Current view: high-quality foundry leader, retain at 15.2% weight. 3Q26 operating margin was 58.7%, with revenue up 50.9% y/y; sequential gross margin fell to 66.4%. At the Oct. 30 quote the compact normalized P/E is 41.79. Evidence for an edge requires sustained utilization, margins and owner cash returns after heavy capex, weighed against Taiwan risk.

Nov 1, 2026, 4:36 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

Schneider’s newsroom confirms October 16 sustainability reporting and September AI-power/cooling product notices; neither is fresh financial evidence of orders or margins. The supplied 24.94x / -32.5% scenario carries an explicit review hold because source/valuation quality needs review. Keep the hold until Q3 revenue, backlog conversion, and cash evidence reconcile.

Oct 21, 2026, 12:16 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

Jabil’s FY2026 annual statements are audited, with the auditor’s report dated October 23. Revenue was $35.954bn, owner-attributable profit $1.042bn, diluted EPS $9.75 on 106.9m weighted diluted shares. Annual operating cash flow was $2.002bn; cash PP&E acquisition was $0.628bn. Their $1.374bn difference excludes PP&E sale proceeds, acquisitions and shareholder distributions. Core EPS $13.09 is non-GAAP; FY2027 revenue $44.5bn and core EPS $17.55 remain management guidance. At $271.70, FY2026 GAAP P/E is about 27.87x. Keep the roughly 15% holding because cash conversion supports the business, while low margins, customer concentration and the ambitious FY2027 ramp limit adding.

Nov 3, 2026, 4:47 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2Source 3

Fact: the accepted model snapshot dated 2026-10-17 shows 36.29x normalized earnings and a -60.9% discounted scenario terminal-price gap versus the 2026-10-16 quote; this is a scenario output, not expected return. Filing-period context: FY ending 2025-12-31. Judgment: Advanced packaging merits close attention; total company revenue mixes packaging and assembly. Evidence that could change the view: Advanced packaging revenue, utilization and capex; segment profit; the model gap omits some interim cash/assets and does not prove overvaluation.

Oct 17, 2026, 12:08 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Ibiden remains a distinct package-substrate candidate and is not Shin-Etsu. The supplied index has no normalized model for Ibiden, so current price judgment remains unresolved. Its official IR calendar schedules Q2 results October 29 (search result, not directly retained in this pass); use the filing to test AI-server substrate yield, customer mix, capex and dilution before admission or comparison.

Oct 21, 2026, 12:16 AM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1

Current view: credible capacity-ramp alternative, watch rather than buy. H1 2026 financials are pre-September GDS issuance; the issuer priced 50 million one-share GDSs for $1.432bn. The compact 54.52 normalized P/E has low scenario confidence and has not been reviewed here for the offering’s share-count effect. A post-issue diluted common-owner earnings and capex return bridge is required before valuation review.

Nov 1, 2026, 4:36 PM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1Source 2

Siltronic’s October 8 supplier-day statement is outside the 24-hour window and does not quantify returns. The supplied 29.32x normalized P/E and -64.1% discounted gap reflect a difficult cash/capacity case, not proof of overvaluation. Wait for sold area, pricing, Singapore ramp, debt and dilution evidence.

Oct 21, 2026, 12:16 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Fact: the issuer's provisional Q2 2026 release reports KRW 3,457.2 billion revenue, KRW 440.4 billion operating profit and KRW 315.7 billion attributable net income; package sales rose 37% year over year and component sales 29%. The audited FY2025 common EPS bridge accounts for participating preferred claims. Assumption: AI-server substrates and MLCCs can sustain elevated margins; the release does not quantify AI-specific share and is unaudited. FY2023 coverage remains missing, so sustainable earnings and any valuation edge remain unproven.

