〔Overnight Brief〕All-A Two-Non Q2 Net Profit +19.3% YoY on AI-Resource Dual Engine, Yet US Risk Spillover and a 96% Home-Ownership Floor Frame a Two-Lane K-Shape - Earnings Beat, Inclusive-Widen Test

0. Weekly Arc

The research tape converges on a K-shape. The 2026H1/Q2 earnings print confirmed the AI-resource dual engine: All-A two-non net profit +17.3%/+19.3% YoY in H1/Q2, with mid-cap growth, cyclicals, and financials leading ^[1]. Yet two K-shape risks surfaced the same week — US-led cross-border risk transmission lifting tech crowding ^[2], and a real-estate governance question on whether property is the more inclusive wealth vehicle ^[3]. The mechanism in both: aggregate beats mask a two-lane distribution.

1. Earnings Mechanism

• **[NEW] Dual-engine beat (券商研报索引 vReport 宏观+策略, 2026-09-05):** 2026Q2 All-A two-non net profit/revenue YoY +19.3%/+6.2%; H1 +17.3%/+6.2%; net margin improvement drove the gap between profit and revenue growth ^[1]. ChiNext/STAR led, main board significantly improved, consumer under pressure, inventory turnover slowdown flagged as a hidden concern ^[1].
• **[NEW] Style structure (vReport 宏观+策略):** mid-cap and large-cap beat small-cap; growth and cyclical styles led; financial growth significantly improved ^[1]. AI internal: upstream pricing power rising; AI + energy up paired with domestic-demand + midstream pressure ^[1]. The report explicitly warns that profit-expansion breadth may cap the total-prosperity improvement height ^[1].

2. Cross-Border Risk Transmission

• **[NEW] US as spillover source (CGS-NDI, 2026-09-05):** since the 2025 geopolitical-narrative shift, risk transmission between representative global stock markets has strengthened, with US markets the dominant risk spillover source; some Asian markets saw significantly increased external risk input ^[2]. Since July's global AI correction, A-shares' overall risk import has receded from 2025 highs but remains elevated in tech ^[2].
• **[NEW] Tech vs non-tech K-shape (CGS-NDI):** non-tech risk import is gradually declining, tech external risk input is relatively high; Japan and Korea markets amplify the tech-channel shock ^[2]. The report cites four major drivers behind the linkage upgrade but does not enumerate them in the packet — flagged in source quality ^[2].

3. Real-Estate Governance and K-Type

• **[NEW] Inclusive-wealth angle (vReport 宏观+策略, 投石问K series seven, 2026-09-05):** the 2019 PBOC survey shows 96% urban household home ownership versus a materially lower equity/equity-fund holding rate; the top 10% hold a much larger share of equity market cap than real estate ^[3].
• **[NEW] Construction chain labor absorption (vReport 宏观+策略):** the construction industry's profit-vs-employment concentration gap is significantly smaller than manufacturing's, supporting the claim that construction demand expansion can better improve employment and primary-distribution share ^[3].
• **[NEW] Nordic precedent (vReport 宏观+策略):** Finland, Sweden, and Norway all experienced 20%+ housing price declines and simultaneous banking shocks around 1990, then adopted fast, transparent clearing policies with strict conditionality ^[3].

4. What Would Falsify It

• Whether the Q2 net-profit beat extends into H2 without a margin reversal once AI capex and resource-supply tailwinds normalize ^[1].
• Whether US risk spillover widens from tech into non-tech sectors as the global AI correction deepens ^[2].
• Whether the inclusive-real-estate thesis holds under a Nordic-style 20%+ price decline and banking-shock scenario ^[3].

5. Source Quality Control

• Items ^[2] four-major-drivers mechanism and ^[3] Nordic policy granularity are single-source, drawn only from the named publisher's research report; the K-shape label is publisher-specific framing, not a consensus read ^[2][3]. ^[1] runs under the 券商研报索引 vReport 宏观+策略 board; ^[2] under CGS-NDI; both dated 2026-09-05, same packet window.