2026-09-02 Daily Quick Read | Hilo Research
The current market presents a landscape where strong AI infrastructure investment coexists with diverging macroeconomic cycles. Companies such as NVIDIA, MediaTek, and Oracle have demonstrated better-than-expected demand and monetization capabilities in AI computing power, custom chips, and cloud transformation, while the rebound in China's manufacturing PMI and the resilience of U.S. auto sales provide fundamental support. However, semiconductor supply chain bottlenecks, rising memory costs, and potential pullback risks in U.S. industrial momentum require investors to allocate precisely between capex beneficiaries and consumption-stressed areas. The Goldman Sachs Conviction List emphasizes that earnings growth rather than valuation expansion is the core driver of current U.S. equities.
AI Infrastructure Demand and the Semiconductor Supply Chain
6 Related reports
Key views
NVIDIA management is highly confident in achieving 70% year-over-year growth in FY28; within the data center business, inference has already surpassed training, and demand is expanding from a few leading customers to a broad ecosystem including hyperscale cloud service providers, neoclouds, and enterprises. Without supply constraints, business growth could more than double.
Dell's F2Q27 AI server backlog reached 950 billion dollars, with single-quarter orders hitting a record 609 billion dollars, and ISG operating margin rose 630 basis points year over year to 15.0%, dispelling concerns of "revenue growth without profit growth" and proving that cloud, hybrid, and on-premises deployment scenarios are all investing heavily in AI.
Within Goldman Sachs' coverage, the median consensus EPS estimates for industrial technology companies for 2026 and 2027 were both raised by 3% after 2Q results, attributable to strong data center capex. However, tight data center supply chains and component availability may constrain delivery and earnings realization for companies such as Vertiv.
Advanced wafers and memory are the main supply-side constraints facing NVIDIA, and the current outlook is constrained by supply rather than demand. Goldman Sachs forecasts that DRAM and NAND will continue to face supply shortages from 2026 to 2028, with capacity expansion in China by CXMT and YMTC being a key uncertainty.
Goldman Sachs expects WFE to accelerate in 2027 after growing about 35% in 2026, with momentum lasting at least through 2028. Bernstein believes the continued shrinking of logic metal pitch ensures structural growth in demand for advanced lithography equipment, reiterating an Outperform rating on ASML with a €2,500 target price.
Goldman Sachs added Applied Materials (AMAT) to the Conviction List, believing that with over 65% of its business exposure in deposition and etch equipment, it will gain share in high-growth equipment segments; it also positioned Celestica (CLS) as a beneficiary of AI data center equipment, forecasting its 2027/28 EPS to be significantly above consensus.
Current market environment
AI infrastructure spending has spread comprehensively from hyperscalers to enterprise and hybrid cloud, as confirmed by Dell's record orders and high margins. However, capacity bottlenecks in advanced wafers and memory are capping chipmakers' revenue upside, making supply chain execution a key differentiator for gaining market share.
Future market changes
AI Computing Architecture Evolving from GPU Dominance to Diversified ASICs
Medium term
Triggers
- Increase in customized workloads
- Rising energy efficiency and cost requirements
Transmission channels
- Commercial solutions dominate -> ASIC and dedicated accelerator share rises -> Chip design engineer shortage intensifies -> Monetization opportunities for EDA vendors' Agentic AI emerge
Indicators to watch
- Rising share of ASIC revenue
- Growth in EDA vendors' AI-related revenue
Invalidation conditions
- GPU versatility continues to outweigh the cost advantages of custom chips
Ultra-high AI server margins recede as front-loading effects fade
Long term
Triggers
- Fading of AI spending front-loading effects
- End of the traditional server refresh cycle
Transmission channels
- Supply scarcity eases -> Price sensitivity rises -> Cost pass-through ability declines -> ISG operating margin correction
Indicators to watch
- Dell ISG quarterly margin declines quarter over quarter
- AlphaWise survey shows IT budgets normalizing
Invalidation conditions
- Enterprise AI demand continues to grow exponentially without cyclical fluctuations
Institutional disagreements
Sustainability of AI server margins
Different views
- Dell achieved simultaneous growth in scale and profit through supply scarcity and cost pass-through to enterprise customers, dispelling concerns of revenue growth without profit growth
- Once the AI spending front-loading effect fades, the current ultra-high margins brought by the traditional server refresh cycle may face a cyclical correction in 2027/2028
Opportunities and risks
Semiconductor Equipment and EDA Toolchain Upgrades
Consensus opportunityCapacity expansion in DRAM, advanced wafer foundry, and advanced packaging is driving accelerated WFE spending, and custom AI chips are exacerbating the design engineer shortage.
Potential beneficiaries
- ASML
- Applied Materials (AMAT)
- EDA vendors
Risks
- Export restrictions
- Rising share of domestic Chinese suppliers
Indicators to watch
- WFE annual growth rate exceeds 35%
- Emergence of EDA vendors' Agentic AI revenue
AI Data Center Hardware and Design Manufacturing
Consensus opportunityHyperscaler and enterprise AI capex continues, with surging demand for equipment such as 800G switches.
