China equity “rotation temptation”: still favoring A-share hard technology, but beginning to watch for an H-share soft technology recovery
AI summary card
China equity “rotation temptation”: still favoring A-share hard technology, but beginning to watch for an H-share soft technology recovery
Goldman Sachs believes the divergence within Chinese equities has reached an extreme level this year. A-share hard technology remains the tactical preference, but H-share internet/soft technology may gradually narrow its performance gap amid valuation de-rating and expectations for future earnings improvement.
- STAR50 has outperformed HSTECH by approximately 68 percentage points year to date, highlighting the significant divergence in trading between Chinese AI hardware and software, suppliers and capital expenditure.
- HSTECH has rebounded 11% over the past two weeks but remains down 14% year to date. Goldman Sachs believes H-share soft technology may improve over the coming months, but sustained gains require confirmation of an earnings recovery in 2Q or 3Q.
- MSCI China earnings declined 8% year over year in 1Q26, mainly due to the internet sector; internet companies have accumulated more than RMB180 billion in subsidy losses since 2Q25.
- Goldman Sachs does not believe Chinese AI stocks as a whole have formed a bubble, but warns that valuation, concentration and leverage risks are rising in semiconductors and some A-share hard technology segments.
- Strategically, Goldman Sachs maintains an overweight on Chinese A-shares and recommends monitoring materials, capital goods, insurance, selected AI portfolios, the GS 15FYP Portfolio and non-AI opportunities.
Report interpretation
Overview
This report summarizes the key questions and feedback from Goldman Sachs’ recent investor conferences in Asia and the Americas, focusing on the major divergence within Chinese equities, A-share/H-share rotation, the shift between hard and soft technology, internet earnings pressure, whether the AI theme is overheated, capital competition in North Asian markets, Hong Kong IPO activity and how to generate returns in the second half of 2026. Overall, the report believes that AI and policy support remain important themes for Chinese equities, while macro factors have temporarily become secondary.
Core views
Goldman Sachs maintains a tactical preference for A-shares over H-shares and hard technology over soft technology, as A-share earnings momentum and the diversification value arising from international investors’ underweight positions remain more favorable. However, after significant valuation de-rating in H-share internet and soft technology, a period of rotation may occur over the coming months to quarters if narrower subsidy losses and opportunities in cloud computing, agentic AI and AI tokens drive earnings improvement. The report emphasizes that low valuations alone are insufficient to support a sustained market trend; the true catalyst remains a visible recovery in profits.
Analysis framework
The report uses a multidimensional framework incorporating investor Q&A, relative performance of regional and market indexes, A-H market rotation models, hard versus soft technology rotation models, earnings sensitivity, comparisons between AI market capitalization and potential economic value, capital positioning, and IPO supply-demand analysis to assess the allocation direction for Chinese equities in the second half of 2026.
Methodology notes
Uses fundamental, valuation and liquidity factors to assess whether the relative performance of A-shares versus H-shares and hard technology versus soft technology may rotate.
The models point to potential improvement in H-share soft technology relative to A-share hard technology over the coming months, but the report also emphasizes that sustained gains still require an earnings recovery for confirmation.
Compares the rise in the market capitalization of Chinese AI stocks with the efficiency gains, new profits and TAM expansion that AI may generate.
The report believes Chinese AI stocks as a whole are not yet in a bubble, but some semiconductors and A-share hard technology segments show localized signs of overheating.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- Chinese A-sharesCore tactical preference
- Strengths
- More favorable earnings momentum, strong exposure to AI hard technology, and undervalued diversification value for international investors.
- Weaknesses
- Valuation, concentration and leverage risks are rising in some hard technology and semiconductor sectors.
- Comparison
- Still tactically preferred by Goldman Sachs over H-shares.
- Risks
- Could face valuation retracement if the AI trade becomes overheated or earnings fail to materialize.
- H-share internet/soft technologyCandidate for gradual accumulation
- Strengths
- Valuations have been significantly de-rated; some negative assumptions regarding AI investment returns may already be reflected in prices, and earnings recovery could drive re-rating.
- Weaknesses
- 1Q26 earnings were weak, while subsidy losses and AI capital expenditure continue to pressure profits.
- Comparison
- Still lagging A-share hard technology in the short term, but may narrow the gap over the coming months.
