UBS China Equity Strategy: AI trade is crowded but fundamentals remain strong, favoring ChiNext, technology hardware, and power equipment

Overview

In this China equity strategy report, UBS answers five of the questions investors are most focused on recently: whether the AI trade is excessively crowded, what factors could disrupt China’s AI technology rally, whether higher PPI will compress corporate margins, whether HSTECH and Hong Kong internet stocks have re-rating potential, and how sector preferences and model portfolios should be adjusted. The report concludes that sentiment is generally positive, with solid earnings from hardware technology companies, marginal improvement in real estate, and some policy or industry catalysts supporting risk appetite, but crowded positioning, global tech linkage, Hong Kong IPO lock-up expirations, and weak consumer conditions remain key constraints.

Core views

The core views are: first, AI remains the most discussed theme among investors, and although crowding is high, it has not recently deteriorated rapidly; fundamentals and earnings momentum in AI hardware remain strong. Second, global spillover risk is an important downside risk for China AI technology trades, especially as the correlation between Chinese and U.S. AI tech stocks has risen and U.S. market moves may affect Chinese tech stocks. Third, power equipment and energy storage can serve as alternative proxies for investors worried about AI crowding, as they indirectly benefit from global AIDC buildout and offer some fundamental protection against higher energy prices. Fourth, higher PPI does not necessarily hurt overall listed-company earnings; historically, stronger PPI often coincides with faster revenue and earnings growth, though downstream sectors with weak pricing power may come under pressure. Fifth, HSTECH may take longer to outperform in the near term, as capital is drawn toward hardware technology, power equipment, and North Asia AI names, while lock-up expirations for LLM companies may create short-term pressure.

Analysis framework

The report uses an investor Q&A framework, combining frequently asked questions from recent roadshows and investor communication with quantitative indicators, historical experience, and industry comparisons. The AI crowding section references crowding scores for popular stocks, the share of TMT turnover, fund allocations to electronics and telecom, and foreign capital flows into AI proxy markets such as Taiwan and Korea. The PPI section compares the relationship between PPI and listed-company revenue, earnings, and industry performance over history. The HSTECH section combines index performance, sector composition, earnings revisions, capital flows, and the pace of IPO lock-up expirations. The sector allocation section adjusts preferences based on earnings momentum, valuation, industry exposure, and policy or flow factors.

Methodology notes

• market crowding analysisAI crowdedness and position monitoring

  Assess crowding in trading by using baskets of popular AI names, mutual fund allocations, turnover concentration, and cross-market capital flows.

  The report notes that the crowding score of popular AI stocks has risen significantly over the past 12 months, but has edged down slightly from the early-May 2026 peak; TMT turnover accounts for about 37% of total A-share turnover, below the roughly 43% historical peak in 2023 but still at a high level.

• macro earnings transmissionPPI and listed-company earnings sensitivity

  Use revenue, earnings, and margin performance in historical PPI upcycles to gauge the impact of imported inflation on corporate profits.

  The report argues that higher PPI has historically been accompanied by faster growth in listed-company revenue and earnings. Beneficiaries include upstream, export, pricing-power, and some midstream manufacturing sectors, while downstream sectors with weak pricing power such as food and beverages, consumer services, and utilities are more exposed to pressure.

• index and flow analysisHSTECH re-rating framework

  Evaluate HSTECH performance through index sector composition, earnings revisions, southbound flows, IPO financing, and lock-up pressure.

  HSTECH faces different headwinds because of its higher weights in internet, autos, and consumer sectors, and the combination of earnings downgrades, slower capital flows, and market supply pressure means near-term outperformance may be delayed.

• valuation methodDCF, Gordon Growth Model, and relative valuation

  Use discounted cash flow, the Gordon growth model, and relative valuation methods such as P/E, EV/EBITDA, and P/BV for Hong Kong and mainland China stocks.

  The report notes that covered stocks may use multiple valuation methods depending on industry and company characteristics; in this strategy report, these methods are mainly cited as part of the research framework rather than to provide a single target price.