Potential beneficiaries
- Celestica (CLS)
- Dell
Risks
- Intensifying competition
- Supply constraints
- ATS demand volatility
Indicators to watch
- Rapid CCS revenue growth
- EPS beats consensus expectations
Related reports(6)
- NVIDIA CorporationJPMorgan · 2026-09-02
- F2Q27 Earnings – Another Blowout QuarterMorgan Stanley · 2026-09-02
- Top takeaways from 2Q26 earnings, plus key debates and areas of focus into the Communacopia + Technology conferenceGoldman Sachs · 2026-08-31
- What to expect at the Communacopia + Technology Conference 2026Goldman Sachs · 2026-09-01
- ASML: High-NA EUV Primer - The shrinking never stopsBernstein · 2026-09-01
- September 2026 Update - Refreshed Views for Back to SchoolGoldman Sachs · 2026-09-01
Custom Chip Partnerships and Strategic Semiconductor Financing
3 Related reports
Key views
MediaTek completed pricing of a 39 billion dollar overseas convertible bond, with NVIDIA investing 35 billion dollars and Alphabet participating. The funds are intended for procuring HBM, ABF substrates, and prepaying foundry capacity, and the two parties are expanding cooperation across cloud AI factories, on-premises AI computing, and automotive via the NVLink Fusion platform.
Bernstein interprets the MediaTek-NVIDIA partnership as a reference design aimed at enabling ASIC customers to move faster from chip design to rack-level systems and interoperate with NVIDIA GPUs, thereby lowering system integration barriers. This arrangement does not constitute material supplier financing, and MediaTek remains a neutral supplier.
Iluvatar CoreX Semiconductor's 1H26 gross margin was weaker than expected, and Morgan Stanley widened its 2026 loss per share forecast to Rmb1.03, but views this as "short-term pain" from strategic low pricing to secure CSP partnerships, raising its 2027 and 2028 EPS forecasts based on a strong order outlook.
Current market environment
Competition in the AI custom chip (ASIC) market is intensifying, and chip design companies are locking in capacity and expanding their customer base through strategic financing and deep technical ties. Some start-up or growth-stage semiconductor companies are willing to sacrifice short-term gross margins in exchange for long-term CSP orders and market share.
Future market changes
Explosive growth in MediaTek's ASIC revenue
Long term
Triggers
- Widespread adoption of NVLink Fusion platform reference designs
- Confirmed Google design wins materialize
Transmission channels
- Reference designs lower customer barriers -> Project win rates and margins improve -> ASIC revenue scale expands -> Financial benefits materialize in 2028 or later
Indicators to watch
- Number of new CSP ASIC customers
- Share of MediaTek's data center-related revenue
Invalidation conditions
- Reference designs fail to effectively convert into actual mass production orders
Opportunities and risks
ASIC Reference Designs and Custom Chip Services
Emerging opportunityStrategic investments from NVIDIA and Alphabet and the launch of the NVLink Fusion platform provide mid-sized chip design companies with a shortcut into top cloud vendors' supply chains.
Potential beneficiaries
- MediaTek
- Iluvatar CoreX Semiconductor
Risks
- Short-term gross margin pressure
- Rising dependence on major customers
Indicators to watch
- CSP partnership announcements
- New order conversion rate
Related reports(3)
- NVIDIA and Alphabet Invest in MediaTek’s CB for Extensive Partnership in ASIC ; Stay OW, Top PickMorgan Stanley · 2026-08-31
- Iluvatar CoreX Semiconductor Co., Ltd. | Asia PacificMorgan Stanley · 2026-08-31
- MediaTek: NVIDIA & Google becoming CB holders; "Reference design" to proliferate ASIC but co-existing with GPU?Bernstein · 2026-09-01
China Internet, Local Services, and AI Cloud Services
6 Related reports
Key views
Meituan's 2Q26 total revenue grew 14% year over year to Rmb104.6bn, and core local commerce operating profit grew 52% year over year to Rmb5.7bn, beating expectations. Food delivery unit economics turned positive to Rmb0.26 per order, and it is expected to maintain about 70% market share among delivery orders with an average order value above Rmb30.
Morgan Stanley expects Meituan's 3Q26 core local commerce operating profit to fall to Rmb4.2bn and food delivery unit economics to drop to Rmb0.1 per order, mainly due to seasonally higher rider costs and shifts in spending timing. New initiatives losses narrowed, the Saudi business reached breakeven, and Keeta International shifted to prioritizing profitability improvement.
Alibaba Cloud currently holds about 40% of China's AI cloud market share. Management stated there is a clear three-year return path for AI infrastructure investment, expressing confidence in achieving RMB30bn in MaaS annual recurring revenue in FY27 and US$100bn in external cloud revenue in 2030.
Tencent management reiterated that AI is the highest capital allocation priority, expecting 3Q capex plus prepayments to remain at a high level, aimed at locking in favorable memory prices and computing power economics to support Hunyuan model development and WorkBuddy commercialization.
