〔Day Digest〕Pre-FOMC Repricing Locks In ~90% Odds for a 25bp Hike; One-and-Done Caps the Path as Diesel, 30Y and AI-Bubble Warnings Stack — Buy the News, or Buy the Dip?

0. Weekly Arc

The pre-FOMC desk has converged: a 25bp September hike is now near-certain, with Huaxin Securities putting the implied probability at ~90% ^[1] and Goldman Sachs explicitly forecasting a 25bp move ^[2]. Yet the magnitude debate — one-and-done versus an extended cycle — is the live one. Sinolink's macro note argues there is room for only one 25bp hike ^[3], while Morgan Stanley's Wilson team flags that quality cyclicals and momentum tend to lead the first-hike phase ^[4]. Stagflation warnings from BofA's Hartnett (diesel crack spread at a $102/bbl record, 30Y at the highest since 2007) and Zeberg's Nasdaq 100 crash call stack against the consensus ^[5][6]. Net: the trade is buy-the-news on the hike, but the post-dot-plot path is unsettled ^[7][8][1].

1. Policy Narrative and the One-Hike Ceiling

• **[ESCALATED] Hawkish baseline — Goldman Sachs:** now predicts a 25bp Fed hike in September, the cleanest pre-meeting commitment on the tape ^[2].
• **[NEW] BoE timing shift — Goldman:** reverses its 2026 hold call, now seeing a 25bp Bank of England hike in November 2026 ^[9].
• **[ONGOING] One-and-done framing — Sinolink Macro:** argues only one 25bp move fits the data, noting that roughly three-quarters of high-sensitivity monthly inflation prints have been offset by subsequent reversals; today's inflation is "not high enough to force consecutive hikes, yet stubborn enough to deliver one" ^[3].
• **[NEW] Contrarian bullish — CICC:** invokes the 1997 analog in which a "preventive" Fed hike marked the peak in yields and was followed by a roughly one-week equity rebound; the firm frames a small hike as potentially "bad news priced in" ^[7].

2. US Equity Positioning Around the Hike

• **[NEW] Quality-over-cycle — Morgan Stanley (Mike Wilson team):** historically, in the first hike of a tightening cycle, quality blue chips and momentum outperform; the team notes the current cycle is being driven by strong growth rather than fiscal or inflation shocks, but flags that a Hormuz Strait blockade plus a China consumption rebound could turn a "moderate, preventive" move into a longer hike regime ^[4].
• **[NEW] Stagflation alert — BofA (Michael Hartnett):** names diesel as the real-economy pressure point with the crack spread at a $102/bbl record; flags the IYT transport ETF as the canary — a break of the 200-day by 80 points would confirm "summer best days are over" has rolled into autumn stagflation ^[6]. Past three weeks saw US equity net outflows of $14.2B, the largest three-week exit since January 2026, with weekly inflows shrinking to $7B from a $52B July peak ^[6].
• **[NEW] Three headwinds — Goldman Sachs Partner Mark Wilson (Weekly Mash):** cites a 0% median S&P return from early August to midterm elections since 1974 as one; warns CTAs and vol-target funds remain near full long — forced-sellers on any drawdown; favours AI infrastructure, commodities (especially copper), German equities, and large banks ^[10].
• **[NEW] Contrarian crash call — Danish economist Henrik Zeberg:** projects Nasdaq 100 at 39,000 (+33% by year-end 2026) followed by a collapse to 10,600 (the 2022 low); single source, treat as scenario not forecast ^[5].