Asset mapping & comparison

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

• A-share ChiNext

  explicitly preferred

  Strengths

  Strong 1Q earnings momentum, higher exposure to hardware technology and power equipment, and better positioned to benefit from AI hardware and electrification supply chains.

  Weaknesses

  If AI hardware earnings slow or risk appetite fades, growth-style valuations may come under pressure.

  Comparison

  Compared with HSTECH, ChiNext is closer to the hardware technology and power equipment themes currently favored by investors.

  Risks

  AI trade crowding, global tech pullback, rising U.S. yields, and slowing earnings growth.

• China AI hardware technology

  consensus long and core theme

  Strengths

  Strong fundamentals, earnings growth remains resilient, and investors continue to look for supply-chain alpha opportunities.

  Weaknesses

  Positioning is crowded, and the market questions how sustainable the rally is.

  Comparison

  Compared with internet software, hardware is viewed as an enabler of AI rather than a potential substitute.

  Risks

  Spillover from U.S. tech stocks, rising 10-year U.S. Treasury yields, slowing earnings growth, and large LLM companies listing as a sign that growth has peaked.

• China power equipment and energy storage

  alternative proxy for the crowded AI trade

  Strengths

  Indirectly benefits from global AIDC buildout and offers some fundamental protection against higher energy prices; valuations are cheaper than international peers.

  Weaknesses

  It does not have direct AI demand exposure, so performance still depends on orders, competition, and the industry cycle.

  Comparison

  Compared with popular AI hardware names, power equipment may be less crowded and more attractively valued.

  Risks

  Competition across the new-energy supply chain, policy changes, overseas demand volatility, and weaker-than-expected earnings realization.

• HSTECH

  short-term outperformance may be delayed

  Strengths

  Catalysts such as DeepSeek and Tencent's new LLM models, easing food-delivery competition, the meeting between the Chinese and U.S. leaders, potential easing of AI chip export controls, and higher oil prices supporting EV demand have already been partly realized.

  Weaknesses

  High weightings in internet, autos, and consumer sectors leave it exposed to regulation, competition, and insufficient consumption stimulus; investor focus is shifting to hardware technology, power equipment, and North Asia AI names.

  Comparison

  Down about 10% year to date and underperformed the MSCI AC World by about 20 percentage points.

  Risks

  Lock-up expirations for LLM companies, earnings downgrades, slower southbound flows, increased IPO and refinancing supply, and EM outflows by foreign investors.

• Hong Kong IPO lock-up related stocks

  short-term pressure at the single-stock level

  Strengths

  At the overall market level, the roughly HK$200bn July lock-up expiration represents a low share of total market cap, so the macro impact appears manageable.

  Weaknesses

  Stocks usually come under pressure ahead of expiry, and recent IPO samples also underperformed the HSI on a relative basis.

  Comparison

  The single-stock impact is significantly larger than the market-wide impact.

  Risks

  Supply pressure and sentiment shocks from upcoming expirations for Zhipu, Minimax, JD Industrials, and others over the coming months.

• Real estate sector

  from underweight to neutral

  Strengths

  Resale home transactions have improved, listing volumes for second-hand homes have fallen, and marginal data is supporting market sentiment.

  Weaknesses

  Confidence in a sustained recovery remains low, and further data confirmation is still needed.

  Comparison

  Compared with the previous underweight view, the report's stance on real estate has improved but has not turned clearly bullish.

  Risks

  Transaction improvement failing to continue, housing-price pressure, weak resident confidence, and policy effects falling short of expectations.

• Consumer sector

  remain underweight

  Strengths

  The sector could benefit if consumption stimulus or employment and wage growth improve.

  Weaknesses

  Youth unemployment remains high, wage growth slowed in 2H25, and the overall consumption environment has been weak recently.

  Comparison

  Compared with upstream, export, and hardware manufacturing, consumer names lack clear earnings elasticity.

  Risks

  Weak employment and wages, insufficient household willingness to spend, and imported cost pressures that are hard to pass through.

