Additions 2026-09-23 — 1913.HK (PRADA S.p.A.), industry stages, 2026-09-23

Additions 2026-09-23 — 1913.HK (PRADA S.p.A.), industry stages, 2026-09-23

Slot cluster-additions-2026-09-23, member 1913.HK only. Split into its own file as the slot note allows ("Split into per-name cluster files if that is cleaner"). Run by hand in a session on the owner's request, not by the routine. HHH is not scored here and stays with the slot. Model-drafted; nothing here is human-verified. Every figure carries the filing it came from; [background] marks background knowledge and [search-summary] marks anything that came from a search result.

Stages: S0 (lens), S4, S6, S7. Instrument: the standard AI-surplus test (SKILL.md §4.1), as the slot note directs for luxury. Reporting currency is EUR. The listing trades in HKD. Every conversion uses EUR/HKD 8.9789 and USD/HKD 7.84335, the 2026-09-22 closes from ROIC.ai FX:EURHKD and FX:USDHKD.

---

§0 — Sources and how quotes were taken

Rung 3 (eu_data, company id 50262) answered everything filing-based:

• the 2025 Annual Report (filing 60295260, consolidated, 426,658 chars, retrieved in full);
• the FY2025 results announcement (32903728, 2026-03-05);
• the H1 2026 results announcement (49775083, 2026-07-30);
• the Versace completion announcement (30119032, 2025-12-02);
• the connected-transaction announcement of 2026-05-19 (46532170);
• the Q1 2026 and Q2 2026 call transcripts (56090428, 56123411).

ROIC.ai HKEX:1913 supplied prices, and three WebSearch queries supplied scuttlebutt. There are no claim rows on the board for Prada, Miu Miu, Versace or any luxury peer.

Each filing's markdown was saved to text. Every quoted span below was located in that text by exact substring match before it was written. Transcript content is paraphrased and never put in quotation marks: the eu_data transcripts are machine speech-to-text, with names garbled ("Mew Miu", "Peter Moulier").

---

§1 — The finding: the lens scores Prada 3/4, and none of the four tests measures the thing that decides it

Under the account skill's wording, Prada scores 3/4. It owns its channel, its store ledger is a physically generated transaction record, and 24 owned factories plus a façon (subcontract) network are a bottleneck that tokens cannot make. It fails liability absorption, because it sells objects and not outcomes. That is a pass on paper. What decides a luxury house is desirability, and desirability is a cycle per brand, not a structure. None of the four tests reads it.

The filings show that cycle turning inside one group in eighteen months:

Sources: FY2025 results announcement, Note 1; H1 2026 results announcement, Note 1 (the Q2 columns); Q1 2026 figures from the CFO's prepared remarks (paraphrased).

So the brand that carried 2025 has stalled, and the brand that lagged is carrying 2026. The margin bill for the third brand arrived at the same time: EBIT Adjusted 22.6% → 17.4% of net revenues, H1 2025 → H1 2026 (H1 2026 announcement, P&L). The CFO attributed about two-thirds of roughly 500bp of dilution to Versace and one-third to FX (Q2 call, paraphrased). Organic margin was described as steady. The instrument does not see any of this, and it scores the same 3/4 in a year when Miu Miu grew 35% as in a year when it grew 2.5%. A brand-cycle variant would need a test for brand heat (full-price like-for-like growth against the brand's own base) and a test for creative-succession risk. That is recorded here for the owner, not improvised as a score.

---

§2 — 1913.HK · luxury · ✅ lens applied — 3/4

Instrument: standard AI-surplus test, account-skill wording (SKILL.md §4.1).

Reference category: none. The account skill's Tier 1 list covers specialty insurance, industrial OEMs, defence, exchanges, payment networks and banks, and aggregates/waste/water/pipelines. It has no consumer-brand or luxury entry, so no category sets the expected score here.

Surplus question — if every competitor gets the same AI cost reduction, who keeps it? The brand does. Luxury prices are not set at cost-plus, and the pricing power sits at the brand's own till:

• Gross margin was 80.3% in FY2025 and 78.3% in H1 2026. The latter was diluted by Versace, and the CFO said it was slightly up excluding Versace (Q2 call, paraphrased).
• Q2 2026 Prada growth came from mix, not price. The CEO said volume was slightly positive and the rest was mix rather than a price effect (Q2 call, paraphrased).
• The company is pushing up the price range. It is aiming at top spenders with higher-priced pieces while holding entry prices stable (same call).

