AI Lifts the Index; Quality Splits the Tape

2. Complete macro-event table

Impact	Time (ET)	Event	Actual vs consensus; prior/revision	Surprise	Immediate reaction / why it mattered
1	All day; oil settlement 2:30 p.m.	Hormuz / U.S.–Iran	White House: no talks under way; no verified operating corridor	More hawkish than early détente pricing	Brent reversed to $89.70 (+2.1%); WTI $83.53 (+1.6%).
2	European session	France fiscal/election stress	CAC 8,319.87 (-1.68%); BNP -4.79%, Société Générale -4.99%, Crédit Agricole -3.97%	France-specific equity underperformance	OAT 4.1011% (+2.3 bp), Bund 3.2500% (+2.4 bp); spread essentially flat at 85.11 bp.
3	8:30 a.m.	U.S. initial/continuing claims	203k vs 208k; prior 207k (revised +1k). Continuing 1.778m vs 1.790m; prior 1.796m (revised -3k)	Labor-strong	Brief yield bounce, little durable USD/equity response.
4	8:30 a.m.	U.S. advance goods trade / inventories	Deficit $118.806bn vs ~$100.5bn; prior $101.407bn. Wholesale +1.3% vs ~+0.2%; retail and ex-auto +0.7%	Growth-negative trade, growth-positive inventory	Muted tape; material Q3 GDP composition and AI-capex signal.
5	10:30 a.m.	EIA natural-gas storage	+15 Bcf vs 19–20 public calendars, but in line with a cited Bloomberg median; prior +16; five-year +33	Mildly bullish, feed-dependent	Oct. gas reached $2.990, closed $2.914 (+~1.4%); cushion still +5.5% vs five-year.
6	1:00 p.m.	$44bn U.S. seven-year note auction	4.512%, at dominant WI; cover 2.50x; indirect 60.78%, direct 26.96%, dealers 12.26%	Neutral headline, mixed quality	Treasury/dollar/equity proxies barely moved.
7	8:30 a.m.	Canada Q2 current account	C$8.84bn surplus vs C$3.5bn; prior C$8.31bn deficit	Stronger external balance	CAD strengthened about 0.2% to C$1.3855/U.S.$; first surplus in four years.
8	U.S. day	U.S.–Canada policy escalation	U.S. Lake Ontario renaming order; 50% U.S. tariffs on C$27.6bn already effective Aug. 22; Canadian retaliation due Sep. 8	Symbolic fresh action, binding adverse context	Limited clean same-day reaction; raises the durable USMCA and supply-chain risk premium.

3. Detailed macro-event sections with opinion clusters

3.1 Hormuz and U.S.–Iran: rhetoric reversed the tape; passage still resolves it

Facts. Brent traded near $86.48 and WTI $80.83 early as the market extended a de-escalation narrative. The move reversed after the White House said no U.S.–Iran talks were under way; Brent settled $89.70 (+2.1%) and WTI $83.53 (+1.6%). Public shipping evidence still showed materially subnormal flows and no broadly insurable, repeatable corridor. (Reuters oil wrap, 27 Aug.; BIMCO shipping analysis)

Cluster A — risk premium remains rational until physical flows normalize. Giovanni Staunovo/UBS, Ole Hansen/Saxo and Niels Rasmussen/BIMCO emphasized constrained passage, inventories and the difference between diplomatic language and operational shipping. The evidence is the late oil reversal and still-low transit counts. Implication: prompt crude and inflation breakevens retain upside convexity over days to weeks. Risk/disconfirmer: several consecutive days of insured, loaded outbound tanker passage. Confidence: high on the operational test, medium on price direction.

Cluster B — headlines can compress price before barrels return. Priyanka Sachdeva/Phillip Nova, June Goh/Sparta Commodities and Matt Miskin/Manulife stressed the market's fast response to negotiations, spare capacity and demand uncertainty. Implication: $5–$10 swings can occur without proportional changes in physical supply. Risk: an enforcement or military escalation makes headline mean-reversion unsafe. Confidence: medium-high.

Consensus / disagreement / resolver. Public experts agreed that diplomacy matters and physical passage is not normal. The sharp disagreement is how much risk premium belongs in price before verification. Vessel transits, insurance terms, sanctions enforcement, Gulf loadings and inventory draws resolve it.

3.2 France: a bank-equity warning without a fresh spread break

Facts. The CAC 40 fell 1.68% to 8,319.87, versus -0.72% for the STOXX 600 and +0.27% for the DAX. BNP Paribas lost 4.79%, Société Générale 4.99% and Crédit Agricole 3.97%. France's 10-year yield rose 2.3 bp to 4.1011%, but the Bund rose 2.4 bp to 3.2500%, leaving the spread essentially unchanged at 85.11 bp. (Reuters European close, 27 Aug.; AFP Paris close, 27 Aug.)

Cluster A — structural credibility premium. Kevin Thozet/Carmignac saw room for the OAT–Bund spread to reach 100 bp; Christopher Dembik/Pictet and David Zahn/Franklin Templeton argued that leading political platforms do not supply a credible consolidation path. Evidence: a 5.1%-of-GDP 2025 deficit, Q1 debt around 117.5% of GDP and €310bn of planned 2026 issuance. Implication: relative OAT and French-bank underperformance through the budget and 2027 election. Disconfirmers: enacted quantified savings and a sustained spread below 75–80 bp. Confidence: high on persistence, medium on >100 bp.

Cluster B — compromise dilution and rating signalling. Isabelle de Gavoty/AllianzGI expected parliamentary compromise to weaken deficit reduction; Olivier Malteste/Yomoni saw an outlook warning as more plausible than an immediate Fitch downgrade. Implication: volatility into the Aug. 28 rating review and September budget. Risk: a stable affirmation and credible arithmetic. Confidence: medium.

Cluster C — sovereign-bank-real-economy transmission. Malteste and Fiona Cincotta/City Index linked easier fiscal policy and higher yields to banks, mortgages, SMEs and investment. The bank tape supports the first link, not yet a credit crunch. Confidence: high on the day's transmission vehicle, medium on multi-quarter credit damage.

Consensus / disagreement / resolver. The consensus was a durable political premium; the disagreement concerned convexity and timing. Fitch's review, the September draft budget, auctions, bank funding and a sustained move above 100 bp are the decisive tests.

3.3 U.S. claims: low layoffs, not a hiring boom

Facts. At 8:30 a.m. ET, initial claims were 203,000 versus 208,000 expected; the prior was revised to 207,000 from 206,000. Continuing claims were 1.778 million versus 1.790 million, with the prior revised to 1.796 million from 1.799 million. The initial four-week average nevertheless rose to 205,500 from 204,250. (U.S. Department of Labor, 27 Aug.; Reuters analysis, 27 Aug.)

Cluster A — broadly balanced labor market. Nancy Vanden Houten/Oxford Economics, Thomas Simons/Jefferies and Kathleen Stephansen/Haver emphasized historically low layoffs and soft labor supply as well as demand. Implication: no recessionary layoff signal and less labor-side urgency for near-term easing. Risk: claims persistently above 230k–250k and rising unemployment. Confidence: high on current layoffs, medium through year-end.

Cluster B — “no hire, no fire.” Neil Sethi and the AP synthesis cautioned that claims measure separation, not job-finding: low layoffs can coexist with weak payrolls and poor entry opportunities. Implication: incumbent income remains supported while confidence and hiring flows stay soft. Disconfirmers: stronger payrolls, vacancies and quits. Confidence: medium-high. (AP labor report, 27 Aug.)

Cluster C — mildly hawkish, not tradable alone. Pablo Piovano/FXStreet observed a small dollar/yield lift; Erik Bregar/Silver Gold Bull said larger Fed/Treasury signals left FX inert. Implication: directionally reduces easing urgency but did not move the modal path. Confidence: high on the muted tape, low on standalone policy magnitude.

Consensus / disagreement / resolver. No public named economist treated the print as a labor break. The live dispute is stable balance versus a low-hire trap; August payrolls, the benchmark revision and the next four claims weeks resolve it.