Kuaishou's Kling plans a major model upgrade, with about two-thirds of the 2026 RMB26bn capex allocated to Kling; next year it expects to independently obtain additional computing power through leasing to ease cash flow pressure.
Range Intelligent's 2Q26 adjusted EBITDA grew 84% year over year to Rmb1.43bn, beating expectations; however, AIDC revenue accounted for 53% of total revenue, 11% below expectations, reflecting a cautious investment strategy amid high GPU costs. Batam is expected to deliver 60MW of liquid cooling capacity in 4Q26, and overseas IDCs are viewed as a seller's market.
Bernstein, citing Ramp data, noted that heavy AI adopters on average grew headcount more than 10% more than lower-adoption peers, and by the 24th month after initial adoption company size was over 50% larger, indicating AI productivity gains are reinvested into growth rather than pure layoffs, supporting the long-term sustainability of per-seat pricing models for application software.
Current market environment
Core profitability of China's local services giants such as Meituan is recovering, but they face pressure from seasonal fluctuations and share loss due to long-term competition. Meanwhile, China's tech giants (Alibaba, Tencent, Kuaishou) view AI as the highest capital allocation priority, locking in computing power and cloud market share through heavy capex, and AI applications are substantively driving corporate scale expansion rather than replacing jobs.
Future market changes
Full-Stack Providers Win in China's AI Cloud Market
Long term
Triggers
- Increasing fragmentation at the model layer
- Surge in enterprise AI demand
Transmission channels
- Full-stack AI providers demonstrate advantages -> Alibaba Cloud consolidates 40% share -> MaaS annual recurring revenue targets met -> External cloud revenue advances toward one hundred billion dollars
Indicators to watch
- Alibaba Cloud MaaS ARR reaches RMB30bn
- Changes in AI cloud market share
Invalidation conditions
- Open-source models or single-function AI tools disrupt full-stack service models
Slow erosion of Meituan's long-term market share
Medium term
Triggers
- Intensifying competition in food delivery and instant retail
Transmission channels
- Competitor subsidies or model innovation -> Meituan's long-term share falls from 70-75% to above 65% -> Unit economics under pressure -> Valuation multiples constrained
Indicators to watch
- Share changes for orders above Rmb30 average order value
- Pace of narrowing losses in new initiatives
Invalidation conditions
- Meituan successfully builds insurmountable fulfillment network barriers
Opportunities and risks
Overseas Data Center Expansion
Emerging opportunityOverseas IDCs are viewed as a seller's market, with expected margins higher than domestic ones, and liquid cooling capacity is about to be delivered.
Potential beneficiaries
- Range Intelligent
Risks
- High GPU costs
- Project phasing delays
Indicators to watch
- Batam 60MW capacity delivered on schedule
- Rising share of overseas revenue
AI-Driven SaaS Seat Growth
Emerging opportunityData proves AI adopters are increasing headcount, refuting the hypothesis that AI reduces white-collar employment.
Potential beneficiaries
- Enterprise application software companies
Risks
- Macroeconomic weakness curbs corporate hiring
Indicators to watch
- Net employee growth rate of heavy AI-adopting enterprises
Related reports(6)
- Morgan Stanley_Meituan(3690.HK)Non~LinearRecovery_20260831Morgan Stanley · 2026-08-31
- Tencent Holdings (0700.HK): Asia Leaders Conference 2026 Takeaways: The Road from Investment to Returns; BuyGoldman Sachs · 2026-09-01
- Alibaba Group (BABA): Asia Leaders Conference 2026 Takeaways: Accelerating cloud with high AI capex return visibility; BuyGoldman Sachs · 2026-09-01
- Range Intelligent (300442.SZ): First Take: 2Q26 EBITDA beat yet sales slightly below; Disciplined AIDC investments against elevated costs; BuyGoldman Sachs · 2026-08-31
- U.S. SMID-Cap and Global SoftwareBernstein · 2026-09-01
- Kuaishou Technology | Asia PacificMorgan Stanley · 2026-08-31
Consumer Electronics Innovation Cycle vs. Memory Costs
2 Related reports
Key views
Bernstein data shows cumulative sales of the iPhone 17 series from 9 to 7 were 15% higher than the iPhone 16 series, with base models up 34% and Pro Max up 22%; Apple has achieved year-over-year sales growth for 17 consecutive months, demonstrating continued gains in high-end smartphone market share.
Morgan Stanley expects Apple to launch its first foldable iPhone on 9 9, with the 512GB version expected to start at 2,399 dollars, C2H26 production of 700-800 million units, contributing about 140 billion dollars in revenue. The iPhone 18 Pro series will adopt TSMC's 2nm process A20 Pro processor to support on-device AI.
Memory has become a constraint on Apple's iPhone production. C1H26 has already consumed nearly 50% more DRAM year over year, and if the bull case plan is fully executed, C2H26 DRAM consumption would increase about 15% year over year. Driven by this, the iPhone 18 Pro starting price is expected to rise by 200 dollars to maintain a gross margin of about 40%.