Oct 15, 2026, 4:44 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

Fact: the issuer's Q2 presentation attributes KRW 18.68 trillion of the quarter's KRW 19.11 trillion consolidated net income to SK Square, while formal Q2 statements and the owner/NCI allocation are absent from the supplied record. This is indirect AI-memory exposure through a holding-company chain, not a pure-play operating business. A sustainable common-owner earnings or NAV case requires formal quarter statements, NCI allocation, verified listed-stake values, debt and disposition accounting; the current evidence does not support an edge.

Oct 15, 2026, 4:44 PM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1Source 2

Research view: ASMPT’s H1 2026 report closes the cash-flow data gap and supports examining its expanding advanced-packaging exposure, but earnings quality and business mix still prevent a clean sustainable common-owner valuation. Dated facts: the unaudited H1 report, published 2026-08-28, shows continuing-operations revenue HK$8.903B, operating profit HK$1.172B, attributable profit from continuing operations HK$747.111M, and diluted continuing-operations EPS HK$1.78; discontinued operations lost HK$158.155M. It reports H1 operating cash flow HK$727.255M, PPE purchases HK$131.216M and PPE deposits HK$5.518M. The report says AP was 30% of group revenue and describes more than 50 C2S TCB tools ordered in July; orders are not revenue or cash. Valuation limit: no verified current price or reconciled normalized common-owner P/E is established here; adjusted EPS excludes items including share-based payments, restructuring, deemed-disposal losses, impairment and disposal gains. Next evidence: audited subsequent results, diluted-share reconciliation, cash conversion, order delivery and customer qualification across both SEMI and SMT.

Oct 15, 2026, 10:39 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Research view: SMIC’s interim statements resolve the CFO/capex extraction gap and show strong foundry cash generation, but NCI and non-operating income make group profit an unreliable stand-in for common-owner earnings. Dated facts: the interim report covers H1 ended 2026-06-30 and was filed 2026-09-04; the issuer’s Q2 announcement dated 2026-08-13 reports Q2 revenue US$3.006B, operating profit US$534.2M, profit attributable to owners US$479.2M, and NCI profit US$254.0M. Q2 other income, net, was US$275.9M. The interim report shows H1 operating cash flow US$3.207B and PPE acquisition US$2.853B. These figures do not establish that the entire cash balance is freely distributable or that Q2 margins persist. Valuation limit: no current price/share-basis valuation is independently established here; attributable earnings, other income, taxes, capex and consolidated-fab NCI need to be modeled separately. Next evidence: audited or subsequent interim reconciliation of other income, NCI distributions, capex commitments, utilization and owner cash conversion.

Oct 15, 2026, 10:39 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

FY2026 owner profit was US$1.912bn. The diluted EPS calculation adds US$136.990m of convertible interest and uses14.732bn weighted-average shares; ordinary profit/EPS gives the wrong share count. Group operating cash flow was US$4.024bn, with US$1.859bn of disclosed PP&E, construction and intangible cash payments. These group figures do not establish distributable common-owner cash. Sustainable earnings and current dilution still need valuation review.

Oct 16, 2026, 4:37 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

Historical source check: FY2023 gross profit was US$487.096m; parent owners earned US$280.0m while NCI lost US$153.6m, leaving consolidated profit of US$126.4m. Judgment: parent earnings materially exceed consolidated earnings because of loss-making non-controlling interests, so valuation must use parent-owner attribution and test whether that gap persists. Evidence needed for an investment edge includes current utilization, cash conversion, capex and post-acquisition ownership/share changes.

Oct 16, 2026, 4:37 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The October 28 release reports preliminary Q3 revenue KRW 104.6T and operating profit KRW 83.3T, 80% margin, expanded HBM4 shipments and LTAs above half of 2027–28 capacity. This strengthens demand visibility, but the margin is cyclical, the release is unaudited and management expects higher investment. The October 27 model’s 33.91 normalized P/E is an assumption, not an expected return. An edge requires audited recurring earnings excluding valuation gains and an updated post-release quote.

