Variant Perception

Figures converted from CNY at historical FX rates — see data/company.json.fx_rates. Ratios, margins, and multiples are unitless and unchanged.

Where We Disagree With the Market

Our sharpest disagreement: the market is pricing Yadea's $2.47B cash pile as permanently stranded and the FY2025 19.1% gross margin as a subsidy-driven peak — together producing a 10.2× headline P/E and a 2.8× operating EV/EBITDA. The report's evidence says the cash discount is already inflecting (FY25 payout +245% YoY with a special, capex collapsed to 2% of revenue, no acquisition runway) and the gross-margin reset is being misread because the post-Nanjing lithium rule is a regulatory ratchet that holds the BoM floor at sub-scale OEMs, not at Yadea or AIMA. The market believes Yadea is a cyclical commodity OEM whose FY2025 print was policy-aided and whose family controllers will not surface the cash; both legs of that belief are testable. The single observable that would resolve the debate over the underwriting horizon is the FY2026-27 capital-return cadence — a second special dividend, a structural buyback authorization of at least $143M, or a stated payout floor would compress the discount toward AIMA's 7.8× EV/EBITDA. If instead the FY26 final dividend reverts to US$0.068 with no special and buybacks stay nominal, the bear's "permanent yield-trap" call is correct and our variant view dies.

Variant Perception Scorecard

A 62/100 variant-strength reflects an edge that is real but not extreme: the operating-business multiple gap to AIMA is unambiguous, the FY25 capital-return inflection is observable, and the regulatory-ratchet thesis on gross margin is grounded in the post-Nanjing lithium rule plus the 2024-25 MIIT Standard Conditions. The 72/100 consensus-clarity score reflects a tape that is doing the work for us — US$1.44 at the 14th percentile of the 52-week range, a fresh death cross on 13-Nov-2025, and SFC reportable shorts that rebuilt +25.7% over seven months tell us where the market sits even though only two sell-side brokers publish. The 65/100 evidence-strength score reflects that the forensic page legitimately flags cash-flow-quality questions and the governance overhang is real; the variant view is calibrated, not maximalist. Resolution is multi-print — 12 to 24 months — because no single H1 2026 disclosure proves capital-return regime change.

Consensus Map

The five-issue map is bifurcated. On issues #1 (cash discount) and #4 (CFO quality), the tape consensus is harder to read than the published-broker consensus because Jefferies (PT US$2.30) and the analyst-screen aggregators (StockAnalysis Strong Buy, PT US$2.23) imply +55% upside that the tape ignores. Where the tape and broker consensus agree — issues #2 (margin reversion) and #3 (Ninebot share threat) — the variant view has the cleanest disagreement.

The Disagreement Ledger

Disagreement #1 — the cash discount. Consensus reads the operating business as a fair-multiple cyclical OEM and prices the $2.47B cash pile as if a 63% Cayman trust will never let it out. Our evidence is that the family has already moved at the margin — capex collapsed from $214M to $108M, ROE printed 27.8% with no incremental reinvestment runway in China, and the FY25 declared distributions of US$0.097 are +245% versus FY24's US$0.028. If the variant view is right, the market has to concede that AIMA's 7.8× EV/EBITDA is the right anchor for Yadea's operating business and that the cash gets distributed rather than discounted; the cleanest disconfirming signal would be a flat US$0.068 FY26 final dividend with no special and a buyback authorisation under $14M.

Disagreement #2 — the gross-margin regulatory ratchet. Consensus prices the FY25 19.1% gross margin as a subsidy-driven peak that reverts to FY24's 15.2% or the FY22-23 average of ~17%. Our evidence is that the post-Nanjing lithium rule and the 2024-25 MIIT Standard Conditions impose a fixed-cost minimum (R&D >=2% of sales, tooling >=$143k, mandatory pack certification) that mathematically removes the sub-scale OEMs whose discount pricing has historically capped Yadea's GM. If the variant view is right, the market must concede that mid-cycle GM lifts to 16-18% and normalised earnings to ~$286M (not the bear's $200M anchor). The cleanest disconfirming signal would be an H1 2026 GM below 17% with AIMA matching at the same level — both confirming that price competition didn't actually withdraw with the long tail.

Disagreement #3 — the Ninebot defender narrative. Consensus, anchored by the 36kr and Yicai post-H1 2025 framing, reads Yadea's "performance adjustment" against Ninebot's +102% E2W revenue growth as share loss to a Xiaomi-ecosystem rival. Our evidence is that Yadea FY25 units grew +25% to 16.3M (above the FY23 peak), Ninebot's H1 2025 E2W base of 2.39M is ~9% share versus Yadea's ~26%, and the H1 2025 China industry grew +29.5% — both companies are riding the same regulatory consolidation wave from opposite ends of the price spectrum. If the variant view is right, the multiple gap between Ninebot (21.3× P/E) and Yadea (10.2× P/E) overstates the threat to Yadea's mass-market position. The cleanest disconfirming signal would be Ninebot E2W units approaching Yadea's pace in H1 2026 (>=40% growth on an already-large base) combined with a Yadea ASP stall or decline.