3.4 Advance goods trade and inventories: an AI-investment deficit

Facts. Census released the report at 8:30 a.m. ET. The July goods deficit widened to $118.806bn versus roughly $100.5bn expected and a revised $101.407bn in June. Exports fell $6.035bn (-2.9%) while imports rose $11.364bn (+3.7%); capital-goods imports rose $14.231bn (+11.3%), while consumer imports rose only 0.1%. Wholesale inventories rose 1.3% versus roughly 0.2%; retail and ex-auto retail inventories each rose 0.7%. (Census advance report)

Cluster A — productive “boom deficit.” Matthew Martin/Oxford Economics and Ershang Liang/PNC attributed the import wave to AI, data-center and high-tech capital spending, while tariffs restrained consumer imports. Implication: net trade may subtract around one percentage point from Q3 GDP, partly offset by inventory and productive-capacity formation; equipment/semiconductor demand persists into 2027. Risks: cancellations, import payback and weak domestic final sales. Confidence: high on composition, medium on persistence. (PNC note, 27 Aug.)

Cluster B — broad two-sided deterioration. Winnie Tapasanun/Haver stressed a third export decline alongside a fifth import increase in six months. Implication: the external drag is not only healthy investment; export softness also matters. Risk: full-trade services or revisions improve the picture. Confidence: medium-high.

Cluster C — tariff strategy failed. Mish Shedlock characterized the widening as tariff backfire. Composition supports a timing/hedging channel but does not prove the counterfactual causal claim. Implication: larger deficits despite tariffs weaken the policy's stated metric. Confidence: medium on the empirical headline, low-medium on singular causation.

Consensus / disagreement / resolver. The consensus is a material Q3 net-export drag cushioned by inventories. The sharp disagreement is productive capex versus tariff-policy failure. The Sept. 3 full trade report, sales/inventory ratios and August–September capital-import persistence resolve it.

3.5 EIA gas storage: tighter direction, comfortable level

Facts. At 10:30 a.m. ET, EIA reported a 15 Bcf injection for the week ended Aug. 21, versus 19–20 Bcf on broad public calendars but equal to a Bloomberg median cited by AEGIS; the prior was 16 Bcf and the five-year seasonal build 33 Bcf. Stocks reached 3,184 Bcf, 30 Bcf below 2025 but 167 Bcf (5.5%) above the five-year average. South Central withdrew 19 Bcf, almost entirely from salt storage. October gas reached $2.990 and closed $2.914, about 1.4% higher. (EIA release; AEGIS, 27 Aug.)

Cluster A — heat and LNG demand now show in stocks. James Hyerczyk/FXEmpire and Thomas Liardon/AEGIS pointed to the seasonal shortfall, hotter forecasts, production softness and feedgas recovery. Implication: prompt support over days to weeks; holding the 50-day average can open $3. Risks: cooling, production rebound or feedgas slippage. Confidence: medium-high.

Cluster B — weather rotation, not supply regime change. Robert Vaughn/SynMax traced the lean build to South Central power burn while production remained near records. Implication: shoulder-season cooling should restore injections. Risk: continued September heat plus LNG demand. Confidence: high for the reported week, medium forward.

Cluster C — supportive one-off inside a cushion. Newsquawk desk commentary treated a 4–5 Bcf miss as normal estimation noise given stocks 5.5% above average. Implication: winter repricing requires repeated shortfalls. Confidence: medium-high.

Consensus / disagreement / resolver. The balance was tight versus seasonality; debate begins with which forecast feed is used and ends with whether another two to four lean builds shrink the cushion materially. Production, feedgas, South Central salt stocks and weather resolve it.

3.6 Seven-year Treasury auction: adequate at a high yield

Facts. Treasury's $44bn new seven-year note stopped at 4.512% at 1:00 p.m. ET, on the dominant real-time WI quote (one vendor implied a trivial 0.4 bp tail), with a 2.50x cover. Indirects took 60.78% versus 65.10% over the prior six; directs 26.96% versus 23.14%; dealers 12.26% versus 11.77%. Minute-level Treasury ETF, dollar and equity proxies showed no meaningful discontinuity. (TreasuryDirect result; TreasuryDirect API)

Cluster A — average with a slight weak tilt. Greg Michalowski/InvestingLive graded the auction C: normal headline demand, softer indirect participation and stronger directs. Implication: neutral to mildly bearish over hours, with no standalone trade. Confidence: medium-high on grade, low on direction.

Cluster B — composition may be less durable. Newsquawk's desk saw direct-for-indirect substitution as a watch item. Indirects are not a clean foreign proxy, so one auction cannot establish withdrawal. Implication: recurrence would raise belly/back-end absorption risk. Confidence: medium.

Cluster C — no-drama normalization versus high-yield unease. ZeroHedge commentary called it a snoozer; Paulo Trevisani/Dow Jones emphasized the near-two-year-high yield and a vendor-implied small tail. The tape supports “no drama” more than a demand shock. Confidence: medium.

Consensus / disagreement / resolver. Adequate absorption and no dealer indigestion are consensus. Repeated indirect shares below roughly 60%–62%, larger dealer awards and genuine tails would validate the bearish quality thesis; a rebound toward 65% would disconfirm it.

3.7 Canada current account

Facts. Statistics Canada released Q2 balance-of-payments data at 8:30 a.m. ET. The current account swung to a C$8.836bn surplus from a revised C$8.310bn deficit (initially -C$7.184bn). Public consensus feeds ranged from -C$2.0bn to +C$3.9bn, so the beat was C$4.9bn–C$10.8bn. The goods balance improved C$18.601bn as exports rose 13.1%; energy exports rose 27.4% to a record C$60.6bn. Goods imports also rose to a record, so this was not an import-collapse surplus. Foreign purchases of Canadian debt reached a record C$110.2bn, but foreign investors sold C$9.6bn of equities. (Statistics Canada release; official table)

CAD was about 0.2% firmer at USD/CAD 1.3855 by 12:51 p.m.; the GoC 10-year was 2 bp higher at 3.677%. Firmer oil confounds causal attribution, and no reliable public five-minute price window was located. (Reuters, 27 Aug.)

Cluster A — high-quality export beat, temporary windfall risk. Statistics Canada's composition and same-day Dow Jones/MT Newswires reporting show a broad export increase with rising imports, but Iran-driven energy prices supplied the largest contribution. Implication: modest support for CAD, nominal GDP and energy cash flow over one to three months. Risks: oil normalization, auto restocking reversal and U.S. tariffs. Confidence: high on Q2 composition, medium-low on persistence.

Cluster B — debt-flow support with a future income cost. Statistics Canada and Nova Scotia Finance highlighted record foreign government-bond purchases, while higher interest paid to foreign bondholders widened the portfolio-income deficit. Implication: near-term duration/CAD funding support, but larger coupon outflows over one to three years. Risk: a term-premium shock or flow reversal. Confidence: medium-high near term.

Cluster C — the tariff/BoC regime still dominates. Tony Valente/AscendantFX said CAD resilience coexisted with downside trade risk; Francesco Pesole/ING expected trade uncertainty to weigh on CAD; Mirza Shaheryar Baig/Desjardins described low volatility as resilience or complacency. The small tape response supports treating the current account as second-tier for policy. Implication: await GDP, CPI, labor and Q3 trade rather than chase the headline. Confidence: high on immediate hierarchy, medium on three-month FX.

Consensus / disagreement / resolver. The release was genuinely export-led and constructive, but experts disagree on durable rebalancing versus a terms-of-trade windfall. Aug. 28 GDP, Q3 trade volumes, oil, tariffs, the Sept. 2 BoC decision and the Nov. 27 current-account release resolve it.

3.8 U.S.–Canada policy escalation

Facts. At roughly 1:30 p.m. ET, President Trump signed an order directing U.S. federal agencies to rename Lake Ontario “Lake America” within 30 days. It changes federal nomenclature but contains no tariff, navigation, border or sanctions provision and cannot compel Canadian, state, private or international usage. The binding backdrop is the 50% Section 338 duty on more than 550 Canadian product lines—about US$20bn/C$27.6bn annually, or roughly 5% of Canadian U.S.-bound merchandise—effective Aug. 22. Canada's rate-for-rate counter-tariffs on C$27.6bn of U.S. goods take effect Sept. 8. (White House order; Canada Finance; AP scope and product map)

No discrete naming-order premium was visible: CAD and the TSX were broadly range-bound, while Magna fell just 0.44% Thursday after losing 6.66% on Monday's auto-tariff threat. The observed hierarchy was current account, oil, bank earnings and Nvidia over symbolism.