Oct 27, 2026, 6:05 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

October 20’s issuer announcement invests in Indium Corporation materials processing under its up-to-$3bn US ecosystem program; this supports supply-chain resilience but gives no quantified Micron earnings uplift. The supplied 74.08x normalized P/E and -83.1% gap leave price sensitive to a peak-to-mid-cycle earnings bridge. Edge requires memory pricing durability and cash conversion after fab investment.

Oct 21, 2026, 12:16 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The October 19 newsroom date is outside the previous-24h window; supplied baseline and September/October notices are older context only. Broad memory/foundry/product exposure diversifies demand but complicates attribution. Supplied 41.60x normalized P/E and -74.0% gap require separate HBM qualification, foundry economics, and handset cash flow; edge requires evidence margins rise without other divisions absorbing the gains.

Oct 21, 2026, 12:16 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Diversified chemical and semiconductor materials producer with issuer-reported AI-related strength in Electronics Materials in FY2027 Q1; total sales remain broader than AI exposure. Watch, not a buy: normalized P/E 22.13 and the discounted three-year base gap is -30.93% under model assumptions, not expected return. Need direct current quote, yen sensitivity and sustained wafer-material owner earnings.

Oct 26, 2026, 12:11 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

Current filed Q3 2026 results strengthen the operating case: revenue was $1.991B, operating income $225M, net income attributable to Amkor $190.613M, and diluted EPS $0.76. Record Computing and Automotive/Industrial revenue and margin expansion support demand, while nine-month PP&E payments of $1.237B, capex payables of $496M, and total debt of $2.5B make cash returns and financing the load-bearing questions. The 11.58x normalized P/E model and its terminal-price gap predate Q3 accounting and are not expected returns. An edge requires sustained common-owner cash flow after capex and dilution, plus a demonstrated return on capacity.

Nov 2, 2026, 4:46 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Fact: the accepted model index contains no normalized scenario for this security; the latest supplied primary period is H1 FY2026 (Jan 1 - Jun 30, 2026), latest filed period; latest full year FY2025 (ended 2025-12-31). Judgment: Low revenue multiple but loss-making; a recovery hypothesis, not cheap earnings. Evidence that could change the view: 300mm utilization, inventory, contract prices, depreciation and net debt; no substitute ticker or model is used.

Oct 17, 2026, 12:08 AM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1

Eaton’s November3 Q3 release reports $8.947bn sales, up28%, including19% organic growth. Owner-attributable income was $1.078bn on389.6m diluted shares, with GAAP EPS $2.77 versus adjusted $3.68. The adjustment includes intangible amortization and acquisition/divestiture charges. Electrical backlog grew46%; backlog is not cash. Reported operating cash flow was $1.31bn and FCF $1.03bn. FY2026 EPS guidance is $10.39–10.49 GAAP and $13.70–13.80 adjusted. At $391.70, midpoint guidance multiples are37.5x GAAP and28.5x adjusted; these are guidance-based measures, not verified future earnings. Demand and cash support retaining the20% target. Actual weight drifted to20.83% after the price move; this review adds no shares. Concentration and execution risk argue against increasing exposure.

Nov 3, 2026, 4:47 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Research view: Keep Vertiv at its current small weight because demand and guidance improved, while order cancellation, backlog realization and working-capital risks remain material. The Oct 28 release reports organic orders growth above 100% and raised guidance. Require sustained audited conversion and cash generation.

Nov 3, 2026, 12:15 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

Exact reconstruction matches the accepted old fingerprint. The only changed typed financial field is the FY2025 auditor identification, now Shinhan Accounting Corp., partner Dongwoon Lee, dated March 17, 2026. All five source documents and their hashes are unchanged. This explains the fingerprint difference but does not complete a current owner-earnings or valuation review. Keep the existing hold; no baseline reset or model admission.