Bernstein notes that rising memory prices may suppress overall smartphone demand and reduce future increases in iPhone DRAM content, although Apple is better able to absorb costs than other OEMs thanks to its profitability.
Current market environment
Apple is at the beginning of a supercycle driven by strong iPhone 17 sales and the upcoming foldable form factor innovation. However, the sharp increase in DRAM and NAND consumption from on-device AI, combined with tight global memory supply and demand, is forcing OEMs to protect gross margins through price increases, which constitutes a potential headwind for shipment growth.
Future market changes
Foldables and On-Device AI Push Up Smartphone ASPs
Short term
Triggers
- Foldable iPhone launch in 9
- Mass production of 2nm A20 Pro processor
Transmission channels
- Form factor innovation and computing power upgrades -> Consumers accept higher prices -> Significant ASP increase -> Offsets memory cost inflation and maintains 40% gross margin
Indicators to watch
- Foldable shipments reach 650 million units in the 12 quarter
- Pre-order data for the iPhone 18 Pro series after price increases
Invalidation conditions
- Demand elasticity to the 2,399 dollar starting price is too high, leading to order cuts
Institutional disagreements
Extent of memory price increases' impact on end demand
Different views
- Apple can fully offset the memory cost burden by raising prices 200-500 dollars, maintaining a 40% gross margin
- Rising memory prices will inevitably suppress overall smartphone demand and limit future increases in DRAM content
Opportunities and risks
High-end smartphones see both volume and price rise
Consensus opportunityStrong iPhone 17 sales lay the foundation, while foldables and 2nm chips provide clear upgrade reasons.
Potential beneficiaries
- Apple
- TSMC
Risks
- Memory costs spiral out of control
- Decline in consumer purchasing power
Indicators to watch
- Consecutive months of year-over-year sales growth
- Revenue share of foldable models
Related reports(2)
- iPhone 18 Launch Preview – Ternus, Foldables, and PricingMorgan Stanley · 2026-09-02
- APPLE TRACKER (July): iPhone demand remains strong with sellthru units +14% YoY and share gains continuingBernstein · 2026-08-31
Enterprise Software Cloud Transformation and Database Monetization
2 Related reports
Key views
Oracle's cloud business mix is shifting significantly toward OCI, with OCI IaaS/PaaS revenue reaching 181 billion dollars in FY26, up 77% year over year. Bernstein estimates FY26 OCI AI revenue at about 66 billion dollars, and it has secured over 3000 billion dollars in AI remaining performance obligations (RPO), with contracts typically spanning multiple data centers and being non-cancellable.
Oracle's multi-cloud database business grew 531% year over year in Q3 FY26. Migrating database customers to multi-cloud environments can deliver more than 3x revenue uplift relative to database support revenue, reduce churn risk, and this revenue may carry SaaS-like high gross margins (about 80%).
Goldman Sachs added Microsoft (MSFT) to the Conviction List, based on Azure capacity and monetization, enterprise AI integration, and accelerating Copilot adoption; paid Copilot seats exceeded 3,000 million at the end of 6, with net additions of about 1,000 million in the quarter. Under a reasonable upside scenario, FY28 operating profit could have 5% upside versus market expectations.
Current market environment
Traditional software and database giants are successfully transforming through non-cancellable long-term AI data center contracts and multi-cloud deployment strategies. Oracle's AI infrastructure business has grown from nothing to become a core driver, while the rapid expansion of Microsoft's Copilot seats proves the viability of enterprise AI monetization.
Future market changes
These reports do not specify a future scenario.
Opportunities and risks
Multi-Cloud Database Migration and SaaS Monetization
Emerging opportunityThe multi-cloud database business achieved explosive growth of 531%, proving that cross-cloud deployment can not only defend against substitution risk but also deliver a structural leap in revenue and margins.
Potential beneficiaries
- Oracle
- Traditional database vendors
Risks
- Exclusive competition among cloud vendors
- Migration implementation complexity
Indicators to watch
- Multi-cloud database revenue growth rate
- Customer retention rate and gross margin levels
Enterprise AI Copilot scaling
Consensus opportunityPaid seats saw a net increase of tens of millions in a single quarter. Supply constraints may reduce the near-term predictability between capital expenditure and Azure growth, but the monetization trend is clear.
Potential beneficiaries
- Microsoft (MSFT)
Risks
- Copilot execution fell short of expectations
- Chip ramp-up delays
- Custom software substitution
Indicators to watch
- Total paid Copilot seats
- Azure revenue growth rate
Related reports(2)
- Oracle FY26: What is really going on under the hood - our latest updateBernstein · 2026-09-02
- September 2026 Update - Refreshed Views for Back to SchoolGoldman Sachs · 2026-09-01
Macro divergence, manufacturing sentiment, and the auto cycle
6 Related reports
Key views
Morgan Stanley believes the U.S. economic cycle is increasingly divergent: the industrial side is strong (7 month durable goods orders +1.1% MoM), driven by reshoring manufacturing, AI, and fixed asset investment; however, housing data lacks conviction (7 month housing starts -13.5% YoY), and consumption indicators are weak. The 8 month Chicago PMI fell to 47.1, entering contraction territory, and tariff-related channel inventory building may make current industrial data stronger than underlying demand.