Design, clienteling, merchandising and supply-chain planning are where cheap intelligence lowers cost, and none of that saving is passed to a buyer who does not know the cost base. The limit on the surplus is not AI. It is the aspirational buyer's willingness to keep paying, which Bernstein raised directly on the Q2 call.

Falsifier — one number. Combined Prada + Miu Miu retail organic growth below zero for two consecutive quarters. It was +3.1% for the group's retail organic in H1 2026 and +5.1% in Q2 (H1 2026 announcement, Note 1). Pricing power in luxury is desirability, and desirability shows up in full-price like-for-like sales before it shows up anywhere else.

Liquidity.

• ADV HKD 57.0m/day (~USD 7.3m; median HKD 31.6m). This is 21 sessions, 2026-08-25 → 2026-09-22, close × volume, from ROIC HKEX:1913, with an average of 1.45m shares a day.
• Last price HKD 37.38 (2026-09-22) = €4.16.
• Market cap HKD 95.6bn = €10.65bn on 2,558,824,000 shares.
• Free float is ~20% (Prada Holding S.p.A. holds 79.98%, H1 2026 announcement), so the floating capitalisation is about HKD 19bn.
• ⚠️ ROIC's volume series was not cross-checked against HKEX. If it under-counts, the ADV is understated. Either way, this is a thin line for a €10bn company, and S8 sizes against it.

Re-tag check (§4.1). luxury holds one member, so there is no mis-clustering to fix. If the owner adds Kering, LVMH, Hermès, Moncler or Cucinelli, the brand-cycle variant proposed in §1 would be worth writing before the second member is scored.

---

§3 — S4 Industry & supply map

The chain, in words.

1. Raw materials. Hides, textiles and metals come from a wide supplier base. The annual report says of it: "The Group has a wide range of raw materials suppliers and external manufacturers." (2025 Annual Report, Directors' Report, relationship with suppliers). About 93% of those suppliers are in the EU, most in Italy (same passage). Prada took a 10% stake in the tanner Rino Mastrotto in 2025 (€29.9m, FY2025 announcement), and the related-party table lists the tannery Conceria Superior and several knitwear and footwear workshops as suppliers.
2. Manufacturing. The 24 owned factories do prototypes, sampling and hide cutting. Façon (subcontract) workshops do the volume assembly.
3. Brand and design. Miuccia Prada with Raf Simons for Prada, Miuccia Prada for Miu Miu, and Pieter Mulier at Versace from 2026-07-01.
4. Distribution.
5. Directly operated stores and e-commerce: 86.4% of revenue.
6. Wholesale: 9.8%, kept selective (CFO), including duty-free and department stores.
7. Royalties: 3.8%, from licences. Eyewear is with Luxottica (2025 Annual Report, history section). Miu Miu beauty is with L'Oréal (same report). Royalties grew +73.5% in H1 2026, helped by Versace's licences.
8. The customer. The CEO's own split is top spenders, who he said are healthy and wealthy, and where he sees the biggest opportunity, against younger and aspirational buyers the industry is failing to attract (Q2 call, closing remarks, paraphrased).

Where the pricing power is. At the brand, exercised at its own retail till. Gross margin of 80% on a vertically controlled chain means the value is captured after manufacturing.

• Suppliers. They are price-takers. The annual report says no customer or supplier has influence: "the Directors do not consider any customer or supplier to have an influence on the Group" (2025 Annual Report, Major customers and suppliers). The five largest customers and suppliers are each below 30% of sales or purchases (same passage).
• Landlords. They are the one counterparty with real leverage. The prime flagships cannot be moved, and €240.5m of the group's lease liabilities at 2025-12-31 were owed to related parties. These are entities of Prada Holding (PH-RE Llc in Tokyo, 720 Fifth USA Llc in New York, Prada RE UK Ltd in London) or other companies controlled by the founders (2025 Annual Report, Note 39 table). The same families that own 80% of the equity own several of the buildings the brand must be in.

Concentration. No customer concentration. The concentrations that matter are:

1. Brand. Prada was 61.6% and Miu Miu 29.0% of H1 2026 retail net sales (H1 2026 announcement, Note 1).
2. Region. Asia Pacific was 35.0%, Europe 28.6%, Americas 21.7%, Japan 11.0% and the Middle East 3.7% of retail.
3. Creative. Two founders aged 77 and 79 (2025 Annual Report, directors' biographies).

