target met (20.3%) is +$15.05.
2. Bear: fixed. The Q4 share step fades to zero by FY2031 (-$20.32; the auditor's bear $146.57 reproduces exactly). The 17-cent opex rule now applies to the bear's later years as well (+$5.32; see the opex fix below). The bear's 25% mature return charged its lost share and price a second time; it is now the central's 48% (+$3.03 in this order; the auditor's +$3.42 was on the first revision's bear). Why the bear's fade value and outside view differ is stated in section 5. Bear $166.89 becomes $154.92.
3. Margin: rule applied, claim disclosed. The central margin is held at the level the latest dated guide supports, the Q4 FY2026 guide: 35.0% at its midpoint, 35.3% on our Q4, with Systems gross margin flat at 55.3%: 35.5% in FY2027, 35.2% in FY2031 (-$11.91). The CFO's margin bridge of 2026-10-13 (value pricing, mix, operational efficiency, 200-300 basis points from AI automation) has no dated action with a number, so management's 40% non-GAAP (39.8% GAAP) is quoted as a claim. The gap is 4.6 points: closing it is +$35.30, and Q3's actual 33.7% is -$11.10 (the auditor's -$25.76 for 33.7% was against the first revision's 36.9%).
4. Opex: fixed and disclosed. The rule is in every case, the bear's later years included. The Q4 opex anchor is worth +$11.67 (see the FY2027 sequential-growth item in round 1) and is kept. 17 cents is an endpoint ratio; the least-squares slope is 15.8 cents (+$2.55). The bull text now matches its FY2030 margin of 38.1%, and the bull reaches the price at about 14.7 cents (section 5).
5. The flat 2030: labelled as our assumption [I]. It is undated, and it moved value toward the price by +$6.85 on the first revision's central (+$6.03 on this one); stated in section 3 and drivers.json.
6. Odds ceiling: fixed. WFE 2031 maximum 272, the highest six-year window in the sector note's own series, in the base and the bull world. P is 8.0% (the auditor, on the first revision's model: 8.2%). The probability concern is answered by number in section 7.
7. Bull 2029: disclosed. Bull 2029 WFE moved from +8% to +10% (+$7.92 on the bull; the advocate's check of changes after the blind commit); its Q4 run-rate step of 4% a quarter is part of the +$15.03 exit-rate change (section 5).
8. Q1 FY2027 thresholds: tied to the cases. Revenue for the 14-week quarter is $11.08B, $11.38B and $11.63B in the bear, central and bull case; less the median 1.6% beat, the guide midpoints are $10.91B, $11.20B and $11.45B (output.txt), replacing $10.6B, $11.15B and $11.7B in record.json and sections 6 and 7.
9. The share count in street_path.py: fixed. EBIT comes from market data net income, whose implied shares are 778M (FY2027) and 750M (FY2028), not 800M; printed in street_path.txt.
10. record.json: fixed. The summary and numbers match this version. The SEMI year-end entry had a placeholder date (2026-12-31); no SEMI date is in the folder and the record needs a date for every catalyst, so it moves from the catalysts to the falsifiers and section 6 says no date is set.

Round 1: model audit and opposite-side review (values as of the first revision, central $319.65, superseded where round 2 says so)

Decisions by our research lead

• 2029-31 WFE path: fixed (SEMI-consistent path, dip disclosed). 2029 +5% becomes +9.6%, 2030 -3% becomes flat, 2031 stays +4%: 2031 WFE $211.8B becomes $227.9B (10.4% to 11.8% a year from 2025). +$23.95. Evidence: SEMI's 300mm outlook of 2026-10-13, $149B in 2026 to $227B in 2029, a 1.52x rise, the same as our 1.52x from $143.9B. The flat 2030 is our own dip assumption (a pause after three double-digit years; no dated source): no dip is +$9.13, the blind path is -$16.63.
• Plain errors: fixed. The street_path.py EBIT field and line 46, the FY2026 base, the tax credit refund, the FY2027 sequential shape and the 53-week year are answered below. other is rejected (the auditor's third item).
• Odds: fixed. Ranges widened to the evidence in the folder, not to the price: WFE 2031 150-265 becomes 150-330 (base), bull world 200-300 becomes 205-330, bear world 125-185 becomes 125-190; and three drivers that were not simulated are now: FY2027 Systems growth a quarter (0-5%), opex per added revenue dollar (12-44%) and the exit multiple (14.2-20x). P 8.7% becomes 15%; 10th to 90th $161-$404 becomes $164-$429.
• Margin, GAAP against non-GAAP: settled from the filings. The gap is 0.1-0.2 point in an ordinary quarter (Q2 FY2026 31.9% against 32.1%, Q3 FY2025 30.6% against 30.7%, Q2 FY2025 30.5% against 30.7%). The gaps of 1.0 (FY2025 29.2% against 30.2%), 3.4 (Q4 FY2025) and 3.9 points (Q1 FY2026) hold a $181M restructuring charge and a $253M legal settlement (releases 2025-11-13 and 2026-02-12). The auditor was right about the ordinary gap; the advocate's 1.0 is a year with charges. The 40% non-GAAP target is about 39.8% GAAP, and our FY2031 35.2% sits 4.6 points below it. Decision: no margin change beyond the opex rule below.