Oct 19, 2026, 4:47 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Exact reconstruction matches the accepted old fingerprint. No current financial field changed. Adding the retained FY2024 annual report changed the document payload from three to four records. Historical reported profit and EPS remain outside current fields; common-owner net income and diluted EPS remain unqualified. Keep the existing hold pending owner/period qualification; no baseline reset or model admission.

Oct 19, 2026, 4:47 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Parent reviewed the retained MOPS H1/Q2 2026 statement, in TWD thousands. It explicitly reports Q2 diluted EPS TWD4.59 and H1 diluted EPS TWD8.48, with owner income 9,418,316 and 17,419,139 respectively. The earlier EPS extraction gap is narrowed, but the exact weighted-average diluted shares and a sustainable owner-earnings model remain unqualified. Do not infer an exact denominator from rounded EPS or annualize one quarter. No canonical financial field or model changed.

Oct 20, 2026, 4:39 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The October 12 unaudited release supports strong Q3 memory earnings, with reported EPS TWD25.40 on 3,450m weighted-average shares. This is quarterly cycle evidence, not sustainable owner earnings. The nine-month denominator differs. Keep Nanya as an alternative to investigate; no scenario or simulated allocation is admitted. Planned future fab capacity is not current qualified production.

Oct 20, 2026, 4:39 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The July 31 issuer-filed Q1 FY2026 report resolves the narrow capex extraction gap, but its exceptional memory-cycle profit and cash generation do not establish sustainable common-owner earnings. The 3-for-1 split effective October 1 also requires a consistent post-split share and per-share history. No current price judgment or model admission is supported until the cycle-normalized earnings bridge and split-adjusted denominator are reviewed.

Oct 21, 2026, 12:16 AM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1

The current pass did not retrieve the DART cash-flow report needed to reconcile owner income, CFO, capex, debt and dilution. A failed issuer-page retrieval is not evidence that the filing or data is absent. Keep the company unmodeled and unvalued until the primary DART statement supports a common-owner earnings bridge.

Oct 21, 2026, 12:16 AM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1

The source-change cause is the reviewed October21 EPS/owner extraction correction, with original document hashes unchanged. H1 basic/diluted EPS12.31/12.19 remain supported. Exact weighted diluted shares and A/H timing are still unresolved; keep the source-change flag and no model admission.

Oct 22, 2026, 4:36 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The H1 statement supports CNY8.682bn filed PRC operating profit and CNY7.529bn owner profit after tax and minority claims. The operating subtotal includes financing and investment items, so it is not a clean US EBIT proxy. The extraction gap is narrowed; sustainable margins, cash conversion, debt and dilution remain unqualified. No scenario or allocation admitted.

Oct 22, 2026, 4:36 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The retained exchange profile reports cumulative H1 EPS44.54 and pretax income TWD25.914bn. It does not qualify diluted EPS, owner net income or cash flow. Treat AVC as a cooling alternative to investigate, without annualizing a strong half-year or multiplying EPS by period-end issued shares.

Oct 22, 2026, 4:36 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Electronics Q1 FY2027 revenue JPY127.7bn and OI JPY31.1bn establish material semiconductor/data-storage exposure, but group cash generation and unrelated-business losses matter. The transcript gives JPY10bn tariff-refund support in the presentation and about JPY1bn net of customer refunds in Q&A; scope needs reconciliation. Do not assign an AI premium or normalize from the electronics segment alone.

Oct 25, 2026, 12:47 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The FY2025 annual report reports owner profit rounded to NT$6.791bn, separate from group profitNT$10.605bn. OCI is a separate loss ofNT$0.545bn. The corrected thousand-unit fields do not establish diluted shares or sustainable earnings. Keep the company under source review and outside the qualified scenario list until dilution and owner-cash checks pass.