Cluster A — symbolic today; tariff clock matters. Tony Valente/AscendantFX, Jack Nguyen/Verecan Capital and the contained tape imply the naming order has no direct cash-flow effect. Implication: retain downside asymmetry in exposed manufacturing/CAD into Sept. 8 rather than trade the label itself. Risk: symbolism triggers a binding retaliation. Confidence: high.

Cluster B — negotiating theater with an off-ramp. Wendy Cutler, Christopher Sands/Johns Hopkins and Scott Lincicome/Cato argued that limited current coverage, mutual costs and the Sept. 8 delay leave room for a deal. Implication: mediation, exemptions or resumed talks would lift CAD and Canadian cyclicals. Risk: domestic political red lines and repeated escalation. Confidence: medium-high.

Cluster C — concentrated stagflationary pain. Trevor Tombe/University of Calgary estimated roughly 85,000 direct and supply-chain jobs at risk; Randall Bartlett and Royce Mendes/Desjardins saw weaker growth but tariff/oil inflation that complicates BoC easing. Implication: underweight directly exposed machinery, electronics, furniture/textiles and suppliers over one to four quarters; front-end rates remain two-sided. Confidence: medium.

Cluster D — USMCA/capex regime shift. William Reinsch/CSIS and Bradley Saunders/Capital Economics emphasized investment uncertainty and the possibility that persistence damages the continental framework more than first-round duties do. Implication: a multi-year risk premium for cross-border autos, steel and logistics. Disconfirmers: rollback, a replacement framework or legal defeat of Section 338. Confidence: medium.

Consensus / disagreement / resolver. Consensus separates symbolic naming from live tariffs. The sharp disagreement is an off-ramp versus regime shift. Sept. 8 implementation, any January auto/parts instrument, Ontario retaliation, BoC inflation guidance, court challenges and cross-border production decisions resolve it.

4. Complete earnings/call table

Basis note: reported and consensus bases are aligned where possible. “Clean” removes disclosed refund/tax distortions; “n/a” means no credible public point-consensus was found. Hong Kong post-close filings have no Aug. 27 post-result cash reaction unless explicitly stated.

Impact	Company (ticker)	Release / call	Result versus public consensus	Guide / central call issue	Reaction
1	Marvell (MRVL)	4:05 / 4:45 p.m. ET	$2.739bn / $0.94 vs ~$2.71bn / $0.92–0.93	Q3 $3.15bn / $1.10 above Street; FY28 $18bn; custom mix lowers GM	$222.67, -7.8% AH
2	Elastic (ESTC)	4:05 / 5:00 p.m.	$478.1m / $0.70 vs $469.5m / $0.58	Cloud, cRPO and FY revenue/EPS raised	$100.83, +20.4% AH
3	Gap (GAP)	4:15 / 5:00 p.m.	$3.651bn miss; clean EPS $0.52 vs $0.48	FY sales narrowed; margin/EPS raised; Old Navy reset	$23.91, +15.0% AH
4	Build-A-Bear (BBW)	6:45 / 9:00 a.m.	$115.3m, -5.2%; EPS ~in line	FY sales cut to $500–525m; wholesale and traffic reset	$28.44, -27.3%
5	Burlington (BURL)	6:45 / 8:30 a.m.	Sales/comps missed; refund-ex clean EPS $2.37 beat	H2 underlying guide unchanged; $55m price reinvestment	$290.00, -7.6%
6	Hormel (HRL)	6:30 / 9:00 a.m.	$2.961bn miss; $0.37 beat; organic volume -7.1%	FY sales cut; Q4 EPS run-rate lowered	$21.28, -10.2%
7	Dollar General (DG)	6:50 / 9:00 a.m.	$11.290bn / $2.48 / +3.5% comps beat; clean EPS ~$2.23	FY comps/EPS raised; refund is ~$0.25	$125.89, +2.5%
8	Dollar Tree (DLTR)	6:30 / 8:00 a.m.	$4.887bn / +3.7%; $2.70 reported, ~$1.39 clean vs $1.13	FY clean EPS ~$7.10–7.45; $173m H2 reinvestment	$127.00, -3.9%
9	Best Buy (BBY)	7:00 / 8:00 a.m.	$9.779bn / $1.47 / +4.1% comps beat	FY raised; $34m refund (~$0.12/share), units weak	$83.56, -4.4%
10	HealthEquity (HQY)	8:00 / 8:30 a.m.	$350.7m / $1.24 vs $349.2m / $1.19	Minimal revenue/EBITDA raise; EPS guide unchanged	~$93.39, -10.5%
11	Rubrik (RBRK)	4:05 / 5:00 p.m.	$427.3m / $0.20 vs $396.4m / $0.04; ARR beat 1.2%	FY revenue/EPS/FCF materially raised	~$95.8, -10.5% AH after +11.4% session
12	SentinelOne (S)	4:11 / 5:00 p.m.	$292.0m / $0.08 vs $290.3m / $0.07	Revenue/operating income up; EPS guide down	$21.72, -4.7% AH after +10.7% session
13	Affirm (AFRM)	4:32 / 5:00 p.m.	$1.166bn vs $1.107bn; GMV $14.1bn; GAAP EPS tax-distorted	FY27 GMV >$64bn; RLTC/GMV ~4.16%	$84.51, +9.1% AH
14	IREN (IREN)	4:10 / 5:00 p.m.	$137.2m; adjusted loss better on one feed; consensus conflicted	>$4bn contracted exit ARR; FY27 capex $25–30bn	$37.39, -7.8% AH
15	Autodesk (ADSK)	4:01 / 5:00 p.m.	$2.046bn / $3.30 vs $2.01bn / $3.12; billings beat	Q3 revenue above, EPS below; FY revenue raised	$257.50, -4.8% AH after +6.2% session
16	Workday (WDAY)	4:01 / 4:30 p.m.	$2.649bn / $2.75 vs $2.64bn / $2.61; cRPO +14.2%	Q3 cRPO slows; FY margin up; FY28 growth ~11%	$194.00, +0.2% AH after initial -5%
17	Ulta Beauty (ULTA)	4:05 / 4:30 p.m.	$3.036bn / $6.55 / +3.8% comps beat	FY raised; H2 comps 2%–3%, transactions flat	$523.00, -3.2% AH
18	Malibu Boats (MBUU)	7:00 / 8:30 a.m.	$295.5m / $0.92 vs ~$263m / $0.75–0.79	FY27 $1.08–1.12bn; dealer stock normalized	$28.80, +8.1%
19	Bilibili (BILI)	6:00 / 8:00 a.m.	RMB7.94bn / RMB1.58 modest beats; ads +28%	Games expected to return to growth in Q4	$16.77, +3.8%
20	RBC (RY)	6:00 / 8:30 a.m.	C$18.538bn / C$4.28 vs C$18.140bn / C$4.08	Premium quality vs fee-heavy mix; CET1 13.5%	U.S. -1.3%; Toronto -1.6%
21	TD Bank (TD)	~5:30 / 8:00 a.m.	Adj. EPS C$2.77 vs C$2.47; PCL C$917m favorable	Cost control/buybacks vs unresolved U.S. consent order	U.S. +1.4%; Toronto +1.2%
22	CIBC (CM)	~5:30 / 7:30 a.m.	C$8.368bn / C$2.73 beat; PCL favorable	Markets mix, U.S. reserve release, higher Q4 costs	U.S./Toronto -2.8%
23	Pernod Ricard (RI FP)	1:30 / 3:00 a.m. ET	€9.404bn slight sales miss; €2.423bn PRO near consensus	FY27 sales broadly flat; no U.S. market growth through FY29	€64.50, -4.6%
24	Prudential plc (PUK/PRU/2378)	Aug. 26 6:00 p.m. / Aug. 27 4:30 a.m. ET	NBP $1.384bn slightly soft; OFSG and margins strong	Double-digit FY growth requires H2 acceleration	HK -1.5%; London -2.5%; ADR -2.0%
25	Mixue Group (2097 HK)	12:05 p.m. HKT; no formal call found	Revenue +2.3%, profit -14.7%; no reliable H1 consensus	Stores +20.7% but sell-in/store proxy -17.2%	HK -8.4% post-print
26	China Life (2628 HK)	4:55 p.m. HKT; webcast Aug. 28	NBV +33.7%; profit +228.6% on fair-value gains	Liability quality vs equity-beta earnings	Next-session reaction
27	China Pacific Insurance (2601 HK)	5:29 p.m. HKT; presentation 7:02	NBV +12.7%; net profit +10.4%; no clean H1 consensus	Better underwriting vs lower comprehensive yield/solvency	Next-session reaction
28	CHALCO (2600 HK)	9:49 p.m. HKT	Revenue +7.7%; profit +67.9%, upper half of guide	Primary aluminum strength vs alumina loss; DPS +124%	Next-session reaction
29	Tianqi Lithium (9696 HK)	9:09 p.m. HKT	Revenue +153%; profit RMB4.24bn near guide top	Price-led leverage vs July lithium normalization	Next-session reaction
30	Wynn Macau (1128 HK)	8:21 p.m. HKT	Revenue/EBITDA +14%; profit flattered by derivative gain	Mass share/dividend vs hold, leverage and capex	Next-session reaction
31	Haier Smart Home (6690 HK)	10:47 p.m. HKT	Revenue -2.8%; profit -14.3%; thin consensus conflicted	HVAC/Q2 stabilization vs white-goods/FX weakness	Next-session reaction
32	OOIL (0316 HK)	8:44 p.m. HKT	Revenue +6.1%; profit -23.7%; no H1 consensus	Q2 freight acceleration vs cost/margin damage	Next-session reaction
33	China Vanke (2202 HK)	8:21 p.m. HKT	Revenue -33.4%; loss RMB14.95bn; thin consensus unusable	Extensions buy time; Deloitte going-concern warning	Next-session reaction
34	Akeso (9926 HK)	9:59 p.m. HKT	RMB1.806bn revenue ~24% below thin consensus; loss vs expected profit	Commercial miss vs ivonescimab catalysts/liquidity	Next-session reaction
35	CCCC (1800 HK)	4:43 p.m. HKT	Revenue -1.2%; profit -24%; orders -8.9%	Overseas +20.6% orders vs domestic/debt stress	Next-session reaction
36	China Longyuan (0916 HK)	10:15 p.m. HKT	Revenue -6.5%; profit -28%; derived Q2 ~43% below thin estimate	Poor wind/tariffs vs capex restraint	Next-session reaction
37	Weichai Power (2338 HK)	8:40 p.m. HKT	Revenue +8.8%; profit +36.5%; consensus feeds conflicted	Engine/data-center volumes vs margin/nonrecurring gains	Next-session reaction