3. China Research Highlights

• **[NEW] 2027 export outlook — Citic Securities:** H2 2026 export resilience intact, full-year growth near 20%; 2027 growth to slow materially though real-volume growth stays stable; quantifies a potential EU tariff drag at -0.3pp and an intensified US transshipment crackdown at -1.7pp on 2027 export growth ^[11]. Structural offsets named: computing hardware, ships, industrial robots, and the "new three" exports ^[11]. August PPI re-acceleration cited; calendar markers include this week's China August CPI/exports, US August CPI, and the September FOMC ^[11].
• **[ONGOING] A-share broker consensus on the hike:** ten-broker roundup led by Citic frames the Fed hike as a "buy point, not a sell point," recommending an AI + energy-chemicals structure with a North America AI supply-chain tilt ^[8]. Shanghai Composite -1.07% on the week, Shenzhen Component -0.34%, STAR50 -1.52%, ChiNext +1.08% ^[8].
• **[NEW] Huaxin Securities strategy:** Fed September probability pushed to 90%, base case 25bp; BoJ 25bp on 9/18; favours a US equity barbell (quality tech growth + defensives), curve bull-flattening into the dot-plot, and gold buy-on-dips above $4,300 ^[1].
• **[NEW] Funding-tape read — Guosen Securities (National Securities):** week-2 September saw a net A-share outflow of 16.7B yuan versus 34.5B the prior week; margin balance down 6.7B while public fund issuance added 10.9B; weekly turnover annualized 390% (70th percentile since 2005); margin-trading share 8.49% (43rd percentile) ^[12].

4. Single-Stock Research Roundup

• **[NEW] Akeso (康方生物, 9926.HK) — Bank of America:** HARMONi-2 OS hazard ratio 0.73 described as "encouraging"; 2026 set as a key year with global and China Phase III readouts; maintains Buy with DCF target HK$136 ^[13].
• **[NEW] Cambricon (寒武纪, 688256.SH) — CITIC CLSA:** initiates at Overweight with target RMB 1,479; relay is thin on the thesis ^[14].
• **[NEW] Pengding Holdings (鹏鼎控股, 002938.SZ) — Guosen Securities:** maintains Outperform; 1H26 revenue 17.22B yuan (+5.14% YoY), net profit attributable to parent 1.284B (+4.11% YoY); AI server and optical module gross margin up 20.59pp YoY; SLP products in 800G/1.6T optical modules in volume production, 3.2T in R&D; 9.6B yuan private placement to support AI interconnect board expansion ^[15].
• **[NEW] Tsingtao Beer (600600.SH) — Huafu Securities:** maintains Buy; 1H26 NP 3.92B (+0.4% YoY), Q2 NP 2.12B (-3.37%); 2026E-2028E NP trimmed to 4.7B/4.9B/5.1B (vs. prior 4.8B/5.2B/5.5B), implying +2%/+5%/+4% YoY; PE ~15x/14x/14x ^[16].
• **[NEW] Chuanjinnuo (300505.SZ) — Pacific Securities:** maintains Buy; 1H revenue 2.112B (+21.1% YoY), NP 124M (-30.2% YoY) on sulphur cost squeeze; 2025-2027E EPS 1.20/1.60/2.86 yuan; flags the Egypt Suez phosphochemical project (800kt sulphuric acid, 300kt wet-process phosphoric acid, 150kt 52% phosphoric acid, 300kt MAP, 20kt sodium fluorosilicate) plus a 100kt/year battery-grade iron phosphate precursor as growth anchors ^[17].
• **[NEW] Precision-manufacturing robotics name (relay truncated):** strong-call research on a company whose robotics business is framed as the breakout, with consumer electronics as a moderate tailwind; the relay was cut before the target and valuation were disclosed ^[18].
• **[NEW] SpaceX (private) — MoffettNathanson:** raises price target to $142 from $131 ^[19].

5. Contrarian and Tail Risks

• The market is split three ways: ~90% probability of a September 25bp ^[1], one-and-done as the consensus ceiling ^[3], and Zeberg's "rally then crash" path to 10,600 on the Nasdaq 100 ^[5]. The falsifier is the dot plot itself plus the September CPI print ^[1].
• Source quality control: the Zeberg crash call is a single-source scenario from a Danish economist, not a bank note ^[5]; the Hartnett diesel/IYT flag rests on a single Flow Show excerpt ^[6]; the Wilson three-headwinds piece truncates before the full prescription ^[10]. Treat the contrarian calls as scenarios; quote the hike-probability band, not a single point.