UBS: A-Share Tech Crowding May Not Have Peaked; Medium-Term Growth Dominance Persists

Overview

This report systematically reviews four typical industry/theme crowded trades and their unwinding processes in A-shares since 2013, distilling three regular characteristics, and uses this to assess the current crowding level of the A-share tech sector. The core conclusion is: under a slow bull scenario, the rise in the tech growth style during this round may not have peaked in terms of magnitude and duration. Although rising short-term trading crowding may trigger market consolidation, the medium-term trend favoring growth styles—driven by the AI industry wave, policies for technological self-reliance, and continuous inflows from high-risk-appetite capital—remains unchanged.

Core views

Essence and Historical Patterns of Crowded Trades: The report defines "crowding" as the result, not the cause, of asset managers expressing views. When industry fundamentals or macro narratives improve significantly, sustained incremental inflows from specific types of capital push up stock prices, leading to excess returns for fund products, attracting more off-market investors, and forming a self-reinforcing positive feedback loop. Conversely, when fundamentals or narratives undergo substantive changes, capital outflows and NAV drawdowns reinforce each other, causing the crowded trade to unwind. Reviewing the four cycles of TMT (2013-2015), Consumption Upgrade (2017-2019), New Energy (2019-2021), and Dividend Defense (2022-2024), the report summarizes three characteristics: First, crowding relies on unique industry/macro narratives and unwinds when the narrative shifts; Second, it is driven by sustained incremental capital from specific investor groups, and slowing or reversing capital triggers unwinding; Third, historically, industry/theme crowding cycles typically last about three years. Multi-dimensional Assessment of Current Tech Crowding: From a time dimension, since the policy shift on September 24, 2024, the outperformance duration of the tech growth style has been less than two years, far shorter than the historical average. From institutional holdings, in Q1 2026, public mutual funds' overweight ratio in the broad tech sector (including electronics, communications, computers, and defense) was 9.9%, not only lower than the 11.6% in Q3 2025 but also far below the historical peak of 14.1% in Q4 2015, and significantly lower than the cyclical highs of 18%-23% for past consumer sectors, indicating room for further allocation by public funds. From leveraged capital, as of June 12, 2026, A-share margin trading balances reached RMB 2.85 trillion, but accounted for only 5.3% of free-float market cap, far below the mid-2015 peak of 9.6%. From a global benchmark perspective, the AI-driven global tech wave is still unfolding; the PE premium of the global semiconductor sector relative to the broader market is about 17%, close to the 10% historical mean, and earnings upgrades over the past three months exceeded the MSCI ACWI index by 10 percentage points, suggesting valuations are not significantly deviating from reasonable ranges. Incremental Capital Structure and Short-term Monitoring Indicators: The source of incremental capital in the current market has shifted from long-term funds like insurance to high-risk-appetite funds such as margin trading, ETFs, and hedge funds. In 2025, new registrations of private securities investment funds grew 241% YoY to RMB 487.9 billion, and in the first four months of 2026, they still grew 114% YoY; the total scale of industry and theme ETFs expanded from RMB 469.4 billion at the end of August 2024 to approximately RMB 1.18 trillion, with higher holding concentration than active funds, benefiting tech leaders. For short-term trading crowding, the report proposes three technical indicators: weekly turnover of the broad tech sector as a percentage of total A-share turnover, rolling four-week net margin buying, and the turnover share of the top 1%/5%/10 stocks by turnover. Backtesting shows that after the tech sector's turnover share exceeds 38%, major indices still recorded positive returns in the subsequent 5-40 trading days, with the STAR 50 Index performing particularly well, indicating that short-term crowding does not necessarily mean the end of the medium-term trend. Potential Unwinding Triggers: The report highlights three risks that could lead to the unwinding of the current tech crowded trade. First, escalating geopolitical tensions in the Middle East driving up global oil prices; US CPI rose to 4.2% in May, and if inflation remains high, it may constrain the Fed's policy space. Second, a shift in the Fed's monetary policy and accelerated IPOs of large global tech companies could drain liquidity from the tech sector. Third, China's PPI recovery driving corporate profits and wage growth; if consumption recovery materializes, some capital may flow from the tech sector to the consumer sector.