5. Detailed company sections with opinion clusters

5.1 Marvell (MRVL): a beat below the whisper bar

Facts. Revenue was $2.739bn (+37%) and adjusted EPS $0.94, modestly above consensus; data-center revenue reached $2.172bn (+46%). Q3's $3.15bn/$1.10 guide exceeded Street, and FY28 revenue rose to $18bn, but custom-heavy mix lowers Q3 gross margin to 57.5%–58.5%. Shares fell 7.8% after hours. (Marvell release)

• Broad AI demand, back-end-loaded custom upside — management, JPMorgan/Harlan Sur pre-print. Optical, switching and custom ramps are real; Google-related FY28 work was already embedded. Implication: positive for connectivity/advanced packaging over 2–8 quarters. Risk: capacity prepayments and customer concentration. Confidence: high on demand, medium on incremental upside.
• Growth quality versus valuation — Goldman/James Schneider and the tape. A 188% YTD rally required more; gross-margin dilution offset the raise. Confidence: high on the day's explanation.

Consensus / disagreement / resolver. Consensus says AI orders strengthened; disagreement is price paid and FY29+ custom upside. The Oct. 6 Investor Day, Q3 data-center growth above 20% q/q and Q4 operating margin at 38%–40% resolve it.

5.2 Elastic (ESTC): consumption conversion finally showed up

Facts. Revenue of $478.1m and EPS of $0.70 beat $469.5m/$0.58; cloud revenue beat by 2.8%, cRPO rose 21% and adjusted FCF margin was 30%. FY revenue/EPS rose; the stock gained 20.4% after hours. (Elastic filing)

• Platform reacceleration — Rosenblatt/Blair Abernethy, management, RBC call framing. Large-customer adds, committed-cloud consumption and AI/search/security use validate durable reacceleration. Implication: positive for hybrid search and SIEM displacement over 2–6 quarters. Confidence: high.
• Proof/valuation skepticism — D.A. Davidson/Lucky Schreiner, JPMorgan/Piper/Wells Q&A. NER slipped to 111%, monthly cloud was flat and solution mix remains undisclosed. Risk: commitments do not convert. Confidence: medium.

Resolver: Q2 annual-cloud/cRPO, NER improvement, disclosed AI-cohort expansion and the promised Q4 exit acceleration.

5.3 Gap (GAP): profit discipline outranked the sales miss

Facts. $3.651bn sales missed, but clean EPS of $0.52 beat and adjusted gross margin was 41.4%. Gap and Banana Republic comps beat; Old Navy missed at -4% and Athleta remained -12%. FY sales narrowed while margin/EPS rose; shares gained 15% after hours. (Gap release)

• Fixable Old Navy execution — management and Dana Telsey's call framing. Weak dresses/shorts/swim and ineffective marketing explain much of the miss; new CEO Michael Francis and early denim improvement are the recovery case. Implication: brand-specific, not consumer-collapse evidence. Confidence: medium-high.
• Margin quality versus weak demand — Goldman/Brooke Roach and Baird/Mark Altschwager Q&A. Inventory/AUR discipline is real, but nearly 60% of sales sit at a weak Old Navy and the EPS raise largely reflects buybacks/tariff assumptions. Confidence: high.

Resolver: Q3 Old Navy traffic/comps without deeper promotions, Athleta stabilization and merchandise margin excluding tariff help.

5.4 Build-A-Bear (BBW): product and wholesale credibility reset

Facts. Revenue of $115.3m missed consensus 5.2%; EPS was approximately in line, gross margin fell 340 bp and FY sales fell to $500m–$525m from $530m–$550m. E-commerce demand dropped 15.6%, a Walmart wholesale program did not recur, and the stock lost 27.3%. (Build-A-Bear release)

• Merchandising execution failure — management, Eric Beder/SCC and Keegan Cox/D.A. Davidson Q&A. Non-core plush concepts drove weak traffic and promotions; a stronger Halloween range could repair it. Confidence: high on diagnosis, medium-low on rebound.
• Channel/credibility break — D.A. Davidson and CJS/Northland Q&A. Wholesale growth fell from 20%+ planned to flat, while the growth-chief exit compounded uncertainty. Implication: treat partner programs as nonrecurring until replenishment proves otherwise. Confidence: high.

Resolver: Q3 traffic/e-commerce, Q4 promotion, wholesale contracts and new-store productivity.

5.5 Burlington (BURL): high-quality margin beat, weak traffic

Facts. Sales and 2% comps missed, but refund-excluded EPS of $2.37 and 100 bp margin expansion beat. Traffic was flat; Q3 reported EPS of $1.60–$1.70 is below Street because a $55m refund is fully reinvested in price. Shares fell 7.6%. (Burlington release)

• Execution/unit-growth bulls — JPMorgan/Matthew Boss, Barclays/Adrienne Yih-Tennant, Telsey/Dana Telsey. Merchandise margin, supply-chain leverage and sub-two-year new-store paybacks support multi-year EPS compounding. Risk: cannibalization and inventory. Confidence: high on Q2, medium forward.
• Top-line skeptics — Wells Fargo/Ike Boruchow, BofA/Lorraine Maikis, Goldman/Brooke Roach. A premium multiple cannot ignore flat traffic, cautious consumer language and off-price peer outperformance. Confidence: high.

Resolver: Q3/Q4 traffic versus basket, comp relative to Ross/TJX and sales lift per reinvestment dollar.