Oct 25, 2026, 12:47 AM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

The H1 2026 filing now provides an exact common-owner and diluted-EPS bridge: parent profitNT$5.221bn, diluted numeratorNT$5.280bn after convertible-interest adjustment, and1.930362bn diluted shares. Carrying-value borrowings areNT$36.670bn before leases, including previously omitted bank loans. Group CFO17.191bn trails PP&E purchases38.448bn. This resolves extraction gaps but not sustainable owner earnings or the return on new fabs; no portfolio allocation or scenario admission follows.

Oct 25, 2026, 4:53 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

Keep the 15% target: established licensing and silicon support the current business, while HBC validation is early engineering evidence and AI250 sampling remains prospective. The edge case requires customer deployment, separately evidenced AI/data-center revenue, durable licensing economics and a price that compensates for handset concentration.

Nov 3, 2026, 4:47 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2Source 3

Q3 2026 supports demand strength: revenue grew 25%, Cloud 103%, Search 12%, and operating margin was 31.8%. Common-owner income was $26.170B after $298M of preferred dividends, versus $34.979B a year earlier; other income/expense moved from +$12.759B to -$8.221B. Nine-month PP&E purchases were $131.953B versus $63.596B a year earlier. The platform remains high quality, but the model’s 29.72x normalized P/E is based on pre-Q3 assumptions and needs review. An edge requires AI monetization and Cloud cash margins to outgrow capex, depreciation, and regulatory costs. Cloud growth includes meaningful TPU systems sales; it is not a pure recurring-service growth rate. Nine-month operating cash flow of $127.149B was $4.804B below PP&E purchases of $131.953B, before acquisitions and shareholder distributions. Investment timing can depress this measure, so it is not a sustainable annual cash forecast.

Nov 2, 2026, 4:46 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1Source 2

At HKD392.20 on 2026-11-02, the current compact scenario screens at 13.76x normalized earnings with a +1.6% discounted three-year base terminal-price gap and low-medium confidence; the gap is not expected return. Q2 2026 unaudited revenue was RMB204.8B (+11%), marketing services grew 22%, and cloud-related Business Services grew. IFRS owner-attributable profit was RMB56.0B (+0.7%), versus non-IFRS RMB68.4B (+9%). Reported quarterly FCF was negative RMB13.8B after capex, content and leases; management’s RMB37.6B pre-compute-prepayment measure is not reported FCF. A quality platform and low scenario multiple merit follow-up, but an edge needs several quarters of common-owner FCF after AI compute investment, with valuation refreshed to match filed cash earnings.

Nov 2, 2026, 4:46 PM PT · Evidence assessment: medium · Separate from scenario confidence and investment conviction.

Source 1

At ¥511.5 on 2026-11-02, the model screens at 14.73x normalized earnings, but this is a diversified chemical company, not a pure AI-material supplier. FY2026 owner net income was ¥60.9B and core operating income ¥208.4B; Q1 FY2027 owner net income was ¥40.8B and core operating income ¥62.3B. A sustained recovery, capital spending and segment cash conversion remain unresolved; do not treat a low modeled multiple or recent quarterly rebound as proof of a durable edge.

Nov 2, 2026, 4:46 PM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1Source 2

Quanta H1 revenue was TWD1.846trn, parent profit TWD49.844bn and basic EPS TWD12.93. Operating margin was 3.10%, down from 4.55% a year earlier. Operating cash flow was TWD45.929bn and cash PP&E purchases TWD21.595bn, leaving TWD24.333bn before other investing flows, leases and distributions. Total investing outflow of TWD56.269bn includes financial investments and is not cash capex. Receivables and inventory each absorbed over TWD270bn, largely offset by trade payables. Profitable growth and positive cash after PP&E favor Quanta; margin compression, supplier financing, customer concentration and Taiwan exposure argue for a measured position decision. At TWD311.50 on November3, retain as a comparison candidate rather than add a position. Terminal-price scenarios cannot establish risk compensation.

Nov 3, 2026, 4:47 PM PT · Evidence assessment: low · Separate from scenario confidence and investment conviction.

Source 1Source 2Source 3Source 4