Institutional Research

Global iPhone 18 Pro lead times held at 23 days while Pro Max lead times eased only one day to 29 days. China improved materially versus the prior-year pattern, with the Pro now ahead year on year and the Pro Max shortfall narrowing.

• Global Week 3 lead times averaged 23 days for iPhone 18 Pro and 29 days for Pro Max, versus 23 and 30 days in Week 2.
• The 18 Pro is modestly ahead of last year globally, at 23 versus 22 days; Pro Max is modestly below, at 29 versus 31 days.
• China lead times fell by only one day for each model; 18 Pro moved ahead of last year and the Pro Max gap narrowed.
• US and European lead times were unchanged from Week 2 and broadly in line with the prior year.

The report forecasts firmer core CPI readings in September and October, driven by energy-linked costs and a seasonal-adjustment distortion, followed by notably cooler November and December prints. Core PCE is expected to remain somewhat stickier than core CPI.

• Core CPI is forecast at 0.24% month-on-month in September and 0.22% in October, then 0.12% and 0.11% in November and December.
• Higher energy and commodity costs are expected to add 3-5bp to monthly core inflation over the next few months; September airfares are nowcast at +3%.
• Goldman Sachs estimates residual seasonality from shutdown-related missing data will add 5bp to October seasonally adjusted core CPI.
• The share of PCE components rising more than 3% annualized over six months is projected to decline from 49% to 43% by the December FOMC meeting.

UBS forecasts 10% average organic growth across its electrical equipment and multi-industry coverage universe and maintained or raised estimates for 75% of previewed companies. The report favors selected data-center, electrical and industrial-recovery beneficiaries while identifying supply-chain and margin pressure at several companies.

• Average organic growth is forecast at 10%, with two-year stacked growth accelerating by 100 basis points.
• UBS expects estimates to move higher for nearly the entire group despite rising oil prices and macro volatility.
• Companies with limited data-center exposure are holding 5% growth versus 2% a year earlier; UBS sees no slowdown in AI-industrial fundamentals.
• JCI, EMR, IR, MOD, VRT and NVT are among the more positive setups.
• Residential HVAC, SPXC and selected margin-sensitive names face near-term cost and supply-chain pressure.

Management highlighted improving US Branded Checkout execution, balance-sheet capacity for credit expansion, BNPL growth and a potential longer-term role in agentic commerce. UBS notes international softness, EU de minimis disruption, competitive intensity and execution risk alongside the company’s strategic opportunities.

• US Branded Checkout momentum is supported by co-marketing, biometrics, redesigned pay sheets and traction with roughly 40 large merchants.
• The EU de minimis change is expected to reduce Q3 2026 Branded Checkout growth by roughly 50-100bps, with adjustments potentially extending through year-end.
• PayPal generated roughly US$40bn of BNPL TPV in 2025; volume grew at roughly the mid-20% rate in 1H 2026.
• Management remains on track for roughly US$400m of gross run-rate savings exiting 2026 and at least US$1.5bn over the next two to three years.
• The board is assessing standalone and alternative value-creation outcomes for Checkout, Venmo and Braintree separately.

Bernstein argues that AI agent harnesses—rather than model intelligence alone—can create durable user stickiness by capturing context, orchestrating tools and improving models through task data. Tencent and Alibaba pursue distinct consumer-ecosystem and enterprise-lock-in routes, while independent labs risk becoming interchangeable model suppliers.

• Harnesses add persistent memory, tools, workflow orchestration and user-specific context to otherwise stateless AI models.
• Tencent's Workbuddy leads early engagement, while Xiaowei could connect Weixin intent, Mini Programs and Weixin Pay.
• Alibaba is targeting enterprise lock-in through Qwen Work, Dingtalk, vertical templates and Alicloud infrastructure.
• Bernstein sees Tencent and Alibaba as not pricing much AI optionality at roughly 11–12x 2027E P/E.
• DeepSeek Harness and ZCode ranked eighth and eleventh among OpenRouter coding-agent apps by trailing 30-day token usage.

AI investment and AI-related wealth are supporting consumption and capital spending, but UBS expects fading fiscal support, higher rates, weak non-AI investment and supply-driven inflation to make 2026-28 uneven.

• UBS forecasts real GDP growth of 2.2% in 2026, 1.9% in 2027 and 2.6% in 2028.
• AI/tech-related real equipment investment rose 20% over the past four quarters, while the rest of equipment investment fell 2%.
• UBS expects 25bp FOMC hikes in September and December 2026.
• Fiscal support from OBBBA is expected to peak in 2026 and turn into a drag in 2027.
• An AI-bust scenario would push unemployment close to 6%, inflation below 2%, and the policy rate back to the zero lower bound.

The report sees both Beijing and Washington seeking to preserve the status quo through 2026, while the PBOC signals greater vigilance against excessive RMB appreciation. Goldman Sachs expects both official manufacturing and non-manufacturing PMIs to rise in September.

• Both sides appear inclined to maintain the bilateral status quo through 2026.
• Rare earths and other critical minerals remain a source of Chinese leverage, with underlying tensions unresolved.
• The PBOC's policy stance was broadly unchanged, but its FX language points to concern about an RMB appreciation overshoot.
• Goldman Sachs forecasts the official manufacturing PMI at 50.4 in September, versus 49.8 in August.

The report expects strong year-on-year revenue growth through 4Q26, supported by AI server-rack ramp-up and PC product-mix upgrades. Higher earnings estimates raise the 12-month target price to NT$1,115, but Goldman Sachs maintains Neutral on valuation.

• September revenue is forecast at NT$98bn, up 19% year on year and 4% month on month.
• 3Q26 revenue is forecast to rise 8% quarter on quarter and 39% year on year; 4Q26 revenue is expected to decline 7% quarter on quarter but grow 27% year on year.
• 2026-28E revenue estimates rise 4% each year, while net-income estimates rise 4%, 7% and 4%, respectively.
• The 12-month target price rises to NT$1,115 from NT$1,109, based on 14.8x 2027E P/E.

MSCI China and CSI300 fell 0.8% and 1.5% during the week, while Goldman Sachs retains 12-month targets of 85 and 5,500. The report highlights trade-policy developments, favorable earnings and valuation metrics, and a preference for selected cyclical and growth-linked sectors.

• MSCI China and CSI300 declined 0.8% and 1.5% for the week.
• A 5-percentage-point reduction in US tariffs could lift Chinese-equity earnings by an estimated 0.3%.
• Goldman Sachs estimates that a one-standard-deviation decline in global policy uncertainty could raise MSCI China’s 12-month forward P/E by 5% over three months.
• Consensus expects 2026/27 EPS growth of 17%/18% for MSCI China and 25%/17% for CSI300.
• The report’s A-H rotation model indicates that H shares may modestly outperform A shares over the next three months.
• Materials, Tech Hardware, Capital Goods and Retailing are overweight sector allocations.

Goldman Sachs models a plausible, though non-base-case, 90-day US diesel export ban beginning in early October 2026. Inventory accumulation would initially pressure US diesel prices lower, but storage congestion and refinery run cuts could lift US gasoline prices and European diesel prices.

• Each week of a ban could initially lower US retail diesel prices by $0.25/gallon, or just under 4% of the current $6.5/gallon.
• National diesel storage could fill in 9-10 weeks under the stylized 1.6mb/d lost-export assumption.
• Once storage is full, each additional week could add $0.30/gallon of pressure to US retail gasoline prices.
• European wholesale diesel prices could rise by $3/bbl per week, although SPR releases could offset roughly half.
• The report reiterates long 2027 European gasoline positions as a geopolitical hedge.