Analysis framework

The report adopts an analysis framework of "Historical Analogy + Multi-dimensional Data Verification." First, by reviewing four complete crowding-unwinding cycles since 2013, it distills three universal characteristics: narrative-driven, capital-driven, and time-cycle, establishing an analytical baseline. Then, placing the current tech sector within this framework, it conducts horizontal and vertical comparisons across four dimensions: time span, institutional holding levels, leveraged capital ratios, and global valuation comparisons, to determine the current position. It also introduces an analysis of incremental capital structure, distinguishing the inflow pace and holding characteristics of funds with different risk appetites, to more accurately assess the sustainability of crowding. Finally, it constructs three quantifiable short-term trading crowding monitoring indicators and verifies their signal significance through historical backtesting, combining qualitative judgment with quantitative tracking.

Methodology notes

• Event Game Theory and Behavioral FinanceReflexivity

  Self-reinforcing positive and negative feedback loops in crowded trades

  The report views crowding as the result of a positive feedback loop of capital inflows, rising stock prices, fund excess returns, and more capital inflows, while unwinding is the reverse negative feedback. This analytical paradigm originates from Soros's theory of reflexivity, emphasizing the bidirectional influence between market prices and participant behavior, helping readers understand why crowding itself is not a peak signal, but the breaking of the feedback loop is.

• Quantitative/Factor/Portfolio TheoryCapital Flow/Chip Analysis

  Crowding assessment based on incremental capital structure and holding concentration

  The report does not rely on a single valuation metric but judges the degree of crowding by dissecting the inflow pace, holding proportions, and historical percentiles of different types of capital such as margin trading, ETFs, hedge funds, and public mutual funds. The core logic of this method is: different capitals have different risk appetites and behavioral patterns; crowding can only unwind when the inflow momentum of dominant capital exhausts.

• Cycle and Prosperity FrameworkProsperity Turning Point Analysis

  Three-year empirical rule for industry/theme crowding cycles

  The report finds that historically in A-shares, industry overweighting from trough to peak usually takes about three years, which is related to the limits of industrial high-growth cycles and capital concentration. This empirical rule can serve as a reference frame for judging the current stage of crowding, but note its statistical nature rather than being a precise prediction tool.

• Valuation methodsPE/PEG valuation

  Historical percentile comparison of global semiconductor sector relative valuation premium

  The report compares the PE premium of the global semiconductor sector relative to the broader market (currently about 17%) with its historical mean (about 10%) to judge whether there is significant overvaluation. This relative valuation method avoids the issue of absolute valuation being affected by interest rate environments and is more suitable for cross-cycle, cross-market horizontal benchmarking of tech sectors.

Asset mapping & comparison

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

• NAURA Technology (002371.SZ)

  Core beneficiary of tech self-reliance, leader in semiconductor equipment

  Strengths

  2026E/2027E EPS growth estimated at 23%/71%, benefiting from accelerated domestic substitution

  Comparison

  Target price RMB 800, implied upside 19%, PE 71.85x

• AMEC (688012.SS)

  Leader in semiconductor etching equipment, beneficiary of AI computing power expansion

  Strengths

  2026E/2027E EPS growth estimated at 73%/70%, outstanding growth prospects

  Comparison

  Target price RMB 403, implied upside 26%, PE 80.82x

• Sungrow (300274.SZ)

  Leader in PV inverters and energy storage, beneficiary of global energy transition

  Strengths

  Relatively reasonable valuation, 2026E PE only 20.82x

  Comparison

  Target price RMB 182, implied upside 19%, EPS growth 13%/27%

• Inovance (300124.SZ)

  Leader in industrial automation and NEV electronic control

  Strengths

  Benefiting from intelligent manufacturing upgrade, steady performance growth

  Comparison

  Target price RMB 86, implied upside 22%, 2026E PE 32.82x

• CR Micro (688396.SS)

