Deleveraging is nearing its end, opening a reallocation window for Chinese technology stocks

Overview

This report assesses the allocation value of Chinese technology stocks after their sharp correction in July. UBS believes that the decline in margin financing balances, the correction in technology stock prices, and the reduction in global leveraged exposure indicate that technical deleveraging is nearing its end. At the same time, results from U.S. hyperscale cloud service providers, demand from AI-native companies, cloud backlog orders, and enterprise AI spending all point to an improving global AI narrative. The fundamentals of China's domestic technology supply chain have not deteriorated significantly due to the market selloff, and with improving domestic GPU supply, AI data center construction is expected to accelerate in the second half of 2026.

Core views

Now is a favorable time to reallocate to Chinese technology stocks, but investors should not simply replicate the concentrated holdings of the first half. AI technology hardware still benefits from earnings upgrades and localization drivers, with semiconductor equipment, network chips, and advanced packaging—segments with relatively weaker cyclicality—being preferred. However, share price volatility and uncertainty around AI commercialization may limit near-term valuation expansion. Market performance is expected to broaden from a small number of AI hardware names to more industries. The internet benefits from improving earnings trends, a stronger AI narrative among global cloud vendors, and low valuations; power equipment benefits from AI data center construction and energy autonomy; non-ferrous metals have strong earnings trends; and companies expanding overseas are expected to benefit from the gradual fading of earnings headwinds caused by currency appreciation.

Analysis framework

The report combines indicators such as individual stock price drawdowns, A-share margin financing balances and collateral ratios, domestic fund holdings, technology turnover share, ETF fund flows, and exposure to leveraged products in Korea and the United States to assess the deleveraging process. It then validates fundamentals through earnings forecast revisions, forward P/E and P/B ratios, global cloud revenue and backlog orders, enterprise AI spending, and compute demand, while comparing valuations and earnings trends across industries to ultimately form sector preferences and model portfolio adjustments.

Methodology notes

• Market technicals and fund flow analysisLeverage, holdings, and fund flow tracking

  Use margin financing balances, collateral ratios, fund holdings, turnover share, and ETF flows to judge whether technical selling is nearing an end.

  When both margin financing balances and technology stock prices have fallen back near the starting point of the previous leveraging phase, while collateral coverage remains healthy, the pressure for further passive deleveraging may decline.

• Earnings and valuation analysisEarnings forecast revisions and relative valuation

  Compare forward P/E, P/B, and earnings forecast changes for Chinese and global AI technology hardware.

  A significant pullback in share prices and valuations, while earnings forecasts are still being revised upward, implies an improved risk-reward profile; the report also compares valuations across industries to identify potential rotation directions.

• Absolute valuationDCF model

  Assess a company's intrinsic value by discounting future cash flows.

  UBS states that the mainland China and Hong Kong-listed companies it covers are valued using DCF models based on industry and company characteristics.

• Absolute valuationGordon Growth Model

  Estimate equity value based on long-term dividend growth and cost of capital.

  This method is one of the valuation tools UBS uses for some covered stocks and is generally suitable for companies with relatively stable cash flows and dividends.

• Relative valuationValuation multiple comparison

  Use P/E, enterprise value multiples, and P/B to compare a company with historical or peer valuations.

  The report explicitly lists methods including P/E, EV/EBITDA, and P/B, and compares current valuations of China's AI technology hardware with historical averages and global comparable assets.

Asset mapping & comparison

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

• Chinese AI technology hardware

  Most preferred

  Strengths

  Technical deleveraging is nearing its end, valuations have fallen sharply, earnings forecasts continue to be revised upward, and domestic supply chain catch-up and AI data center construction provide medium-term growth drivers.

  Weaknesses

  Recent share price volatility is high, investors' risk tolerance has declined, and the sector's near-term performance may no longer be as concentrated and strong as in the first half.

  Comparison

  Current valuations are only slightly above historical averages, and fundamentals have not clearly deteriorated compared with before the global selloff.

  Risks

  AI commercialization progress falls short of expectations, improvement in domestic GPU supply is delayed, or leverage-driven liquidation or valuation compression occurs again.

• Semiconductor equipment, network chips, and advanced packaging

  Top picks within AI hardware

  Strengths

  Benefit from semiconductor localization and AI infrastructure expansion, with weaker cyclicality than other AI-themed segments.

  Weaknesses

  Valuations of some names remain not low, and earnings delivery depends on capital expenditure and domestic substitution progress.

  Comparison

  Compared with more cyclical AI hardware segments, earnings visibility and defensiveness are better.

  Risks

  Slower capital expenditure, technology iteration, supply chain bottlenecks, and intensified competition.

• Chinese internet

  Preferred

  Strengths

  Improving earnings trends, inexpensive valuations, and benefits from a stronger AI commercialization narrative among global hyperscale cloud service providers.

  Weaknesses

  Growth is still affected by the macro consumption environment and competitive landscape.

  Comparison

  Compared with the recently crowded AI hardware trade, internet valuations are lower and there is room for fund inflows to return.

  Risks

  AI investment returns fall short of expectations, consumption recovery is weak, and regulatory uncertainty.