5.6 Hormel (HRL): cost-supported EPS did not cure volume

Facts. Revenue missed at $2.961bn; adjusted EPS beat at $0.37, but organic volume fell 7.1% and Retail 9%. FY sales fell to $12.1bn–$12.2bn and the implied Q4 EPS midpoint dropped to roughly $0.37. Shares fell 10.2%. (Hormel release)

• EPS resilience versus volume deterioration — Newsquawk and Zacks editorial research. Lower SG&A/advertising and commodity relief produced the beat; volumes and throughput define the next leg. Implication: negative for branded packaged-food utilization. Confidence: high.
• Delayed pork benefit versus freight/beef drag — management; Barclays/BofA/JPMorgan call framing. Pork helps with a lag, but lower inventory turns, fuel, beef and plant absorption can consume it. Confidence: medium.

Resolver: Q4 Retail volume, gross margin, freight/fuel and FY27 organic sales excluding exits.

5.7 Dollar General (DG): a real turnaround inside a refund-assisted headline

Facts. Sales of $11.290bn, +3.5% comps and EPS $2.48 beat; a net tariff refund added about $0.25/share, leaving inferred clean EPS near $2.23, still roughly 11% above consensus. FY comps and EPS rose; DG closed 2.5% higher after a large premarket fade. (Dollar General release)

• Operational turnaround — Oppenheimer/Rupesh Parikh, Jefferies/Corey Tarlowe. Positive traffic, non-consumable breadth, shrink improvement and controlled inventory support estimate/multiple improvement over 2–6 quarters. Confidence: high.
• Refund/valuation skepticism — Goldman/Kate McShane, Telsey/Joe Feldman, closing tape. The refund is 43% of the EPS-guide midpoint increase and disappears in H2. Confidence: high.

Resolver: ex-refund H2 margin, Q4 storm-lap comps, promotion, fuel and actual buyback accretion.

5.8 Dollar Tree (DLTR): the reported guide was optically worse than the clean guide

Facts. Sales of $4.887bn and 3.7% comps beat. Reported EPS was $2.70, but management's $1.31 net tariff item leaves about $1.39 clean versus $1.13 consensus. FY clean EPS is roughly $7.10–$7.45; Q3's reported $0.80–$0.95 becomes $1.30–$1.45 before planned reinvestment. The stock fell 3.9%. (Dollar Tree filing)

• Traffic/execution inflection — management, Telsey/Joe Feldman, JPMorgan/Matthew Boss. Traffic turned positive early, store standards improved and clean EPS beat. Confidence: high on Q2, medium on durability.
• H2 margin air pocket — UBS/Michael Lasser, Barclays/Seth Sigman, the tape. $173m of H2 reinvestment, Q4 gross-margin pressure and essentials-heavy mix create reported volatility and a 2027 proof burden. Confidence: high on timing, medium on structural drag.

Resolver: Q3 traffic, discretionary holiday, clean gross margin, fuel surcharges and returns on the $210m program.

5.9 Best Buy (BBY): dollars grew faster than units

Facts. Revenue, EPS and 4.1% comps beat; FY revenue/comps/EPS rose. Yet the $34m tariff refund was about $0.12/share—larger than the headline EPS beat—and another similar refund is assumed in Q3. Computing ASP rose mid-teens while units fell high single digits; shares lost 4.4%. (Best Buy release)

• Innovation/replacement cycle — management, Telsey/Joe Feldman, EMARKETER/Zak Stambor. TV >10%, emerging devices and August strength support specialty-tech share gains over 2–6 quarters. Confidence: medium-high.
• Lower-quality beat — Citi/Steven Zaccone and the tape. Refunds, memory-price inflation and event timing cap the multiple until unit/margin proof arrives. Confidence: high.

Resolver: computing units and product margin, refund-ex Q3 margin, TV/emerging contribution and Ads/Marketplace profit—not GMV.

5.10 HealthEquity (HQY): good quarter, almost no incremental guide

Facts. Revenue of $350.7m and EPS of $1.24 beat; HSAs reached 10.739m and assets $37.921bn. Yet the revenue midpoint rose only $1m, EBITDA $3m, and EPS guidance remained unchanged around $4.70 versus a roughly $4.72 Street view. Shares fell about 10.5%. (HealthEquity release)

• Structural HSA growth — management and the call's large-bank analysts. HSA growth, 24% new-sales acceleration and healthcare inflation support multi-year adoption. Confidence: high.
• Rate/service-price skepticism — call Q&A and the tape. Only 2% cash growth, hedges that delay rate upside and service-price pressure make the guide look conservative or earnings momentum thin. Confidence: medium-high.

Resolver: cash balances/yields, HSA net adds, service revenue per account and an actual EPS-guide raise.

5.11 Rubrik (RBRK): strong raise, stronger expectations

Facts. Revenue of $427.3m beat 7.8%, EPS $0.20 beat $0.04 and ARR beat 1.2%; FY revenue, EPS and FCF rose materially. Shares first gained 11.4% in the session, then fell about 10.5% after hours. (Rubrik release)

• Organic resilience acceleration — Guggenheim/John DeFucci and management. Revenue/ARR/profitability improved without Strata contribution; data/identity resilience budgets remain durable. Confidence: high.
• H2 deceleration/option-value skeptics — Citi/Fatima Boulani, BMO/Keith Bachman. Guidance implies slower net-new ARR, cloud ARR narrowly missed, and Agent Cloud has only 15+ paying customers. Confidence: medium-high.

Resolver: Q3 net-new/cloud ARR, normalized revenue growth and Agent Cloud/identity monetization.

5.12 SentinelOne (S): AI products are real; consolidated acceleration is not yet

Facts. Revenue/EPS beat, ARR rose 22% and RPO 45%; non-endpoint products exceeded half of ARR. But Q3/FY EPS guidance fell, quarterly FCF remained negative and gross margin dropped 200 bp. Shares reversed a 10.7% session gain to -4.7% after hours. (SentinelOne results)

• AI-security platform inflection — Needham/Mike Cikos and management. Prompt/Purple AI ARR tripled and Flex passed 10% of ARR; enterprises are funding AI guardrails. Confidence: high on product demand, medium on companywide acceleration.
• Profit/FCF skepticism — UBS/Roger Boyd, Jefferies/Joseph Gallo and the tape. Net-new ARR grew only 4%; reinvestment slows H2 margin expansion. Confidence: high.

Resolver: total ARR above 22%, quantified AI ARR, NRR, gross-margin stabilization and positive quarterly FCF.

5.13 Affirm (AFRM): operating credit beat, tax-distorted EPS

Facts. Revenue was $1.166bn (+33%), GMV $14.1bn (+36%) and RLTC $589m, all materially above relevant bars. GAAP EPS of $4.62 was dominated by a $1.448bn tax benefit; the operating signal was RLTC holding 4.2% of GMV as transaction costs fell. FY27 GMV is above $64bn; shares gained 9.1% after hours. (Affirm results)

• Scale/margin flywheel — management and Goldman/Will Nance call framing. Card growth, user frequency and funding execution support >30.5% adjusted margin. Risk: gain-on-sale mix and a lost promotion. Confidence: high on Q4, medium forward.
• Controlled credit, limited macro signal — Wells Fargo/Jason Cooperberg and RBC/Dan Perlin Q&A. Delinquencies improved q/q but 30+/60+ remained slightly worse y/y; Affirm's underwriting is selective and not the whole consumer. Confidence: high.

Resolver: vintages, approval rates, charge-offs/provisions, ABS spreads and RLTC excluding securitization gains.

5.14 IREN (IREN): contracted AI future, enormous funding bridge

Facts. AI Cloud revenue doubled q/q to $70.5m and overtook Bitcoin mining at $66.7m, but adjusted EBITDA fell to $19.2m and FY27 capex is $25bn–$30bn. Management says contracted December-exit ARR exceeds $4bn and mining shuts by December; shares fell 7.8% after hours. (IREN release)

• Scarce-infrastructure bulls — management and customers' prepayments. Accepted Microsoft capacity, higher revenue/MW and 45%–55% GPU prepayments validate demand. Implication: positive for GPUs, networking, cooling and private credit over 2–8 quarters. Confidence: medium-high.
• Funding/execution skeptics — JPMorgan/Richard Choe. Uncontracted/speculative GPUs, delayed recognition, $40m–$50m sequential SG&A and the remaining funding gap create dilution/leverage risk. Confidence: high.