The façon layer is the exposed link. Milan prosecutors asked 13 luxury brands, Prada and Versace among them, to hand over supply-chain governance documents on 2025-12-04, in a labour- exploitation investigation of subcontractors [search-summary]. Other houses (Loro Piana among them) were earlier placed under temporary judicial administration [search-summary]. The annual report, the FY2025 announcement and the H1 2026 announcement were searched for "prosecut", "investigat", "judicial" and "subcontract" and contain no mention of it. The annual report's own risk language on the façon layer is quoted in the dossier's S3.

Who wins share, from whom.

• Prada is gaining. The CEO said the Prada brand has grown above the market for five years on a like-for-like basis (Q2 call, paraphrased), and he named the US and China as the two regions where it is still below its fair share (Q1 call, paraphrased).
• Miu Miu is normalising, but was not a store-count story. Its five years of growth came with few new stores (Q2 call, paraphrased: roughly 155 → 165).
• Who is losing is not established from Prada's filings. The share donors at the aspirational end are commonly said to be Kering's brands [background].

---

§4 — S6 Scuttlebutt

Outside evidence 1 — sell-side, [search-summary].

• Morgan Stanley downgraded Prada to Equal-weight on fashion-trend risk: a possible swing from minimalist to maximalist aesthetics, and renewed creative energy at competing houses.
• UBS downgraded to Neutral with a HK$50 target, down from HK$74. It cited early signs of maximalism and competition from Dior and Chanel under new creative directors.
• Both notes date from late 2025 per the search summary. The articles were not read.
• The consensus cited in trade press has 2026 group operating profit at about €1.37bn, described as broadly unchanged on the year.

Outside evidence 2 — regulatory, [search-summary]. Italy's labour-exploitation investigation of luxury supply chains covers Prada and Versace (2025-12-04 document request). Police raided other brands' offices in July 2026. None of this is in Prada's filings (§3).

Outside evidence 3 — the analysts' own questions on the two calls, which say what the street is worried about:

1. whether Miu Miu's normalisation becomes fading (Bernstein, Q2);
2. the aspirational middle-class buyer being priced out (Bernstein, Q2);
3. Versace losses and the store count (Citi, BofA, Kepler, Q2);
4. capital return given the discount to peers (Kepler, Q2);
5. a Milan dual listing (Morgan Stanley, Q2);
6. the traveller share of sales (Intesa Sanpaolo and DBS, Q1).

Short interest: not reached. HKEX short-selling turnover was not fetched this run; no tool on the ladder serves it. No short report found in three searches, and a failed search is not evidence that there is none.

Board claim rows: none. ledger/pipeline/gold/ holds no row mentioning Prada, Miu Miu, Versace, Kering, LVMH, Hermès or "luxury".

Product test / expert call: ⏳ OPEN (user). For Prada the useful checks are:

• an in-store visit to a Miu Miu and a Prada flagship, looking at full-price sell-through and whether the new Miu Miu leather-goods line is on the floor;
• an expert call with a former Italian façon workshop owner on audit practice after the Milan probe.

---

§5 — S7 Thesis and framework test

Framework used: a plain brand-cycle / adoption read, per brand, each on its own desirability S-curve. The _clusters/ai-capex-2026-09-20.md lens table is for AI and semis and does not apply. Bear case first.

Bear case.

• Earnings are going down, not up.
• EBIT Adjusted fell −14.3% in H1 2026 (€618.5m → €530.2m).
• Net profit to owners fell −15.3%.
• EPS went €0.151 → €0.128.
• Miu Miu has normalised, perhaps permanently. It went from +34.8% in FY2025 to +2.5% in H1
• The CEO now speaks of 5–10% as its long-term range, not the double digits of early 2026 (Q2 call, paraphrased).
• Versace is a multi-year cost.
• Management guides to operating losses in 2026 "of not dissimilar magnitude" to 2025's (FY2025 results announcement, Versace section).
• It also guides to a top-line contraction in 2026.
• Mulier's first collection launches in May 2027, and it will not be in every store until late 2028 (Q2 call, Lorenzo Bertelli, paraphrased).
• About 350bp of group margin dilution is guided for the full year (Q2 call, CFO, paraphrased).
• The balance sheet has been used. It went from a €599.6m net cash position at end-2024 to a €692.7m net financial deficit at 2026-06-30, excluding €3.1bn of leases.
• Governance cuts one way. The company is 80% controlled; the founders are 77 and 79; and there is a steady flow of real-estate transactions with family vehicles (the scorecard's S5).
• Management declined both value unlocks the street asked about in Q2: a buyback and a Milan listing.
• Fashion may turn against the house's aesthetic (Morgan Stanley and UBS, [search-summary]).
• The unfiled overhang: the Milan labour probe names both Prada and Versace.