  Power semiconductor IDM leader, core beneficiary of domestic substitution

  Strengths

  2026E EPS growth estimated at 92%, high elasticity

  Comparison

  Target price RMB 83.4, implied upside 16%, PE 75.08x

• Universal Scientific Industrial (USI) (601231.SS)

  SiP packaging leader, beneficiary of consumer electronics and AI terminals

  Strengths

  Target price implies 40% upside, highest among recommended stocks

  Comparison

  Target price RMB 51.5, 2026E PE 36.40x

• Silan Microelectronics (600460.SS)

  Power semiconductor and MEMS sensor company

  Strengths

  2026E EPS growth estimated at 129%, strong expectation of performance turnaround

  Comparison

  Target price RMB 46.2, implied upside 21%

• Putailai New Energy (603659.SS)

  Leader in lithium battery anode materials and coated separators

  Strengths

  Low valuation, 2026E PE only 17.81x, implied upside 45%

  Comparison

  Target price RMB 43, EPS growth 51%/28%

• Kedali (002850.SZ)

  Leader in lithium battery precision structural components

  Strengths

  2026E EPS growth 40%, benefiting from battery capacity expansion

  Comparison

  Target price RMB 268, implied upside 32%

• Shuanghuan Driveline (002472.SZ)

  Precision gear and robot reducer company

  Strengths

  Benefiting from dual demand from robots and new energy vehicles

  Comparison

  Target price RMB 53, implied upside 25%, 2026E PE 24.47x

Key data

• Public Mutual Funds' Broad Tech Overweight Ratio9.9%Q1 2026 data, lower than 11.6% in Q3 2025 and the Q4 2015 peak of 14.1%, far below the historical high of 18%-23% for consumer sectors
• Margin Trading Balance as % of Free-Float Market Cap5.3%As of June 12, 2026, far below the historical peak of 9.6% in mid-2015
• Global Semiconductor Relative PE Premium~17%Close to the historical mean of 10%, no significant overvaluation; earnings upgrades over the past three months exceeded MSCI ACWI by 10 percentage points
• Total Scale of Industry and Theme ETFs~RMB 1.18 TrillionSignificant expansion from RMB 469.4 billion at the end of August 2024, with higher holding concentration than active funds
• New Registration Scale of Private Securities Funds in 2025RMB 487.9 BillionYoY growth of 241%; still grew 114% YoY in the first four months of 2026
• A-Share Average Daily TurnoverRMB 2.65 TrillionData for the first five months of 2026, a significant increase from the full-year average of RMB 1.06 trillion in 2024

Impact & implications

The report believes that under the current slow bull scenario, the medium-term dominance trend of the tech growth style has not ended, providing continuous valuation support for related tech stocks. For market participants, the rise in short-term trading crowding may bring phased consolidation, but should not be overly interpreted as a trend reversal signal. The shift in incremental capital structure from defensive funds like insurance to offensive funds like margin trading, ETFs, and hedge funds means market volatility may remain high, but it also provides more elastic liquidity support for the tech sector. If catalysts such as easing Middle East tensions, clarification of Fed policy, or validation of domestic consumption data emerge subsequently, the upside potential for the tech sector is expected to expand further.

Risks

• Escalating Middle East geopolitical tensions driving up global oil prices, exacerbating inflation pressure and constraining Fed policy space
• Unexpected Fed monetary policy shift towards rate hikes, or accelerated IPOs of large global tech companies draining liquidity from the tech sector
• China's PPI recovery driving consumption recovery, potentially diverting funds from the tech sector to the consumer sector
• Systemic risks such as a hard landing in the real estate market, capital outflows, and slow progress in structural reforms

What to watch

• Whether the weekly turnover of the broad tech sector as a percentage of total A-share turnover continues to exceed 38% and its subsequent trend
• Whether the rolling four-week net margin buying of the broad tech sector breaks the RMB 80 billion threshold
• Trends in the turnover share of the top 1%/5%/10 stocks by turnover
• Evolution of the Middle East situation and progress in US-Iran negotiations impacting global oil prices
• Fed statements at the June FOMC meeting and subsequent inflation data
• Whether China's non-financial corporate profit growth meets the 11% expectation, and the linkage effect of wage and sales expense growth