China Equity Strategy: UBS sees foreign investors cautious on China equities but favors banks and selected AI-hardware supply-chain winners

Overview

This China equity strategy Q&A summarizes feedback from Asian and US investor meetings. UBS describes a cautious foreign-investor backdrop but retains a barbell approach centered on banks and selected technology, hardware and supply-chain opportunities.

Core views

UBS finds that sentiment toward Chinese equities has become materially more cautious following weaker-than-expected GDP growth, restrictive property and taxation policies, and uncertainty around the AI technology supply chain. Interest in China remains elevated, but foreign investors' positions are lighter than in June 2026 and UBS expects deployment to take time without clearer policy support and technology-earnings visibility. Sentiment is negative but not at the 2023/2024 trough. Investors were disappointed by the limited policy response to weak data, the July technology sell-off's lack of diversification, FCC-related uncertainty and the absence of sufficient domestic HBM supply. UBS does not expect international investors to add actively at current levels, despite elevated oil prices and US rates potentially improving China's relative appeal against more vulnerable markets. The report expects policymakers to favor staged easing rather than a sharp pivot like September 2024. UBS Economics expects remaining planned broad fiscal expansion equivalent to 0.7% of GDP relative to H225 to be accelerated through the rest of the year, with quicker fiscal disbursement and policy-bank financing directed toward infrastructure, AI-related investment and industrial production. Consumption support is expected to remain mild; an LPR or RRR cut would be a positive surprise. A further fiscal package is viewed as probable but data-dependent and potentially later in the year. Weak GDP-linked indicators, particularly industrial production, could increase policy urgency, with the October Politburo meeting and December Central Economic Work Conference identified as potential announcement windows. UBS identifies five potential equity catalysts: measures that ease local-government financing stress or raise household income; SOE purchases of A-shares; relief in China's memory and HBM supply; a meaningful rebound in A-share trading volume, which historically has signaled additional inflows; and technology earnings stronger than expected. The report says property stimulus has so far disappointed. It also expects China's AI data-center deployment to keep accelerating; as US AI data-center rollout slows in 2028, China's domestic rollout could become relatively more attractive versus global peers sometime in 2027. Banks are UBS's key defensive preference as investors seek diversification after the July pullback. The report argues that Chinese banks are suited to subdued macro conditions and persistently low domestic bond yields. Revenue growth has accelerated to about 10% in recent quarters as net interest margins stabilized; the dividend-yield premium over the 10-year government bond yield remains above its historical average; and P/B valuations are materially cheaper than global peers. UBS also highlights lower institutional ownership, continuing insurance-company inflows, greater share-price diversification from AI-related sectors, and rising payout ratios that could increase further if loan growth slows. Technology remains constrained near term by component shortages, geopolitical uncertainty and previously crowded positioning, but UBS notes that crowding in a China AI-hardware basket has fallen from its June peak back to early-2026 levels. Onshore retail sentiment has weakened since July, evidenced by declining A-share turnover and margin-financing balances returning to April levels. Foreign institutions were only 0.9% underweight MSCI China in 2Q26, their highest positioning in five years, partly reflecting an estimated 4–5 percentage-point active-fund overweight in A-shares versus benchmark, allocation-driven flows from Korea and Taiwan, foreign A-share inflows, and participation in Hong Kong IPOs and placements. However, China represented only 7.6% of active managers' portfolios in absolute terms in 2Q26, and UBS's top-40-investor tracker indicates positions have been broadly unchanged since June. For AI supply-chain exposure, UBS favors memory companies benefiting from HBM shortages and disciplined global capacity expansion; semiconductor fabs whose high-end capacity determines available compute; semiconductor-equipment suppliers enabling domestic capacity growth and import substitution; industry leaders protected by high barriers to entry in advanced components; and exporters benefiting from higher overseas profitability and established overseas competitiveness. UBS's preferred-name list spans Baidu, Fuyao Glass, Zijin Mining, NAURA Technology, Tencent, CATL, Shengyi Technology, Alibaba, BYD, CNOOC, Zijin Gold International, Shenzhen Kedali, Bank of Jiangsu, China Construction Bank, SMIC, CXMT, GDS, JCET Group, China International Capital and Tianshan Aluminum.

Analysis framework

UBS combines investor-meeting feedback, macro-policy expectations, positioning and turnover indicators, bank operating and valuation comparisons, and sector-level supply-chain analysis. It uses its quant crowding data and market-positioning trackers to assess sentiment, then applies a barbell strategy across defensive banks and selected growth, technology and supply-chain beneficiaries.