- Risks
- If the earnings recovery is unclear in 2Q or 3Q, the rebound may prove unsustainable.
- Chinese AI stocksHigh-conviction theme requiring selectivity
- Strengths
- China has an important position globally in AI market capitalization, revenue, model usage, power, infrastructure and physical AI.
- Weaknesses
- International investor holdings remain insufficient, but some subsectors are already showing signs of overheating.
- Comparison
- The report believes the sector as a whole does not constitute a bubble, but risk-reward differentiation has increased.
- Risks
- Rising valuation premiums, crowded trading, leverage and thematic concentration.
- Hong Kong IPOsSource of alpha
- Strengths
- New listings have delivered strong short-term post-listing returns year to date; large offerings, standalone H-share listings and moderate cornerstone investor ownership support performance.
- Weaknesses
- A long IPO pipeline has raised market concerns about liquidity absorption.
- Comparison
- Compared with substantial dividends and buybacks by listed companies, the pressure from new supply is considered manageable.
- Risks
- IPO returns and liquidity could come under pressure if risk appetite declines or offering sizes exceed expectations.
Key data
- STAR50 year-to-date performance relative to HSTECHApproximately +68 percentage pointsReflects the extreme divergence between A-share AI hard technology and H-share internet/soft technology.
- ChiNext year-to-date performance relative to SHCOMP/CSI300Approximately +19/+17 percentage pointsShows that growth and hard technology-related sectors have significantly outperformed broad-based indexes.
- Recent HSTECH performanceRebounded 11% over the past two weeks, but remains down 14% year to dateThe short-term rebound was driven by low valuations, improving news flow and progress in AI applications and monetization.
- MSCI China 1Q26 earningsDown 8% year over yearThe internet sector was the main drag and accounts for approximately 35% of index earnings weight.
- Internet subsidy lossesCumulative losses exceeding RMB180 billion since 2Q25Narrowing subsidy losses are an important variable for future earnings improvement.
- China’s global AI shareApproximately 11% of AI-related market capitalization and 18% of revenueGoldman Sachs believes China remains an important component of the global AI equity landscape.
- Hong Kong IPOs year to date100 companies listed, raising approximately US$35 billionMedian returns one month and three months after listing were approximately 32% and 30%, respectively.
- Estimated remaining new equity supply for the yearApproximately US$25 billion in IPOs and approximately US$45 billion including follow-on offeringsGoldman Sachs believes concerns about liquidity absorption may be exaggerated relative to the scale of listed-company dividends and buybacks.
Impact & implications
For investors, the report recommends not simply chasing AI hard technology stocks that have already risen sharply, but seeking earnings-driven opportunities while controlling valuation and concentration risks. Allocation should continue to center on a tactical preference for Chinese A-shares, while gradually building positions in selected large-cap H-share internet and soft technology proxies. Materials, capital goods, insurance, shipbuilding, new consumption, healthcare/biotechnology, the property chain, brokerages and the hog cycle are also viewed as potential sources of alpha in the second half of 2026.
Risks
- The recovery in H-share internet earnings occurs later than expected, preventing the valuation recovery from being sustained.
- Localized overheating in A-share hard technology, semiconductor and AI-related sectors triggers a valuation retracement.
- Returns on AI capital expenditure fall below market expectations, weighing on the profits of internet and platform companies.
- Hong Kong IPO supply exceeds expectations or market liquidity deteriorates.
- Macro factors re-emerge, including uncertainty surrounding real estate, consumption, the renminbi, exports, oil price shocks and policy stimulus.
- International capital continues to favor North Asian AI hardware markets such as Korea and Taiwan, weighing on the beta performance of Chinese equities.
What to watch
- Whether leading Chinese internet companies show a visible earnings recovery in 2Q26 and 3Q26.
- Whether subsidy losses narrow and whether new businesses such as cloud, agentic AI and AI tokens can contribute incremental profits.
- Whether the performance gap between HSTECH and STAR50 and A-share hard technology continues to narrow.
- Whether valuation, leverage and trading concentration in semiconductors and A-share hard technology continue to rise.
- The supply-demand balance between Hong Kong IPOs and follow-on offerings versus funds available for dividends and buybacks.
- The July Politburo meeting and subsequent policy stimulus signals.
- Changes in overseas long-only and hedge fund net exposure, gross exposure and alpha-trading preferences toward Chinese equities.