Model auditor, round 1

1. FY2027 sequential growth: fixed. FY2027 Systems $32.11B (four times the $8.03B Q4 forecast) becomes $34.41B at 2% a quarter from the Q4 level, with R&D and SG&A also starting from the Q4 run rate: +$11.90 on the central. 0% a quarter is -$12.46 and 4% is +$12.81. Evidence: the Q4 FY2026 forecast and guide ($7.9B, +12% on Q3's $7.04B), management's "another strong growth year" (Q3 call 2026-08-13). The auditor's +$8.5, +$20.2 and +$30.6 roll SEMI's growth off the higher base on the blind opex; ours carries the new level as a permanent share step and charges the opex that comes with it, so it adds less.
2. FY2026 base on $7.9B: fixed. Systems $26.05B becomes $26.18B ($8.03B Q4), AGS and Other and gross profit likewise; base share 18.7% becomes 18.8%: +$1.80.
3. Other at four times $0.51B: rejected. FY2027 stays $1.6B. The CEO said on 2026-10-13 that the two businesses are at "around $400 million per quarter" and that he would not model anything near term beyond that (data/calls/); the Q4 guide of $0.51B is one quarter above it. At $2.08B in FY2027-28 the value is +$0.52 (the auditor's +$0.71 was on the blind model).
4. A fall assumed in 2030 without a dated cause: fixed with disclosure. See the research lead's decision above: -3% becomes flat, +$23.95 together with 2029; the flat 2030 is labelled as our assumption in the report and in drivers.json.
5. Margin from opex: fixed. From FY2027 R&D and SG&A rise 17 cents for each dollar of revenue added (Applied's FY2016-25 pass-through, Market data annual; 12 cents in FY2016-22, 44 cents in FY2022-25). FY2031 operating margin 37.7% becomes 36.9%; opex 14.5% of FY2031 revenue becomes 15.7% (a ten-year low of 16.1%; FY2027 starts at the Q4 FY2026 ratio of 15.2%): -$17.92. The auditor's FY2031 margins of 36%, 35%, 34% and 32.1% through opex are -$7.40, -$15.58, -$23.76 and -$39.30 on this model; at 44 cents the value is -$66.65. The same rule now applies to the bull, whose opex no longer outruns its revenue (bull FY2031 margin 41.4% becomes 38.4%; the rule is -$32.43 on the bull).
6. Where the fade starts: disclosed. The fade starts at the return on fixed and intangible capital; the auditor's 56.7% with working capital and -$1.07 to -$1.64 are not re-derived here. It is our standard method (fade-peer-2026-10-16), so it is not changed in one report.
7. EBIT in street_path.py: fixed. the vendor's ebitAvg ($13.44B for FY2026, Applied's own GAAP $11.0B) is replaced by EBIT implied by the Street's EPS (section 7): FY2026 $11.23B, FY2027 $16.92B, FY2028 $22.76B, FY2029 $24.89B, FY2030 $33.68B. The Street's margin now adds $0 through FY2027 and $3 through FY2029; the FY2027 step including the Street's margin is +$20.34, of which the margin is +$0.05.
8. Our standing cost-of-equity calibration: disclosed. Without the -1.00 point the value is $294.93 (-$24.72); the auditor's $275.65 was on the blind model. It is now stated in section 3 as a calibration setting without direct evidence.
9. $1.2B investment tax credit refund: fixed. Added to the bridge from 10-Q Note 8, undiscounted, timing not stated: +$1.50.
10. Small bridge items: fixed and disclosed. Tax on unrealized gains -$0.2B and the pension deficit -$0.146B together -$0.47 a share; debt at principal $6.6B against $6.544B -$0.07. The $3.0B operating cash and $1.6B Q4 cash kept stay as labelled assumptions; each $1B is $1.25 a share.
11. Capex: disclosed. 9M FY2026 capex was 8.3% of revenue against our 6.5% in FY2027; a path of 8%, 7%, 6%, 5.5% and 5% is -$3.82. Not changed: the Singapore campus build is lumpy and revenue grows 28% in FY2027, and the effect is small.
12. Tax 17% for FY2028-31: rejected. The CFO's estimated tax rate for 2030 is 15% (investor meeting 2026-10-13); 17% is -$4.07.
13. Peer set: disclosed, model matches. The outside view uses the median of LRCX, KLAC and ASML ten-year medians (16.7x), as the report says. Without ASML it is 15.4x and -$10.41.
14. Bull mature-return text: fixed. "FY2021-25 above 45%" becomes "FY2021-24 above 49%; FY2025 43.0%" in model.py.