Disagreement #4 — the horizon mismatch. Consensus and our own catalysts page concentrate decision weight on the August 2026 H1 interim print as the single resolving event for cash-flow quality, gross margin, and capital-return cadence. Our evidence — drawn from the long-term thesis page — is that the durable underwriting variable (capital-return regime) resolves over three signal events: the 17-Jun-2026 AGM mandate vote, the H1 2026 interim dividend declaration (Aug 2026), and the FY26 final dividend declaration (Mar 2027). If the variant view is right, a PM who waits for August 2026 alone is buying optionality on a multi-print sequence; the August print partially answers the question but cannot fully resolve it. The cleanest disconfirming signal would be a clean H1 2026 print where bills-payable contracts, GM holds at 19%+, and the family declares no interim special — proving that one print does in fact resolve both the cash-quality and the regime-change questions in a single document.

Evidence That Changes the Odds

How This Gets Resolved

The eight signals cluster into two windows. Signals #1, #4, and #5 resolve within 90 days and stress-test the H1 26 cash-flow quality question that the bears prioritise. Signals #2, #3, #6, and #7 resolve over 12-24 months and stress-test the durable capital-return and competitive variables that the variant view actually depends on. Signal #8 (SFC SPR direction) is the implementation-side tell — it does not update the thesis but it telegraphs whether the institutional book is reading the same evidence the way we do.

What Would Make Us Wrong

The cleanest refutation of the variant view is governance, not operations. A third Big-Four auditor exit, an expanded Deloitte Key Audit Matter on revenue cut-off or payables classification, or a Cayman trust restructuring that telegraphs a partial-sale path would crystallise the worst-decile ISS QualityScore as the correct read on this name. The structural argument — that the family will surface the cash because there is no reinvestment runway — depends on the family choosing the path that makes economic sense. If governance fails, the family is revealed to have chosen optionality over economics, and the cash discount becomes permanent regardless of capex intensity, ROE, or capital-allocation math. The forensic page's elevated 48/100 risk score is the institutional-grade flag for this scenario; we should not pretend it doesn't exist.

The second-cleanest refutation is competitive. The Ninebot share-loss narrative looks weak from H1 2025 data alone, but the test runs for several more years. If Ninebot's E2W units exceed 4M in H1 2026 (growth holding at 60%+) and Yadea's ASP stalls or contracts, the trade-press framing was correct and the moat is narrowing at the premium end. The variant disagreement on this would die. Equally, if AIMA prints H1 2026 operating margin at or above Yadea — repeating the FY24 outperformance pattern in a non-trough year — then Yadea's claimed "scale + integration = better economics" thesis is empirically falsified and the AIMA-discount peer-rerate math collapses. Both of these are observable in 12 months.

The third refutation is the one the forensic and short-interest pages already point to. If H1 2026 shows bills payable expanding past $1.57B with the supplier-finance book growing past $286M, and CFO ex-Δ-bills tracks below 0.8× of NI, then the institutional short positioning is correct that FY25 CFO was bank-mediated and the durable FCF base is materially below what the variant view requires. The bull case for cash return depends on a sustainable $570-714M mid-cycle FCF; if the FY25 print was working-capital release plus subsidy-peak earnings plus supplier-finance window-dressing all stacked, then the FY26 trough hits before the FY27 recovery and the family delays cash return as a defensive measure. This scenario is internally consistent with everything the bears have written; the report's lean-long verdict assumes it doesn't happen.

The fourth and quietest refutation is what we get wrong if all three of the above are mild: the published consensus (Jefferies Buy US$2.30, ValueInvesting screens US$2.94-3.70, StockAnalysis Strong Buy US$2.23) could already be correct, the tape is wrong, and the variant view is just a slower version of consensus. In that scenario the August 2026 print prints clean, the family does declare a second special, and the H1 2027 stock prints US$1.92-2.17 mechanically — but the variant view added no edge because the consensus call was right all along. The honest read of this risk is that disagreement #4 (horizon mismatch) is partly a hedge for the case where the consensus is right; if it is, the right action is to sell into the print rather than wait for FY26 final.

The first thing to watch is the 17-Jun-2026 AGM resolutions for any buyback authorisation language and any chairman-statement payout-floor commitment — that is the single highest-leverage signal in the next 18 days and the one that most cleanly distinguishes the variant view from the bear's permanent yield-trap call.