Technicals — Closed at HK$121.90, off 0.57% on 0.61x volume — classic pre-earnings standoff. Short/medium-term structure intact: holding above the 20-day (HK$118.76) and 50-day (HK$111.44), but the 200-day at HK$136.88 still looms ~12% overhead, the same ceiling that's capped this tape since early August. RSI at 56.4 and 7-day RSI at 51.4 are neutral; ADX 20.8 shows the trend coiling into the print. A 3.6% weekly give-back after a 10.5% monthly run is healthy digestion. Just 34% up the 52-week range, ATR HK$4.84 (~4%) — budget 2–3 ATRs of movement on Aug 20; an HK$10–15 intraday swing is on the table.
Fundamentals — A 20.2x P/E isn't demanding for a platform with an accelerating cloud business, but earnings quality is at its ugliest window: consensus sees June-quarter revenue of RMB 242.65bn, up just 2.6%, with profit down roughly 65% on instant-commerce subsidies — and the company has missed EPS four quarters running. The bull case lives in the cloud: ~RMB 33bn of cloud revenue last quarter, +26% YoY, with AI-related revenue compounding at triple digits and Goldman expecting AI to exceed half of external cloud revenue within a year. The bear case lives in negative free cash flow and commerce margins torched by the subsidy war. With subsidies rolling off under the regulator's anti-involution push since August, whether the loss-reduction inflection shows up in this print decides if 20.2x gets re-rated. The 0.84% dividend is a rounding error.
News — Today's -0.57% carried no new bad news — it's position-squaring ahead of the Aug 20 pre-market print. The real news sits at both ends of the battlefield: the instant-commerce war is officially cooling — Meituan, Ele.me and JD jointly pledged on Aug 1 to end predatory subsidies, and the market regulator's ten-point subsidy rulebook is out for comment — while Taobao Instant Commerce holds 80m average daily orders in August with 300m monthly buyers, lifting Taobao DAU 20%. Translation: share captured, cash burn regulated away — a clean tailwind for loss reduction. On AI, the 2.4-trillion-parameter Qwen3.8 launched Aug 3, with the Max version reportedly open-sourcing imminently. Into earnings, Morgan Stanley trimmed its ADR target to $180 from $190 but kept BABA a top pick expecting cloud acceleration to beat, and Goldman reiterated Conviction Buy.
Position call: hold through earnings. Rationale: a -0.57% drift on 0.61x volume, sitting on the 20-day at 118.76 — no one is running for the exits; the subsidy-rolloff-plus-cloud-acceleration script is a tailwind for this quarter's loss line, and 50.6% consensus upside from 33 analysts provides margin of safety. Entry zone anchors the 20-day up to spot; stop at HK$113 sits a step above the 50-day (111.44), ~1.8 ATRs below spot — a close through it on earnings day means take the loss and walk. First target HK$136.9 is the 200-day, the ceiling since early August; a clean print makes it the first stop. Eyes open: four straight misses means this is a stopped position, not a faith trade — and no fresh longs four sessions before the flip.
China's twin leader in AI compute and models: cloud growing 26% with AI revenue at triple digits, and Qwen firmly in the global top tier with an open-source ecosystem flywheel. The commerce base defended its traffic moat through the instant-commerce war (DAU +20%, 300m monthly buyers), and with regulators winding down the subsidy fight, profit leverage cuts the other way. At 34% of its 52-week range, 20.2x earnings, and an 11% discount to the 200-day, the stock still pays a penalty for the commerce baggage — the cloud/AI re-rate is far from done.
- Loss reduction in instant commerce disappoints; the subsidy war reignites under share pressure, and commerce margins keep bleeding
- Heavy AI capex keeps free cash flow negative; any cloud deceleration triggers a double de-rate
- Recurring China regulatory and geopolitical overhangs keep a persistent liquidity discount on the HK line
- 07-02LongHK$94.5 → closed HK$94.107-04-0.42%
- 07-08LongHK$99.6 → open → HK$121.9+22.39%
- 雪球 · Xueqiu's BABA board has split into its usual two tables before the print: one reciting the cloud-and-Qwen re-rating scripture, the other flipping through four straight quarters of misses. The flat, thin tape says neither table is pressing — everyone's waiting for Aug 20 to make the other side fold first.
- 华尔街共识 · 33 analysts, 1.17 composite rating (strong buy territory), average target HK$183.6 — 50.6% implied upside, among the fattest odds on the board. But this consensus has been slapped by the print four quarters running; Aug 20 is its credibility test.
- 摩根士丹利 · Kept Alibaba a top pick into earnings: trimmed the ADR target to $180 from $190 on softer e-commerce EBITA, but stays Overweight — expects cloud acceleration to beat expectations with the margin-expansion story intact, at roughly 13x F28 earnings.Source ↗
- 高盛 · Reiterated Conviction Buy: models cloud revenue up 38% YoY and sees AI-related products exceeding 50% of external cloud revenue within a year — one of the most aggressive bull calls into the print.Source ↗