• Power equipment

  Preferred

  Strengths

  Low valuations, and benefits from AI data center expansion, increased electricity demand, and the trend toward energy autonomy.

  Weaknesses

  There may be a time lag in project orders and earnings realization.

  Comparison

  Combines AI infrastructure exposure with relatively low valuations, making it a broadening direction outside technology.

  Risks

  AI data center construction falls short of expectations, project delays, and intensified industry competition.

• Non-ferrous metals

  Preferred

  Strengths

  Strong earnings trends and potential benefits from resource demand and supply constraints.

  Weaknesses

  Earnings are sensitive to commodity prices and the macro cycle.

  Comparison

  Compared with some high-valuation growth sectors, earnings momentum is more direct and valuation support is clearer.

  Risks

  Weakening global demand, falling commodity prices, and exchange rate volatility.

• Companies expanding overseas

  Preferred

  Strengths

  The earnings drag previously caused by currency appreciation is expected to gradually fade, and overseas markets can provide sources of growth.

  Weaknesses

  Overseas operations, trade policies, and exchange rate changes increase earnings volatility.

  Comparison

  When domestic demand is uncertain, overseas revenue provides some business diversification.

  Risks

  Trade frictions, overseas regulation, tariffs, and changes in the RMB exchange rate.

• Construction, automobile OEMs, and consumption

  Least preferred

  Strengths

  If policy stimulus or demand recovery exceeds expectations, there may be a phased rebound.

  Weaknesses

  Currently relatively lacking clear earnings upgrades and structural growth catalysts.

  Comparison

  Compared with technology hardware, internet, power equipment, and non-ferrous metals, the risk-reward appeal is weaker.

  Risks

  Property downturn, slow consumption recovery, price competition, and policy effects falling short of expectations.

Key data

• July performance of Chinese AI technology hardware stocks-32%Average monthly decline of selected stocks tracked by UBS.
• Proportion of stocks with sharp declines36%Share of tracked names that fell by 40% or more in July.
• A-share margin financing balanceRMB2.6 trillionHas fallen from a peak of about RMB3 trillion, close to the April 2026 level.
• A-share margin collateral ratioAbout 280%Indicates that overall collateral coverage remains relatively healthy.
• Decline in assets of Korean leveraged ETFsAbout 50%Used to illustrate that global technology-related leverage has declined significantly.
• Decline in U.S. leveraged ETF exposureAbout 30%Shows that leverage risk in the U.S. market has also been released to some extent.
• Quarter-on-quarter increase in average enterprise AI spending25%The report views this as evidence of accelerating enterprise AI adoption.
• BABA.N model portfolio dataClosing price 126.81; target price 195.00UBS rates it Buy, with FY26E P/E of 20.0x.
• 600584.SS model portfolio dataClosing price 77.75; target price 135.00UBS rates it Buy, with FY26E EPS growth of 72%.

Impact & implications

The easing of technical selling pressure and improvement in global AI demand together increase the allocation appeal of Chinese technology stocks, but investment strategy should shift from a highly concentrated AI trade to a fundamentals-driven multi-industry portfolio. Investors can retain core exposure to the domestic AI supply chain while increasing exposure to the internet, power equipment, non-ferrous metals, and companies expanding overseas to diversify AI commercialization and high-volatility risks. UBS's model portfolio adds Alibaba Group, 3908.HK, 603308.SS, and 600584.SS, and lists technology hardware, internet, companies expanding overseas, non-ferrous metals, and power equipment as the most preferred areas.

Risks

• A hard landing in China's property market could drag on economic growth and equity valuations.
• RMB depreciation could trigger capital outflows and may exacerbate market volatility.
• Structural reform progress may be slow, or policies may fail to adequately address key risks.
• Excessive stimulus may hinder the economy's transition from investment-driven to consumption-driven growth and increase debt at the government and state-owned enterprise levels.
• Uncertainty remains around AI commercialization and investment returns, which may limit valuation recovery in technology stocks.
• If margin trading heats up again and the market falls once more, it may trigger a new round of technical selling.
• Domestic GPU supply and AI data center construction progress in the second half of 2026 may fall short of expectations.
• Historical performance and target prices in the report do not guarantee future returns.

What to watch

• Whether A-share margin financing balances stabilize around RMB2.6 trillion and whether margin collateral ratios continue to remain healthy.
• Changes in technology sector turnover share, domestic fund holdings, and fund flows into STAR 50 and ChiNext-related ETFs.
• Whether earnings forecasts for China's AI technology hardware continue to be revised upward and whether valuations remain within a reasonable range.
• Improvement in domestic GPU supply and the start, delivery, and capital expenditure progress of AI data center projects in the second half of 2026.
• Cloud revenue, backlog orders, and AI commercialization data from global cloud service providers such as Microsoft and Amazon.
• Whether the growth rate of enterprise AI spending and compute demand can be sustained.
• Whether market upside breadth broadens from AI hardware to the internet, power equipment, non-ferrous metals, and companies expanding overseas.
• Changes in real estate, the RMB exchange rate, capital flows, and structural reform policies.