The 8 month RatingDog China manufacturing PMI headline rose to 51.5, with output, new orders, and new export orders improving simultaneously. However, the pricing environment is unfavorable for profits: the input price index rose to 52.2, while output prices fell to 49.6, leaving manufacturers facing a "scissors gap" caused by rising upstream costs and intense downstream competition.
U.S. light vehicle sales SAAR in 8 month was approximately 1680 ten thousand units, above consensus expectations, implying upside risk to the full-year forecast of 1600 ten thousand units. However, internal divergence in electrification is severe: BEV sales fell about 48% YoY, hybrid sales rose about 21% YoY, and Tesla's estimated sales fell 31% YoY.
Competition among the Detroit three automakers for pickup truck market share has intensified, with GM and Ford losing large pickup share while Stellantis's share rose to 21.5%. Goldman Sachs expects U.S. auto sales in 2026 to decline slightly YoY, and global light vehicle production is expected to fall 2-3%.
Global fuel, LNG, and oil inventories have all declined year-to-date, with Asian JKM LNG prices holding around US$20/mmbtu. Morgan Stanley expects the global LNG market to remain undersupplied until 2H26, prompting Indian city gas distributors to further raise CNG prices.
Current market environment
The global macroeconomy exhibits significant structural divergence. Manufacturing activity in both China and the U.S. shows signs of stabilization, but both face margin squeeze issues—the U.S. constrained by the unsustainability of tariff front-running effects, and China by the input-output price scissors gap. Total auto market volume shows resilience, but the pure electric transition faces cyclical headwinds, making hybrids and traditional ICE pickups the focus of competition. Energy markets remain supply-tight due to inventory destocking.
Future market changes
Pullback in U.S. short-cycle industrial momentum
Short term
Triggers
- End of tariff-related channel inventory building and order front-loading
Transmission channels
- Front-running effect fades -> true demand exposed -> PMI contracts further -> industrial sector earnings estimates revised down
Indicators to watch
- Chicago PMI consecutively below 50
- Durable goods orders excluding defense and transportation declined
Invalidation conditions
- U.S. fixed asset investment and AI capex are sufficient to fully offset the fading of tariff front-running
Institutional disagreements
Sustainability of the U.S. industrial recovery
Different views
- Industrial strength is driven by reshoring manufacturing, AI, and fixed asset investment, with fundamental support
- Current industrial data is stronger than underlying demand, raising doubts about the sustainability of short-cycle momentum once tariff front-running effects fade
Opportunities and risks
Revaluation of energy and LNG infrastructure cash flows
Emerging opportunityA global LNG supply shortage is pushing up gas prices, with floating LNG vessels being deployed under long-term contracts. Goldman Sachs added Golar LNG (GLNG) to its Conviction List, expecting contracted EBITDA to grow more than fourfold to approximately 1,200mn dollars by 2030; ConocoPhillips (COP) is expected to shift from heavy investment to a harvest phase.
Potential beneficiaries
- Golar LNG (GLNG)
- ConocoPhillips (COP)
Risks
- Project ramp-up and execution risks
- Commodity price volatility
Indicators to watch
- Commercialization progress of GLNG's fourth vessel
- COP free cash flow inflection point
Transitional dividend from hybrid models
Consensus opportunityBEV demand faces headwinds, while hybrid model sales grew against the trend by 21%, becoming consumers' top choice.
Potential beneficiaries
- Traditional automakers (Toyota, Ford, GM, etc.)
Risks
- Policy tilts back toward pure electric
- Sharp declines in battery costs erode hybrid advantages
Indicators to watch
- HEV/PHEV sales mix continues to rise
Related reports(6)
- Inventories Unwinding = More Fuel Price HikesMorgan Stanley · 2026-08-31
- US AUTOS & INDUSTRIAL TECHGoldman Sachs · 2026-09-01
- CoTW: 2H26 Durability Questions Grow as August Chicago PMI Enters Contraction TerritoryMorgan Stanley · 2026-08-31
- Top takeaways from 2Q26 earnings, plus key debates and areas of focus into the Communacopia + Technology conferenceGoldman Sachs · 2026-08-31
- China: Unofficial manufacturing PMI rose in AugustGoldman Sachs · 2026-09-01
- September 2026 Update - Refreshed Views for Back to SchoolGoldman Sachs · 2026-09-01
Insurance industry earnings leverage and financial/healthcare recovery
3 Related reports
Key views
Morgan Stanley notes that all covered Hong Kong and China insurers achieved double-digit earnings growth in 1H26, benefiting from stock market rallies, with annualized operating ROE maintaining a healthy level of 14-21%. Most insurers achieved 11-34% VNB growth, but only AIA China and CPIC posted positive 2Q growth, reflecting diverging channel strategies.