Resolver: Horizon acceptance, March-quarter revenue conversion, 2027 contracts, data-center finance terms and capex/MW.

5.15 Autodesk (ADSK): organic execution versus weak EPS translation

Facts. Revenue, EPS and billings beat; Q3 revenue was above Street but EPS below, while FY revenue rose and the FCF midpoint narrowed lower. Autodesk fell 4.8% after hours, nearly erasing its session gain. (Autodesk release)

• Organic productivity/vertical AI — Barclays/Saket Kalia, Baird/Joe Vruwink and management. Renewals, seller attainment and design/manufacturing growth support seat plus metered-machine usage. Confidence: high on execution, medium on AI monetization.
• Low-quality translation — Wolfe/Joshua Tilton and Jefferies/Brent Thill Q&A. MaintainX is loss-making; billings' headline raise equals acquisition contribution and AI workloads compress gross margin. Confidence: medium-high.

Resolver: Q4 EBA renewals, organic billings, Sept. 15–17 Autodesk University packaging and FY29 margin.

5.16 Workday (WDAY): measurable AI, unchanged growth regime

Facts. Revenue/EPS beat modestly; subscription revenue topped management's guide and cRPO rose 14.2%. Q3 cRPO slows to 11%–12%, preliminary FY28 subscription growth is about 11%, and the stock recovered an initial 5% after-hours decline. (Workday release)

• AI demand is measurable — Evercore/Kirk Materne and management. AI SKUs near $600m ARR and >25% of new ACV are unusually concrete. Risk: Flex credits are consumption-based and revenue lags. Confidence: high on bookings, medium on recognition.
• System-of-record durability, structurally slower SaaS — Goldman/Gabriela Borges. 97% retention supports durability, but 11% FY28 growth shows AI has not changed recognized growth. Confidence: high.

Resolver: Flex adoption, AI ARR-to-revenue, Q3 cRPO and Oct. 13 analyst-day guidance.

5.17 Ulta Beauty (ULTA): beauty resilience with flat transactions

Facts. $3.036bn sales, $6.55 EPS and 3.8% comps beat; FY sales/comps/EPS rose. But transactions were roughly flat, promotions increased and H2 comps slow to 2%–3%; shares fell 3.2% after hours. (Ulta release)

• Execution/share-gain bulls — Oppenheimer/Rupesh Parikh and Deutsche Bank/Krisztina Katai. Prestige share, controlled inventory and broad category strength validate the Unleashed plan. Confidence: high.
• Beat-quality skeptics — BofA/Lorraine Hutchinson, Loop/Anthony Chukumba, JPMorgan/Chris Horvers. Ticket/mix and promotion, not traffic, drove the beat; gross-margin upside remains limited. Confidence: high.

Resolver: Q3 transactions, gross margin excluding Space NK, promo cadence and digital fulfillment economics.

5.18 Malibu Boats (MBUU): channel normalization is not demand recovery

Facts. $295.5m revenue, $0.92 EPS and 1,456 units beat; every segment exceeded a small two-analyst unit panel. Dealer/aged inventory improved and FY27 revenue is $1.08bn–$1.12bn; shares gained 8.1%. (Malibu release)

• Channel normalization/share gains — Truist/Gregory Miller, B. Riley/Anna Glaessgen Q&A. Clean inventory and broad unit beats reduce destocking risk. Confidence: high.
• Two-speed consumer/Saxdor execution — KeyBanc/Noah Zatzkin, Baird/Craig Kennison. Premium demand holds; payment-sensitive buyers and acquired-growth margins remain weak points. Confidence: high on split, medium on Saxdor upside.

Resolver: retail registrations, financing conversion, wholesale/retail alignment and Saxdor margin toward 10%–11%.

5.19 Bilibili (BILI): ads carried the quarter; games carry Q4

Facts. Revenue and adjusted EPS modestly beat; ads grew 28%, while games fell 14%. BILI gained 3.8% in New York. (Bilibili release)

• Engagement/ad flywheel — management. Search doubled and time spent rose 14%, supporting share gains among young/high-intent users. Confidence: high on platform share, medium on China's macro ad pool.
• Macro/title-cycle skepticism — Morgan Stanley/Yang Liu. The firm retained Overweight but cut target/estimates; Q4 game recovery is a dated promise. Confidence: medium-high.

Resolver: Q3 ad growth, Sept. 17 Lumi Master retention and Q4 San Wang/game revenue.

5.20 RBC (RY): premium franchise, premium hurdle

Facts. Revenue and EPS beat; CET1 held 13.5%, but PCL rose to C$1.0bn and fee businesses supplied much of the upside. Both listings reversed an early gain and closed lower. (RBC release)

• Quality franchise — Jefferies/John Aiken. 17.9% ROE, record PPPT and capital generation justify a premium. Confidence: high.
• Good, not enough — First Avenue/Brian Madden and Scotiabank/Mike Rizvanovic. Personal Banking was light and valuations discount execution; episodic wholesale provisions remain. Confidence: high.

Resolver: Q4 retail NIM/operating leverage, impaired PCL, CRE/utilities formations and GTB/AI KPIs.

5.21 TD Bank (TD): the broadest Canadian bank beat

Facts. Adjusted EPS beat 12%; PCL of C$917m was materially favorable, every major segment grew and CET1 was 14.3%. TD outperformed peers. (TD release)

• Earnings/cost-control bulls — Jefferies/John Aiken. Positive operating leverage, U.S. NIM and >C$13bn potential FY27 buybacks support the best peer reaction. Confidence: high on Q3, medium on durability.
• Regulatory/credit skeptics — Canaccord/Matthew Lee and TD's guarded CRO framing. The U.S. consent order remains, branch approvals are conditional and tariff reserves may still be needed. Confidence: medium-high.

Resolver: asset-cap/consent milestones, FY27 expense/PCL, approved U.S. openings and actual capital return.

5.22 CIBC (CM): real beat, lower-quality mix

Facts. Revenue/EPS/PCL beat and operating leverage was positive, but Capital Markets rose 34%, performing provisions reversed in the U.S., expenses reaccelerate and CET1 fell q/q. Shares lost 2.8%. (CIBC release)

• Resilient bank — Jefferies. Broad segment growth and Canadian retail NIM support mid-teens ROE. Confidence: high.
• Quality/valuation skepticism — Jefferies and TD/Mario Mendonca's peer framing. Market-sensitive revenue, reserve release and wholesale growth receive a lower multiple. Confidence: high.

Resolver: Q4 expenses, U.S. NIM/deposits, impaired PCL near 37 bp and Capital Markets in a quieter tape.

5.23 Pernod Ricard: cash bridge versus structural U.S./China reset

Facts. Sales slightly missed, recurring operating profit was essentially in line and cash conversion improved to 91%. FY27 sales are broadly stable, but management assumes no U.S. market growth through FY29; shares fell 4.6%. (Pernod official results)

• Cyclical trough/cash bridge — Bernstein/Trevor Stirling; Jefferies and JPMorgan firm views. H2 improved, savings accelerate and India offsets part of destocking. Confidence: medium.
• Structural demand reset — RBC and AlphaValue/Théodore Duval-Segard. U.S. stagnation, China prestige weakness, low A&P and 3.7x leverage question brand economics. Confidence: high.

Resolver: Q1 U.S./China sell-in versus sell-out, Mid-Autumn Martell, A&P, FCF/leverage and FY28–29 growth arithmetic.

5.24 Prudential plc: cash/margins beat, growth arithmetic did not

Facts. NBP and APE were slightly soft, while NBP margin, OFSG, EPS, dividend and buyback were strong. Hong Kong, London and ADR listings all underperformed, including versus AIA. (Prudential results)

• Capital/quality bulls — Matt Britzman/HL, Richard Hunter/interactive investor, Henry Heathfield/Morningstar. Protection mix and cash generation support higher returns and ASEAN read-through. Confidence: high.
• China/growth skeptics — Britzman, Hunter, Christopher Akers/Investors' Chronicle. Double-digit FY growth requires acceleration amid mainland expense rules and cross-border uncertainty. Confidence: medium-high.