Methodology notes

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  AI hardware supply-chain bottleneck analysis

  The report links HBM availability, high-end fab capacity and semiconductor equipment to compute supply and identifies the companies positioned along those bottlenecks.

• Event-Driven and Behavioral FinanceFund-Flow and Positioning Analysis

  Positioning, crowding, turnover and margin-financing indicators

  UBS uses investor positioning, AI-hardware crowding, A-share turnover and margin financing to judge whether sentiment and allocation conditions may support a recovery.

• Valuation methodsDCF (Discounted Cash Flow)

  DCF valuation

  UBS states that it uses DCF models for stocks across Hong Kong and mainland China in its broader valuation work.

• Valuation methodsDDM (Dividend Discount Model)

  Gordon growth model analysis

  The Gordon growth approach values equity using dividend growth assumptions and is one of the valuation approaches UBS states it uses.

• Valuation methodsP/E and PEG Valuation

  Relative valuation using P/E, EV/EBITDA and P/BV multiples

  UBS compares companies using market multiples; its bank thesis specifically emphasizes Chinese banks' P/B discount to global peers.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Chinese banks, including Bank of Jiangsu (600919.SS) and China Construction Bank (0939.HK)

  Preferred defensive allocation within UBS's barbell strategy.

  Strengths

  Stabilizing net interest margins, c.10% recent revenue growth, elevated dividend-yield gap, low ownership and potential payout growth.

  Comparison

  P/B valuations are significantly cheaper than global-bank peers.

  Risks

  Subdued macro conditions and slower loan growth remain relevant to the sector.

• China AI hardware and semiconductor supply chain

  Preferred growth exposure focused on supply-chain bottlenecks and domestic substitution.

  Strengths

  HBM shortages, disciplined capacity expansion, high-end fab capacity constraints, domestic equipment demand and high entry barriers support selected companies.

  Weaknesses

  Near-term performance is impeded by component shortages and geopolitical uncertainty.

  Comparison

  AI-hardware crowding has normalized from its June peak to early-2026 levels.

  Risks

  Memory/HBM supply constraints and uncertainty around domestic AI deployment.

• Baidu (9888.HK), Tencent (0700.HK), Alibaba (BABA.N), NAURA Technology (002371.SZ), CATL (300750.SZ), Shengyi Technology (600183.SS), SMIC (0981.HK), CXMT (688825.SS), JCET Group (600584.SS) and GDS (GDS.O)

  Listed among UBS's most preferred or preferred technology-related names.

  Strengths

  Exposure to AI, semiconductors, compute capacity, data centers or related hardware supply-chain themes.

  Risks

  Component shortages, geopolitical uncertainty, HBM constraints and technology-sector positioning.

Key data

• Planned fiscal expansion0.7% of GDPRemaining planned broad fiscal expansion expected during the rest of the year, relative to H225.
• Chinese bank revenue growthc.10%Accelerated in recent quarters as net interest margins stabilized.
• MSCI China foreign institutional positioning-0.9% underweightIn 2Q26, the highest level in the last five years.
• Active-fund A-share positioning4–5ppt overweightUBS estimate versus benchmark on average.
• China allocation in active-manager portfolios7.6%Absolute portfolio allocation in 2Q26.
• US AI data-center rollout2028UBS expects it to slow, potentially improving the relative appeal of China's domestic rollout in 2027.

Impact & implications

UBS's strategy conclusion is conditional rather than broadly risk-on: stronger policy action, liquidity, technology earnings and HBM supply could draw investors back, while the current environment supports diversification. The institution sees banks as relatively well placed and directs growth exposure toward supply-chain bottlenecks, domestic substitution, technology leaders and exporters.

Risks

• UBS identifies a hard landing in the property market, capital outflows tied to currency depreciation and slow structural reform as risks to Chinese equities.
• Policies that fail to address these risks could shock the market.
• Excessive stimulus could impede the shift from investment-led to consumption-led growth and add to government and SOE debt.

What to watch

• Policy measures that ease local-government financing pressure or improve household income.
• The October Politburo meeting and December Central Economic Work Conference for possible policy announcements.
• SOE purchases of A-shares and a meaningful rebound in A-share trading volume.
• Availability of memory and HBM supply in China.
• Technology earnings relative to expectations and progress in domestic AI data-center rollout.