Goldman Sachs raised China Life's FY26-28E net profit forecasts by 2-4%, reflecting strong investment income in 1H26; however, it simultaneously lowered shareholder equity forecasts by 5-6% due to unrealized equity investment losses and mark-to-market changes in liabilities. The two offset each other, and the Neutral rating is maintained.
Goldman Sachs added UnitedHealth Group (UNH) to its Conviction List, forecasting average annual EPS growth of 17%+ through 2028, and expecting adjusted EPS growth of about 23% in 2026 (versus a 40% decline in 2025), based on Medicare Advantage margin recovery, portfolio rationalization, and cost efficiency improvements. It also added Citizens Financial (CFG), expecting it to reach the lower end of its ROTCE target range in 2H27.
Current market environment
Against the backdrop of recovering capital markets, the insurance industry has demonstrated strong earnings leverage, with investment income boosting reported profits, though balance sheets are pressured by market volatility. Life insurance new business value growth has begun to slow from a high base, and differences in channel strategy determine relative performance. Meanwhile, managed care and regional banks are undergoing a recovery cycle as margins bottom out and rebound.
Future market changes
These reports do not specify a future scenario.
Opportunities and risks
Managed care earnings cycle recovery
Emerging opportunityAfter scaling back low-return businesses, UNH is seeing Medicare Advantage margin recovery, with 2026 EPS poised for a strong 23% rebound.
Potential beneficiaries
- UnitedHealth Group (UNH)
Risks
- Persistently high utilization rates
- Regulatory scrutiny
- Medicaid margin pressure
Indicators to watch
- Medicare Advantage margins improve QoQ
- Adjusted EPS growth delivered
Related reports(3)
- China Life Insurance Co. (2628.HK): Updating post 1H26 resultsGoldman Sachs · 2026-08-31
- 1H26 Wrap-up: Strong Results as Expected, while Several Nearterm Watchpoints RemainMorgan Stanley · 2026-08-31
- September 2026 Update - Refreshed Views for Back to SchoolGoldman Sachs · 2026-09-01
Advanced manufacturing, medical devices, and battery equipment cycles
2 Related reports
Key views
Wuxi Lead Intelligent's 2Q26 revenue grew 28% YoY, broadly in line with expectations, but gross margin fell 2.5 percentage points YoY to 30.6% due to raw material inflation and higher transportation costs. New orders in 2Q exceeded Rmb9bn, up over 40% YoY, with about 80% from lithium battery equipment. Management expects this upcycle may last 3-5 years.
Mindray Bio-Medical's 3Q26 revenue growth is expected to slow to low single digits compared to 2Q, due to a higher comparison base in the China market and pressure from declining gross margins and FX losses. However, against the backdrop of negative growth in China's overall IVD market, 1H26 China IVD revenue grew over 10%, with the goal of increasing combined market share from 14% to 20%, and an ultimate target for overseas revenue contribution of 70%.
Current market environment
China's advanced manufacturing sector is characterized by strong orders but pressured margins. The lithium battery equipment industry is in the early stage of an upcycle, with leading customers' capacity utilization near full load supporting expansion willingness; medical device leaders are growing against the trend driven by import substitution and overseas expansion, but need to digest quarterly volatility from high bases and cost inflation in the short term.
Future market changes
These reports do not specify a future scenario.
Opportunities and risks
Lithium battery equipment upcycle and domestic substitution
Consensus opportunityNew orders grew over 40% YoY and the cycle is expected to last 3-5 years, with overseas demand providing incremental growth.
Potential beneficiaries
- Wuxi Lead Intelligent
Risks
- Raw material inflation continues to squeeze gross margins
- Downstream battery plant overcapacity
Indicators to watch
- Quarterly new order growth rate
- Gross margin stabilizes and recovers
Medical device overseas expansion and share gains
Emerging opportunityAchieved over 10% counter-trend growth amid negative growth in the domestic IVD market, with accelerating overseas recognition for hematology and ultrasound series.
Potential beneficiaries
- Mindray Bio-Medical
Risks
- Domestic policy headwinds
- Exchange rate fluctuations
- Overseas expansion falls short of expectations
Indicators to watch
- Rising share of overseas revenue
- Biochemistry and CLIA market share advancing toward 20%
Related reports(2)
- Morgan Stanley_MindrayBio~Medical(300760)2Q26EarningsCallTakeaways~RecoveryOnTrack,but3QtoSlowonATougherBase_20260831Morgan Stanley · 2026-08-31
- 2Q26 Earnings Largely in-line; Strong Order OutlookMorgan Stanley · 2026-08-31
Humanoid robots and autonomous driving commercialization
2 Related reports
Key views
Global humanoid robot shipments reached 19,000 units in 1H26, up 272% YoY, with Chinese companies accounting for 97%. Morgan Stanley expects China's shipments to reach 50,000 units in 2026E, and by 2050, global cumulative adoption will reach 10.19 hundred million units, with annual market revenue reaching 7.5 trillion dollars.
