Goldman Sachs is optimistic about A-shares but downgrades H-shares to neutral: As other North Asian markets outperform, the cost of waiting for an H-shares recovery has increased.
```
This year, the Chinese stock market does have opportunities, but the indices and themes have diverged onto two different paths.
On June 3rd, according to news from Zhuifeng Trading Desk, Goldman Sachs strategy team analysts including Kinger Lau released their latest strategy report on the China stock market. In short: they are optimistic about the China AI story, but not about the index that carries the story—especially H-shares.
A-shares and H-shares have become two separate worlds.
Since the beginning of the year, the internal divergence within China's stock market has been extremely pronounced. The MSCI China Index (mainly H-shares) fell 8% this year, while the CSI 300 Index (A-shares, Shanghai and Shenzhen 300) rose 5%, a difference of 13 percentage points; the ChiNext Index outperformed the Hang Seng Tech Index by 40 percentage points.
There are four reasons behind this divergence: index composition, the division between hard technology and soft technology, policy tilting direction, and profit realization capability. 35% of the MSCI China Index’s weight is concentrated in nine internet giants such as Tencent, Alibaba, and Meituan, and these nine companies have seen an average drop of 22% this year. In contrast, "hard tech" sectors such as AI hardware, semiconductors, and infrastructure have performed strongly, contributing 85% of the AI market cap growth in China since DeepSeek.
The analysts wrote: "We like the story of China stocks, but not the indices; continue to overweight A-shares, and downgrade H-shares to neutral."
H-shares still have room for recovery, but other North Asian markets (Japan, South Korea, Mainland China, and Taiwan, etc.) offer stronger cyclical profit growth and higher visibility. If funds continue to overweight H-shares, they will have to bear the opportunity cost of delayed profits.
The Core Problem of H-shares: Repeated Profit Delays
Analysts pointed out the crux of H-shares:
Losses from subsidies for food delivery and instant retail are expected to reach RMB 34 billion in Q1 2026, with a cumulative total of RMB 180 billion since Q2 2025—far exceeding our and the market’s previous expectations in both scale and duration.
This has directly caused the profit recovery timetable for the internet sector to be delayed by 3-6 months compared to expectations at the start of the year. In Q1 2026, overall offshore market profits declined 7% year-on-year, and 30% of companies, by index market cap, reported earnings below market expectations.
On this basis, analysts have lowered their MSCI China EPS growth forecast for 2026 from 12% to 8%, and for 2027 from 14% to 12%. The target PE was compressed from 13x to 12x, and the 12-month target price was lowered from 95 to 85—implying about 11% upside, but this 11% is likely to be realized only once profits improve by year-end.
A-shares: The Logic Still Holds
By contrast, the investment logic for A-shares is much clearer. Analysts have raised the target price for the CSI 300 (Shanghai-Shenzhen 300) from 5,300 to 5,500 and maintain overweight.
This is the second upward revision this year, mainly due to profit accumulation and time carry. The A-share profit growth expectation for 2026 EPS was also raised from 16% to 20%.
Reasons include: higher commodity prices, resilient demand for Chinese exports, improved profit growth (2026 EPS growth raised to 20%), a loose liquidity environment, diversification value for international investors, and direct exposure to hard tech in AI.
A-shares not only performed better in terms of gains this year but also better in return/volatility metrics. They are relatively less sensitive to external shocks, and their correlation with major global markets is also different. This diversification value has not yet been fully priced in for international funds.
More importantly, A-shares have greater exposure to hard tech AI. Sectors like power equipment, infrastructure, semiconductors, advanced manufacturing, and physical AI are more concentrated in A-shares or the related A-share industry chain. This explains why A-shares can outperform with independent trends when MSCI China is weak.
But this does not mean all sectors within the index benefit equally. The advantages of A-shares mainly come from profit improvement and hard tech themes, not a full broad-based re-rating.
AI: China Is One Part of the Global Landscape, But Is Underweighted
Calculations show China accounts for 10% of the global AI-related market cap and 16% of revenue, but the actual allocation by global funds to Chinese AI stocks is only about 1.2%, much lower than its weight in the global AI landscape.
In AI sub-segments, power, infrastructure, semiconductors, and physical AI (robots, autonomous driving) are seen as more certain directions, with stronger policy support and more prominent advantages in China’s manufacturing sector.
H-shares Still Have Room to Recover, But the “Opportunity Cost” of Holding Them Is Rising
H-shares are not expensive in valuation; downgraded to neutral mainly because profit realization is too slow.
The internet/application layer has rebounded recently, but whether it can be sustained depends on whether earnings season from mid to late August can bring a profit inflection point.
MSCI China's target PE is therefore lowered from 13x to 12x. The target level for the next 12 months is lowered to 85 points, still implying about 11% upside, but lower than the previous target of 95 points.
This means offshore China assets are not without opportunities. The A-H rotation model still suggests H-shares have room for recovery relative to A-shares. The relative outperformance projected in the chart is 6%. But this recovery relies more on a profit inflection point, not just on cheap valuation.
The problem is that waiting costs have risen. Other North Asian markets offer stronger cyclical profit growth and higher visibility. If funds continue to overweight H-shares, they have to bear the opportunity cost of profit being delayed.
North Asia usually refers to Mainland China, Hong Kong, Taiwan, Japan, and South Korea.
A more likely rhythm: wait until the internet profit outlook is clearer later this year, especially closer to Q3 2026, and then re-evaluate the beta opportunities in the offshore market.
~~~~~~~~~~~~~~~~~~~~~~~~
The above highlights are from Zhuifeng Trading Desk.
For more detailed analysis, including real-time insights and frontline research, please join [Zhuifeng Trading Desk · Annual Membership]
Risk Disclaimer and TermsThe market is risky, investment needs caution. This article does not constitute individual investment advice, nor does it take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are appropriate for their particular circumstances. Investing based on this article is at your own risk. ```