Bull case.

• The Prada brand itself, 61.6% of retail, is re-accelerating on like-for-like, full-price sales: +0.4% in Q1 → +6.3% in Q2 2026. The Chinese cluster was up double digits in Q2 (CFO, paraphrased).
• Growth has been continuous. It is the 22nd consecutive quarter of organic growth at group level (Q2 call, CFO, paraphrased).
• The organic business is holding its margin (steady organic EBIT Adjusted margin, H1 2026 announcement). The dilution is purchased, and the purchase was small: Versace was bought for USD 1,395m (FY2025 announcement) on €684m of 2025 net revenues. That is about 1.7x sales for a globally known name.
• The debt is modest. Net financial debt is 0.43x LTM pre-lease EBITDA (1.80x including leases), and the bridge was refinanced with a €300m, 10-year US private placement.
• The price is low. 13.4x LTM EPS, 9.2x EV/EBIT Adjusted (excluding leases) and a 4.0% dividend yield at HKD 37.38. The sector's peers have long traded at higher earnings multiples [background], and Kepler said on the Q2 call that the stock trades at one of the lowest multiples in the sector.

Variant perception. The market prices 1913.HK on its two noisy lines: Miu Miu decelerating and Versace diluting, both visible in a 23% fall in H1 EBIT Adjusted margin. The variant is that both are time-bound and already in the numbers:

1. Miu Miu has moved from hypergrowth to a normal brand. It has not faded: its 2026 growth is positive on a +40% comparable, with few added stores.
2. Versace's loss is guided flat on 2025, and its first real test is not until 2027–28.
3. The core Prada brand, which the market spent 2025 worrying about, is growing faster than Miu Miu for the first time in years.

If that is right, H1 2026's 17.4% is the trough margin. FY2027–28 earnings recover as Versace's loss narrows, on a flat share count, and the multiple has room to move back toward the sector's. If it is wrong, the reason will be desirability, and it will show first in Prada's full-price like-for-like sales.

Kill criteria (specific, testable):

• Prada-brand retail sales below 0% at constant FX in any reported half or quarter. H1 2026 was +3.3% and Q2 +6.3% (H1 2026 announcement, Note 1).
• Miu Miu retail sales negative at constant FX for two consecutive quarters. Q1 was +2.4% and Q2 +2.6%.
• Versace's FY2026 operating loss clearly larger than FY2025's. Management guided "not dissimilar", and the FY2026 annual report (window 2027-03-28 → 2027-04-23, eu_data) will state both years.
• Net financial deficit (excluding leases) above 1.0x LTM pre-lease EBITDA. It is 0.43x at 2026-06-30; a second debt-funded acquisition would be how it gets there.
• Any Prada or Versace entity placed under judicial administration in the Milan labour- exploitation investigation. It would be an HKEX inside-information announcement if material.

Time horizon. Two to three years, to FY2028. That is the first full year with Mulier's Versace in every store, per management's own timeline.

Conviction: ⏳ OPEN (user).

---

§6 — What this pass could not do

• Short interest and the HKEX register. Neither was reached. The Central Clearing and Settlement System (CCASS) participant breakdown and HKEX short-selling turnover are not on the source ladder.
• Q3 2026 trading. Not yet reported; the release is expected around late October on last year's cadence (see S9 in the scorecard).
• A reconciliation of Versace's H1 loss. Citi's arithmetic put it near €90m. The CFO said it was far less than half of that and took the question offline (Q2 call, paraphrased). No filing splits it.
• The eu_data normalized financials cannot be used for FY2025. The "FY2025" row is extracted from the parent-company Separate Annual Report (revenue €3,043m), not the consolidated one (€5,718m). All figures here come from the consolidated documents.
• §1's brand-cycle variant is a proposal only. It was not scored.