Morgan Stanley defines the industry's current state using Moravec's paradox: locomotion capabilities have improved significantly, but perception, reasoning, and dexterous manipulation in variable physical environments remain difficult. Training data is seen as the core bottleneck in global competition, with multiple players targeting 100 ten thousand usable data hours by the end of 2026 or before the end of 2027.
Based on NHTSA data, Goldman Sachs estimates that Tesla experiences one accident per 5 ten thousand to 7 ten thousand miles in fully driverless Robotaxi operations, while Waymo has one per 16 ten thousand to 18 ten thousand miles. The Waymo app's MAU growth accelerated to 54% YoY in 7 month, accounting for roughly 12%-13% of U.S. ride-hailing app downloads.
Ride-hailing platforms are building demand and fleet partnership layers around AV supply. Uber has committed to investing over 100 billion dollars in the future in AV equity investments, infrastructure, and vehicle offtake. Goldman Sachs believes Uber and Lyft can benefit by aggregating AV supply, becoming asset-light third-party marketplace platforms.
Current market environment
The humanoid robot industry is moving from proof-of-concept to early commercialization, with China holding absolute dominance on the manufacturing side, but breakthroughs in embodied intelligence depend on the accumulation of massive training data. In autonomous driving, Robotaxis are building an early user base, with safety metrics remaining the core threshold for regulatory approval, while traditional ride-hailing platforms are transforming into asset-light AV demand aggregators through capital commitments.
Future market changes
Humanoid robot cost reduction and scaled deployment
Long term
Triggers
- Training data breaks through the one million hour threshold
- Embodied AI models effectively transfer to real-world tasks
Transmission channels
- Perception and dexterous manipulation capabilities improve -> production efficiency validated -> ASP drops from 20 ten thousand dollars to 7.5 ten thousand dollars -> trillion-level end market opens up
Indicators to watch
- Annual shipments reach the hundreds of thousands of units level
- Real-world task success rates improve
Invalidation conditions
- The cognitive and manipulation bottlenecks in Moravec's paradox cannot be broken through with existing AI architectures
Opportunities and risks
Embodied intelligence training data and computing power infrastructure
Emerging opportunityData output and diversity have become core bottlenecks, with Figure planning to invest over 10 hundred million dollars in the next 12 months.
Potential beneficiaries
- Computing power providers
- Data collection and annotation companies
- Leading robotics companies
Risks
- Data quality fails to meet standards
- Payback period for computing power investment is too long
Indicators to watch
- Usable data hours reach the million-level target
Ride-hailing platforms' AV aggregation model
Emerging opportunityUber has committed over ten billion dollars to the AV ecosystem, and partners have already committed approximately 12 ten thousand vehicles, reducing its own R&D risk.
Potential beneficiaries
- Uber
- Lyft
Risks
- Robotaxi safety incidents lead to regulatory halts
- AV technology iteration slower than expected
Indicators to watch
- Number of AV fleets connected
- Increase in Robotaxi order share
Related reports(2)
- Humanoid Horizons: Moravec's ParadoxMorgan Stanley · 2026-09-01
- US AVs: Robotaxi and AV Tracker - Deployments update and key safety and usage metricsGoldman Sachs · 2026-09-01
Space launch infrastructure and orbital computing
1 Related reports
Key views
Bernstein believes SpaceX's valuation is highly dependent on achieving high Starship launch frequency and full reusability. The planned Starbase Louisiana covers 125,000 acres and is ultimately expected to have at least 10 launch pads, supporting over 30 flights per day, serving as critical infrastructure needed for orbital data centers.
Starlink's aviation business is the clearest recent evidence of commercial leadership, with 46 airlines already signed. Starlink has over 11,000 satellites in orbit, whereas Amazon's competing LEO constellation has only 396, which is insufficient to operate services.
Bernstein intentionally adopts a slower launch trajectory than SpaceX management; if each launch pad launches 3 times per day, all 15 launch pads could support far more than 15,000 launches annually, broadly consistent with SpaceX's forecast of launching approximately 50 GW of orbital data center capacity in 2031.
Current market environment
Commercial space is transitioning from launch capability building to vertical monetization (such as in-flight internet). SpaceX holds an absolute lead thanks to Starlink's network effects and Starship's payload potential, but institutions remain highly vigilant about schedule delay risks in space projects when valuing them.
Future market changes
These reports do not specify a future scenario.
Related reports(1)
- SpaceX: Hot topics - Louisiana Starbase and five othersBernstein · 2026-09-01
Regulatory litigation and supply chain vulnerabilities
2 Related reports
Key views
The FTC and attorneys general from 22 states filed a lawsuit against Amazon, alleging it introduced undisclosed "soft floors" in Sponsored Products auctions, interfering with 70-80% of auctions and causing advertisers over 200 hundred million dollars in losses since 2018. Amazon denies this. The lawsuit involves its advertising business, which totaled approximately 2600 hundred million dollars cumulatively from 2020-2025.