Key data

• TMT trading value shareapproximately 37%A-share TMT trading value as a share of total turnover remains at a high level, but is below the roughly 43% historical peak in 2023.
• Crowding score of popular AI stocksrose significantly over the past 12 months and edged down slightly after the early-May 2026 peakUsed to measure position concentration in popular China AI names.
• ChiNext 1Q earnings momentum+23% in 1Q26One reason the report favors A-share ChiNext is its higher exposure to hardware technology and power equipment, as well as stronger earnings momentum.
• HSTECH year-to-date performanceabout -10% YTDIt underperformed the MSCI AC World, which is near historical highs, by about 20 percentage points.
• HSTECH earnings downgrade over the past 12 months-37%This was mainly driven by food delivery competition, with Meituan, JD, and Alibaba contributing more than 70% of the earnings downgrade.
• Expected July lock-up expiration size in Hong Kongabout HK$200bnThis is about 0.4% of Hong Kong's total market cap and roughly three-quarters of average daily turnover, which appears manageable at the overall market level.
• Pre-expiry stock performancehistorically, relative underperformance versus the broader market of about 25% in the three months before lock-up expirationIn the recent 18-month IPO sample, the effect is weaker, with the median stock underperforming the HSI by about 6% in the three months before expiry.
• Weight of sectors more affected by energy pricesabout 10% of CSI 300Downstream sectors more exposed to high energy costs, such as food and beverages and utilities, account for a limited share of the index.

Impact & implications

For portfolios, the report suggests that Chinese equity risk appetite can still be supported by earnings improvement and the AI hardware theme, but allocation should shift from simply chasing crowded AI trades toward more granular opportunities with better earnings protection and valuation appeal. A-share ChiNext, technology hardware, power equipment, and energy storage are relatively attractive; HSTECH and Hong Kong internet stocks may still have catalysts, but may need more time to recover because of flow preferences, earnings downgrades, and lock-up pressure. Higher PPI is not necessarily negative for aggregate earnings, but industry dispersion is likely to widen, with stronger pricing power, upstream, and export chains outperforming weak-pricing-power consumer and utility names.

Risks

• AI-theme positioning is crowded, and a reversal of positions could magnify downside moves.
• Correlation between U.S. tech stocks and Chinese AI tech stocks has risen, so a global tech correction could spill over into the Chinese market.
• Further increases in U.S. yields, especially if driven by higher energy prices, could weigh on growth-stock valuations.
• If AI hardware earnings growth slows, the current trading logic may need to be reassessed.
• A U.S. listing by a large LLM company may be interpreted by some investors as a sign that LLM commercialization growth has peaked.
• HSTECH faces an unfavorable industry structure, earnings downgrades, weak capital flows, and supply pressure from IPOs and refinancing.
• Lock-up expirations for Hong Kong IPOs and LLM companies may create short-term selling pressure in individual stocks.
• It remains uncertain whether China can continue to emerge from deflation, and weak employment and wage conditions may weigh on consumer spending.
• Imported inflation may pressure margins in downstream sectors with weak pricing power.

What to watch

• Whether the crowding score of popular AI names continues to fall from the early-May 2026 peak or starts rising again.
• Whether A-share TMT turnover as a share of total turnover approaches or exceeds the roughly 43% peak reached in 2023.
• U.S. tech stock performance and changes in its correlation with Chinese AI tech stocks.
• The 10-year U.S. Treasury yield and energy price trends.
• Whether earnings growth in China AI hardware and U.S. tech hardware companies continues to accelerate.
• The pace and market interpretation of large LLM companies listing in the U.S.
• HSTECH capital flows, southbound flows, internet platform competition, and earnings revisions.
• The roughly HK$200bn Hong Kong lock-up expirations in July and stock performance for Zhipu, Minimax, JD Industrials, and others.
• Whether higher PPI translates into revenue and profit growth for listed companies, especially with attention to industry dispersion.
• Second-hand home transactions, listing volumes, and the sustainability of the real estate recovery.
• Youth unemployment, wage growth, and changes in the consumer environment.