Resolver: Sep.–Dec. Hong Kong visitor sales/lapses, mainland par mix/margin, agents and buyback execution.

5.25 Mixue Group: network growth outran franchise throughput

Facts. Revenue rose 2.3%, profit fell 14.7% and stores rose 20.7%; a derived goods-sell-in-per-average-store proxy fell 17.2%. Overseas stores contracted, closure intensity rose and the stock fell 8.4% despite a special dividend. (Mixue filing)

• Scale/moat bulls — broad positive but stale broker set. Value pricing, vertical supply and RMB21.6bn liquidity preserve expansion optionality. Confidence: medium.
• Unit-economics bears — Haitong and JPMorgan pre-print stances. Store growth, cannibalization and quality/marketing spend are masking weaker throughput. Confidence: high.

Resolver: same-store GMV, franchisee payback/profitability, openings/closures, overseas net growth and 2H margin recovery.

5.26 China Life: franchise improvement inside an equity-market windfall

Facts. Premiums rose 2.2%, first-year regular premium 24.7%, NBV 33.7% and individual margin 600 bp. Profit surged 228.6%, but fair-value gains rose to RMB135.8bn; interim DPS increased 50.4%. (China Life filing)

• Structural value inflection — positive pre-print broker consensus. Regular-pay mix, agent productivity, persistency and lower guaranteed rates support durable NBV economics. Confidence: high.
• Peak earnings/equity beta — public FY forecast dispersion. H1 profit exceeded the stale FY mean because mark-to-market gains dominate; low rates remain a valuation sensitivity. Confidence: high.

Resolver: Aug. 28 webcast, normalized investment yield, H2 NBV/margin, equity allocation and dividend policy. Post-result analyst notes were not yet public.

5.27 China Pacific Insurance: operating quality versus investment yield

Facts. NBV rose 12.7%, margin 2.5 points, P&C combined ratio improved 1.3 points and profit rose 10.4%; CPIC initiated an interim dividend. Comprehensive yield fell 0.6 points and group solvency declined. (CPIC filing)

• Liability/underwriting quality bulls — management and constructive public pre-print forecasts. Regular premiums and underwriting profit are repeatable drivers. Confidence: high.
• Investment/capital skeptics — solvency filing and low-rate sensitivity. Weaker comprehensive yield and falling ratios limit valuation credit. Confidence: medium-high.

Resolver: 2H NBV margin, P&C catastrophe losses, normalized yield and solvency stabilization; same-night independent commentary was sparse.

5.28 CHALCO: exceptional smelter economics, weak alumina

Facts. Revenue rose 7.7%, profit 67.9% and gross margin to 26.0%; primary-aluminum pretax profit more than tripled, while alumina swung to loss. Interim DPS rose 124%. (CHALCO filing)

• Scarcity/cost bulls — 12-analyst pre-result Buy consensus. China's capacity ceiling plus centralized procurement support smelter spreads and cash return. Confidence: high on H1, medium on H2.
• Spread-normalization bears — Zhongyuan Futures. Alumina oversupply, inventories and Middle East restarts threaten peak economics. Confidence: medium-high.

Resolver: Q3 SHFE aluminum/alumina/power spreads, unit costs, inventories and dividend follow-through.

5.29 Tianqi Lithium: low-cost leverage meets a price reset

Facts. Revenue rose 153%, profit reached RMB4.24bn near the top of guidance and gross margin was 64.3%. H1 delivered 69% of a stale FY profit mean; cash improved and inferred net debt fell. (Tianqi filing)

• Resource/operating-leverage bulls — Dongguan Securities and management. Greenbushes, compound volume and China storage demand support margins. Confidence: high on cost position.
• Commodity-cycle skeptics — July spot-price evidence and consensus dispersion. SC6 fell materially from May, U.S. EV sales weakened and SQM contributes meaningful profit. Confidence: high on cyclicality.

Resolver: Aug. 28 briefing, realized prices/volumes, CGP3 ramp, SQM contribution and H2 price sensitivity.

5.30 Wynn Macau: mass-share gain versus hold and leverage

Facts. Revenue/EBITDA rose about 14%; mass and slots outgrew the market, VIP fell, and Palace dominated. Profit included a HK$691m derivative gain; DPS rose 20.5%, while net debt was HK$39.8bn. (Wynn Macau filing)

• Share-gain bulls — JPMorgan and Citi. Q2 GGR share reached a two-year high; mass drop and Palace support operational gains. Confidence: medium-high.
• Hold/capital skeptics — UBS. Above-normal hold, weak VIP/non-gaming and rising Enclave capex reduce extrapolation quality. Confidence: high.

Resolver: hold-normalized EBITDA/share, mass drop, Palace/peninsula mix and 2H free cash flow versus capex/dividend.

5.31 Haier Smart Home: HVAC stabilization, white-goods weakness

Facts. Revenue fell 2.8% and profit 14.3%; thin public consensus conflicted. Q2 revenue returned to slight growth, HVAC revenue/profit rose, but white-goods profit fell and FX swung sharply adverse. No interim dividend was proposed. (Haier filing)

• Stabilization/HVAC bulls — management. Commercial cooling, GE Appliances Q2 improvement and emerging markets point to selective strength. Confidence: medium-high.
• Core-demand/quality bears — same-day filing recaps. China demand, U.S. housing/tariffs, higher selling expense and one-off rebates challenge the rebound. Confidence: high.

Resolver: Q3 China sell-through, GE U.S.-dollar growth, HVAC margin without rebates, FX hedge and full-year target detail.

5.32 OOIL: Q2 freight acceleration versus H1 margin damage

Facts. Revenue rose 6.1% and liftings 5.2%, but profit fell 23.7% and liner gross margin compressed. Q2 revenue/TEU accelerated as ships filled; no credible half-year consensus existed. (OOIL release)

• Exit-rate bulls — management and higher-end pre-print targets. Fully loaded long-haul sailings and Q2 pricing support peak-season cash. Confidence: medium.
• Post-peak/cost bears — lower-end broker targets. Bunker/capacity costs and Hormuz-driven rerouting can lift rates and costs simultaneously; H1 profit conversion deteriorated. Confidence: high.

Resolver: Q3 revenue/TEU, bunker spread, FBX/WCI, utilization and dividend/cash conversion. Same-night post-result broker notes were unavailable.

5.33 China Vanke: extensions are not normalized credit

Facts. Revenue fell 33.4%, contract sales 48.2% and loss reached RMB14.95bn. Half of debt is due within a year against cash covering roughly 30%; Deloitte flagged material going-concern uncertainty. (Vanke filing)

• Marginal stabilization — management. Positive operating cash flow and service revenue provide a narrow cushion. Confidence: low-medium.
• Credit/liquidity bears — Fitch and public equity consensus. Bond extensions, Shenzhen Metro support and asset sales buy time, but sales track Fitch's -50% view and refinancing remains scarce. Confidence: high.

Resolver: the next extension's upfront cash, unrestricted cash, disposal proceeds, bank renewals and monthly sales/margin.

5.34 Akeso: financial miss, unchanged clinical convexity

Facts. Revenue of RMB1.806bn missed a thin one-analyst estimate 24%; Akeso lost RMB424m versus expected profit, though commercial sales grew and liquidity remained RMB9.16bn. (Akeso filing)

• Commercial-quality skeptics — the thin consensus and filing. Slower sales, gross-margin erosion and negligible license revenue deserve estimate cuts. Confidence: high on direction, medium on benchmark magnitude.
• Pipeline/licensing bulls — CMBI, Nomura and SPDB pre-print stances. Ivonescimab's global trials/PDUFA and Summit economics dominate valuation. Confidence: medium-high on catalysts, low on approval outcome.

Resolver: 2H product-level sales, margins, HARMONi-3 data and Nov. 14 FDA action.

5.35 China Communications Construction: overseas growth could not fix domestic cash

Facts. Revenue fell 1.2%, profit 24%, orders 8.9%; overseas orders rose 20.6% while domestic fell 16.4%. Net debt rose 20% to RMB659bn and operating cash stayed deeply negative. (CCCC filing)

• Overseas resilience — four-analyst Buy consensus and management. Better overseas growth/margin and a large backlog support activity. Confidence: medium-high.
• Fiscal-transmission/balance-sheet bears — Citi's Hold and official China investment data. Selective debt swaps are not broad infrastructure reflation; margins and collections remain weak. Confidence: high.