Morgan Stanley warns that the deterioration in container shipping and car carrier supply-demand may last longer than reflected in share prices and faces cost inflation risks, thus maintaining Underweight ratings on Nippon Yusen, Mitsui O.S.K. Lines, and Kawasaki Kisen.
Current market environment
Large tech platforms and cyclical transport industries are facing downside risks in their respective dimensions. Amazon's advertising monetization model is directly challenged by antitrust regulation, while the shipping industry's cyclical downside risks and cost inflation have not yet been fully priced in by the market.
Future market changes
These reports do not specify a future scenario.
Related reports(2)
- TransportationMorgan Stanley · 2026-08-31
- FTC Files Lawsuit re: Amazon’s Advertising BusinessGoldman Sachs · 2026-08-31
Valuation framework adjustments and institutional conviction lists
11 Related reports
Key views
Goldman Sachs updated its U.S. Conviction List in 9 month, adding Vertex Pharmaceuticals (VRTX) and removing Interactive Brokers. The list contains 23 stocks with "Buy" ratings, with median target price upside of 27% and median FY2 revenue growth of 6%. The report emphasizes that the S&P 500 is up 12% YTD but still maintains a 20x NTM P/E, with the stock market driven by earnings rather than valuation expansion.
Morgan Stanley cut Meituan's target price by 8% from HK$120 to HK$110, based on a DCF base case corresponding to 19x 2027E P/E, maintaining Overweight but preferring Alibaba. It maintained Mindray at Overweight with a Rmb210 target price; for Wuxi Lead, it applied a 2.5x market cap/new orders multiple to maintain a Rmb54 target price.
Morgan Stanley maintained MediaTek at Overweight and Top Pick, with a target price of NT$5,588 (implying 42% upside), using a residual income model. It lowered Iluvatar CoreX's target price to HK$668, applying 50x 2026e P/S in the base case. Goldman Sachs maintained Tencent's HK$670 and Alibaba's US$186 SOTP target prices.
Goldman Sachs maintained Range Intelligent at Buy with a Rmb117 target price, applying a 18x EV/EBITDA discount to 2030E. Bernstein reiterated ASML at Outperform with a €2,500 target price, based on a P/E multiple of 40x Q5-8 EPS estimates. Goldman Sachs fine-tuned China Life H-shares' target price to HK$32.0, using a 12-month ROA valuation.
Current market environment
Institutions show highly targeted valuation methods: applying SOTP or forward EV/EBITDA discounts for AI and cloud businesses to capture option value, order multiples for cyclical manufacturing, and DCF or normalized P/E for mature tech. Goldman Sachs' Conviction List update highlights a bottom-up strategy of seeking earnings beat candidates during a valuation plateau (such as VRTX's pipeline value and AMAT's equipment share).
Future market changes
These reports do not specify a future scenario.
Opportunities and risks
Healthcare pipeline and biotech innovation premium
Emerging opportunityVRTX was added to the Conviction List, possessing up to five multi-billion-dollar growth pillars, with catalysts such as the IgAN PDUFA approaching and not yet fully factored into market models.
Potential beneficiaries
- Vertex Pharmaceuticals (VRTX)
Risks
- Clinical setbacks
- Regulatory approval delays
- Pricing or reimbursement pressure
Indicators to watch
- IgAN PDUFA results on 11 month 30 day
- Phase 2/3 data readout
Related reports(11)
- Morgan Stanley_Meituan(3690.HK)Non~LinearRecovery_20260831Morgan Stanley · 2026-08-31
- Morgan Stanley_MindrayBio~Medical(300760)2Q26EarningsCallTakeaways~RecoveryOnTrack,but3QtoSlowonATougherBase_20260831Morgan Stanley · 2026-08-31
- 2Q26 Earnings Largely in-line; Strong Order OutlookMorgan Stanley · 2026-08-31
- September 2026 Update - Refreshed Views for Back to SchoolGoldman Sachs · 2026-09-01
- China Life Insurance Co. (2628.HK): Updating post 1H26 resultsGoldman Sachs · 2026-08-31
- ASML: High-NA EUV Primer - The shrinking never stopsBernstein · 2026-09-01
- NVIDIA and Alphabet Invest in MediaTek’s CB for Extensive Partnership in ASIC ; Stay OW, Top PickMorgan Stanley · 2026-08-31
- Tencent Holdings (0700.HK): Asia Leaders Conference 2026 Takeaways: The Road from Investment to Returns; BuyGoldman Sachs · 2026-09-01
- Alibaba Group (BABA): Asia Leaders Conference 2026 Takeaways: Accelerating cloud with high AI capex return visibility; BuyGoldman Sachs · 2026-09-01
- Range Intelligent (300442.SZ): First Take: 2Q26 EBITDA beat yet sales slightly below; Disciplined AIDC investments against elevated costs; BuyGoldman Sachs · 2026-08-31
- Iluvatar CoreX Semiconductor Co., Ltd. | Asia PacificMorgan Stanley · 2026-08-31