Opposite-side advocate, round 1

1. WFE growth equals the weakest six-year window: disclosed, ranges widened. Our 2025-31 pace is 11.8% a year after the path change (it was 10.4%, the weakest). The median window, $252.6B in 2031, is +$23.41; the flip fact of $250B is +$20.95 (WFE of $258B reaches the Hold line, 14.1% a year). Not adopted as central: the windows are the advocate's reading of a sector note not in this folder, and beyond 2029 there is no dated source. Its extremes set the new range (to $330B) and the bull world.
2. The 2030 fall of 3% is undated: fixed (see the auditor's item on the 2030 fall).
3. 2029 +5% against SEMI's 300mm outlook: fixed. +9.6%; the advocate's +10% with 2031 at $238B is +$6.90 on this model.
4. SEMI's 2029 forecast rose 32% in six months: fixed in the odds. Wider ranges (the research lead's decision above).
5. FY2027 has 53 weeks and a 14-week Q1: fixed. Applied's year ends on the last Sunday of October: FY2026 on 2026-10-25, FY2027 on 2027-10-31, so FY2027 has 53 weeks and Q1 FY2027 (to 2027-01-31) has 14. Systems, AGS and opex carry the 53rd week in FY2027 and drop it in FY2028: +$0.32. The Q1 FY2027 guide forecast of $10.6B (range $10.0-11.2B) implied a 5.5% weekly decline from Q4; it is restated to $11.15B (range $10.6-11.7B): our Q1 revenue of about $11.3B for 14 weeks less the median guide beat of 1.6% (Q4 flat per week is $11.21B). The bull threshold is now $11.7B, not $11.2B.
6. Bug in street_path.py line 46: fixed. base[k+1] replaces base[k]; with the revised model the Street's FY2027-29 revenue on our margins is worth $340-$372 (the report had $375-$400 from the buggy script and the blind model).
7. Street margin and the "Systems gross margin 65%": fixed (see the auditor's item on EBIT in street_path.py). The 65% was an artifact of the Market data EBIT field; the implied Systems gross margin now peaks at 62% in FY2030-31.
8. FY2027 Street step is a higher base carried forward: confirmed. +$20 carried, +$0.27 as a one-year effect.
9. Margin target about 39% GAAP: rejected, settled (the research lead's decision above): the target is 39.8% GAAP. semi_gm FY2031 60.1% is +$17.59; not adopted. Our 57.3% is about half of the company's own gross-margin pace since FY2015 (0.8 point a year, 40.9% to 48.7%) and management's 1 point; both rose in good years and fell in bad ones.
10. Services: disclosed, not changed. AGS growth of 20%, 16%, 14%, 14% and 14% is +$7.16. Q3 growth of 22% and "above $12B by calendar 2030" are management's; AGS grew 10.6% a year over FY2014-25 and the 2026 rate includes high utilisation. Ours is $10.4B in FY2030; +2 points a year is +$2.61.
11. Share +2 points: rejected. +$9.14. It is management's target, taken at half (the 2025 share fell 0.8 point).
12. The Street's FY2030 is indefensible: confirmed (section 7): $324B of calendar 2030 WFE against the Street's own $238B for 2029.
13. Truncated simulation: fixed. WFE base max 330 and bull max 330 (the advocate asked for 300 and 350); FY2027 level, opex and exit multiple are now drivers. P is 15% (the advocate's re-run: 14.5%); 90th percentile $429 (the advocate's: $431).
14. Affiliates Rule date: disclosed (section 7). Unconfirmed; the 8-K date is kept.
15. Peers' latest five-year EV/EBIT of 20.9x: disclosed, not used. It is price-derived. The reference-only lines in output.txt stay: the peers' trailing 39.0x gives $506.63 and forward 25.3x gives $391.52.

The world where the price is right. Seven things at once: Systems growth 2.7% a quarter, 2029 +10%, 2030 +6%, FY2030 share at FY2025 + 2 points, Systems margin 60.1%, AGS growth 15.6% and opex at 12 cents. The advocate's value was $404.5 on the blind model; our model gives $379.51 for the same recipe (revision-output.txt). The recipe of round 2 is the bull world's mode on all seven drivers, $411.23 (section 7).

The flip fact. WFE growing 14% a year to about $250B in 2031 gives $308.62 here (the advocate: $348); the Hold line needs $297B.

9. Sources (read 2026-10-18)

• Applied 8-K earnings releases, Q3 FY2024 to Q3 FY2026 (masked; data/edgar/ and the sector note's data/edgar/), 10-Q for the quarter to 2026-07-26, 10-Ks FY2024 and FY2025 (sector note data/edgar/).
• 8-K 2025-10-02 (Affiliates Rule estimate) and 8-K 2026-08-27 (board appointment, not material), data/edgar/.
• Q3 FY2026 call, 2026-08-13 (sector note data/calls/); investor meeting 2026-10-13 (Quartr via market data, market paragraphs removed, data/calls/).
• Market data statements for Applied (income, balance sheet, cash flow, annual and quarterly; segment and geographic revenue), data/market-data/; peers' key metrics, statements and estimates, data/peers/.
• SEMI mid-year forecast 2026-07-14 and 300mm outlook 2026-10-13, via the sector note data/industry/.
• Damodaran implied equity risk premium and the 10-year Treasury, 2026-10-01 (our standing market inputs).
• Sector note docs/research/sectors/semi-equipment/note.md; our ASML report docs/research/2026-10-15-asml/assumptions.md (operating sections only).