Resolver: 2H overseas gross profit, domestic order recovery, receivable collections and deleveraging toward an already-demanding FY consensus.

5.36 China Longyuan: bad wind exposed structural tariff pressure

Facts. Revenue fell 6.5%, profit 28%, wind output 6.7% and realized wind tariff 4.2%. A derived Q2 profit was about 43% below a thin public estimate; capex fell 53%. (Longyuan filing)

• Weather normalization bulls — positive pre-print forecasts. Better wind and offshore commissioning could restore fixed-cost leverage; cheaper funding helps. Confidence: medium.
• Structural squeeze — DBS Hold view. Tariff competition, curtailment and capacity growth faster than demand can keep profits weak even when weather improves. Confidence: high.

Resolver: wind hours, tariffs/curtailment, solar profit, H2 commissioning and whether low capex is discipline or slippage.

5.37 Weichai Power: volume strength, compressed engine economics

Facts. Revenue rose 8.8%, profit 36.5% and recurring profit 18.9%; engine/data-center-generator/truck volumes were strong, while engine segment margin compressed and investment gains boosted the headline. Public consensus feeds conflicted. (Weichai filing)

• Structural volume/KION bulls — CMBI. Truck replacement, exports, new energy, data-center backup power and KION normalization support growth/dividend. Confidence: medium-high.
• Quality/margin bears — filing economics and KION orders/cash. Competition/raw materials and nonrecurring income lower the quality of a 36.5% headline. Confidence: high.

Resolver: Aug. 28 estimate revisions, H2 engine margin/AIDC units, truck sales after subsidy and KION orders/FCF.

6. Cross-event themes and notable contradictions

1. AI demand is broad; its economics are concentrating. Nvidia lifted the indexes; capital-goods imports widened the trade deficit; Marvell, Elastic, Workday, SentinelOne, IREN, Autodesk and Weichai all reported tangible AI demand. Yet Marvell's mix pressure, IREN's $25bn–$30bn capex, Workday's unchanged 11% FY28 growth and Autodesk's AI gross-margin warning show that demand growth does not guarantee incremental return on capital.
2. Reported EPS was often the least useful number. Refunds distorted BBY, DG, DLTR and BURL; GAP's GAAP EPS contained a large tariff recovery; Affirm's included a deferred-tax release; Wynn/China Life/Weichai included market or investment gains. Clean operating bridges explained the tape far better than headline EPS.
3. The consumer is selective, not uniformly weak. DG/DLTR traffic and value trade-down were healthy; BBY's high-ticket demand worked around need/innovation; Ulta's beauty demand held; Malibu's premium consumer remained resilient. Conversely, BBW, BURL traffic, Hormel retail volume and Mixue store productivity showed clear pressure. The contradiction resolves by value proposition, urgency, novelty and starting income—not a single “consumer strong/weak” label.
4. Capital return could not fully offset quality risk. China Life, CPIC, CHALCO, Wynn, Mixue and several U.S./Canadian companies raised dividends/buybacks. Markets still discounted Mixue's unit economics, RY/CM's revenue mix and software companies' guide quality. Payouts supported valuation floors; they did not erase operating proof burdens.
5. The commodity complex split between scarcity and normalization. Hormuz restored an oil premium; U.S. gas storage tightened; CHALCO benefited from capped aluminum supply and Tianqi from lithium's H1 spike. OOIL, Hormel and the more cautious CHALCO/Tianqi clusters warned that fuel/logistics costs and price normalization can reverse the same earnings benefits.
6. China's policy transmission is selective. Weichai's trucks/generators and CHALCO's primary aluminum benefited from policy/AI/grid demand; CCCC domestic orders, Vanke sales, Longyuan tariffs and Haier white goods did not. This contradicts a broad China reflation thesis and favors bottom-up exposure to funded networks, replacement programs and export franchises.
7. Rates were a second-order event today—but not an irrelevant one. Claims and the seven-year auction did not reprice policy, while France/Canada supplied more idiosyncratic sovereign and FX signals. HealthEquity's cash-yield setup, insurer investment yields and Canadian bank NIM show how even a quiet rate tape continues to redistribute earnings.

7. Coverage audit

Calendars and source sets checked

• Macro calendars: New York Fed August indicator calendar, Econoday, Investing.com, Benzinga, Myfxbook, Newsquawk and TreasuryDirect auction schedule/results.
• Official macro sources: U.S. DOL, Census advance indicators, EIA storage, U.S. Treasury/TreasuryDirect, Statistics Canada, White House/USTR/Canada Finance, INSEE, Agence France Trésor, Eurostat and relevant exchange/market releases.
• Earnings inventories: Earnings Whispers' Aug. 27 before-open and after-close sets (40 names on the public snapshot), Kiplinger/Briefing, Reuters/LSEG/FactSet market wraps, SEC EDGAR, company IR feeds, HKEX title search and the issuers' results announcements.
• Reaction/opinion sources: company calls/transcripts, Reuters, AP, Dow Jones, Bloomberg-linked public recaps, public broker actions, firm research pages, rating-agency releases and exchange/quote histories. Private or paywalled notes were not claimed.

Borderline macro events excluded

• Jackson Hole opening/program: the symposium began, but no substantive Chair Warsh address occurred on Aug. 27; the speech is Aug. 28 and was only a positioning catalyst.
• Kansas City Fed manufacturing survey: composite 10 vs 9 prior and broadly in line; limited independent cross-asset reaction and lower information content than national data.
• Fed balance sheet: scheduled weekly update; no discrete surprise or same-day market impulse identified.
• ECB accounts / euro-area M3 and lending: informative for the policy backdrop but did not plausibly drive the day's U.S./global tape; France-specific fiscal risk had clearer impact.
• Germany GfK, Brazil current account/unemployment and other secondary releases: screened; surprises and observed cross-asset responses did not clear the materiality bar.

Earnings names screened but excluded

• PagerDuty, Domo, Ambarella, Shoe Carnival, Big Lots, Canada Goose, CooperCompanies and other calendar constituents: either sub-large-cap with no outsized move/sector read-through, or results did not create a material market signal by the cutoff.
• Duplicate/formal filings: names whose HKEX interim report repeated already-released operating results without a new call, guidance, balance-sheet or payout signal were excluded as duplicate events.
• Lower-impact HKEX filings: Mao Geping, China Merchants Port, Zoomlion, Shanghai Pharmaceuticals, Fosun, Beijing Enterprises, HashKey, Skyworth and several smaller issuers were screened. They lacked a combination of large capitalization, outsized contemporaneous move, broad sector read-through and usable new consensus/call evidence sufficient to outrank the 15 global filings covered in detail.

Data gaps and availability limits

• Public named post-result sell-side notes were sparse for many late U.S. prints and most Hong Kong post-close filings. In those sections, named call-question framing or dated pre-print recommendations are explicitly labeled and are not represented as post-result ratings.
• Aug. 27 cash reactions do not exist for HKEX filings posted after the Hong Kong close; “next-session” is stated rather than inventing an after-hours move. China Life's Aug. 28 webcast and several Asia briefings were not yet complete at the report cutoff.
• No credible public point-consensus existed for several segment metrics, half-year Hong Kong results, EIA storage (survey feeds diverged), or certain FCF/RPO lines. The report uses company guidance, year-on-year comparison or a clearly labeled thin/full-year proxy instead of false precision.
• Automated public transcripts can contain speaker-name errors; management claims and Q&A themes were cross-checked to issuer materials where possible.
• Market reactions around 8:30 a.m. macro releases were confounded by simultaneous claims/trade data, Nvidia and Jackson Hole positioning. Causality is described as muted/indeterminate where no price discontinuity was observable.

Completeness result

Covered in dedicated research: 8 qualifying macro events and 37 qualifying earnings/releases/calls. Every item has a primary-source fact set, market reaction or next-session status, opinion clusters, consensus/disagreement and explicit resolver. No qualifying category was empty.