Goldman Says China AI Stock Correction Has Cleared Key Risks, Urges Diversification Across Four Themes
China’s AI hardware and technology stocks staged a strong rally in the first half of the year. But after June, as the global AI trade unwound, Chinese AI stocks also suffered a deep pullback. On Aug. 9, a team led by Goldman Sachs’ chief China equity strategist, Liu Jinjin, published a research report titled “AI Changes the Game: Correction, Rotation and Diversification.” The report reviewed the drivers of the latest rise and fall in China’s AI sector, used 10 quantitative signals to assess the maturity of the correction, and outlined an allocation strategy. Goldman analysts said the latest decline has absorbed the sector’s key earlier risks. They cautioned, however, that volatility in technology stocks listed in Hong Kong and mainland China is likely to remain high in the near term. Investors should move beyond a single-track approach, the analysts said. The report said China’s AI hardware stocks tracked by Goldman rose an average of 33% in the first half of the year, with computing power and semiconductors emerging as the strongest market themes. After global AI stocks began correcting in June, however, those shares fell sharply from their June highs. The STAR Market 50, ChiNext Index and CSI 1000 each dropped more than 20% from recent highs. To assess the latest AI stock correction, Liu’s team developed 10 indicators covering market pricing, risk appetite, fund flows, positioning structure, industrial policy and corporate fundamentals. The indicators were designed to measure the current stage of the correction, potential downside risks and medium- to long-term return potential for AI hardware stocks. Goldman said the decline had clearly absorbed four major risks that had built up earlier in the sector: speculative long positions had been largely closed, overall valuations had moved away from historical highs, onshore leverage had been substantially reduced, and crowded positioning in the sector had eased significantly. Although those risks have been released in stages, Goldman warned that volatility in A-share and Hong Kong-listed technology stocks would remain elevated in the short term. The report advised against concentrating heavily in a single AI hardware segment. Instead, its core approach is to diversify across assets inside and outside the AI supply chain that have low correlations with one another, using a balanced portfolio to offset market swings. The report identified four differentiated investment themes: First, investors could add selected Hong Kong-listed internet and AI software companies in stages. After a prolonged correction, Hong Kong internet stocks offer more attractive valuations, while the rollout of AI applications could support an earnings recovery. Goldman said the group could provide balance outside AI hardware and help offset volatility in hardware stocks. Second, investors should consider policy-supported themes while covering both domestic-demand and export-oriented AI companies. Goldman said investors should focus first on the Goldman Sachs 15-year investment portfolio as well as inward-focused and export-oriented artificial-intelligence companies. Third, the report called for identifying nontechnology stocks that have been oversold and whose earnings expectations are being revised higher. In a high-volatility environment for technology stocks, some traditional-industry companies have undergone deep corrections and, combined with improved earnings expectations, may have room for valuation recoveries. Goldman described them as defensive core holdings that could smooth portfolio returns. Fourth, investors could seek excess returns from initial-public-offering opportunities in the primary market. The new-stock market offers Alpha opportunities independent of the performance of AI stocks in the secondary market, the report said. Goldman highlighted listed companies with stable dividends and ample cash flow as the focus of that screening. Liu’s team concluded that the current correction in AI hardware stocks represents a clearing of risks after a major rally, rather than the end of the industry’s long-term trend. In the short term, however, competition for capital is intensifying and the risk-reward appeal of a pure sector-concentration strategy has declined. The report said investors should move away from a one-sided bet on AI hardware and instead combine software and hardware, balance technology and nontechnology holdings, and link primary- and secondary-market exposure. That approach could capture the AI industry’s long-term benefits while limiting the risk of short-term drawdowns, according to the report. Goldman’s latest report points to a clear shift in allocation priorities for the second half of the year toward “correction, rotation and diversification.” The report said companies where valuation, earnings and policy support reinforce one another could have greater potential to generate excess returns.
On Aug. 9, a team led by Goldman Sachs’ chief China equity strategist, Liu Jinjin, published a research report titled “AI Changes the Game: Correction, Rotation and Diversification.” The report reviewed the drivers of the latest rise and fall in China’s AI sector, used 10 quantitative signals to assess the maturity of the correction, and outlined an allocation strategy.
Goldman analysts said the latest decline has absorbed the sector’s key earlier risks. They cautioned, however, that volatility in technology stocks listed in Hong Kong and mainland China is likely to remain high in the near term. Investors should move beyond a single-track approach, the analysts said.
The report said China’s AI hardware stocks tracked by Goldman rose an average of 33% in the first half of the year, with computing power and semiconductors emerging as the strongest market themes. After global AI stocks began correcting in June, however, those shares fell sharply from their June highs. The STAR Market 50, ChiNext Index and CSI 1000 each dropped more than 20% from recent highs.
To assess the latest AI stock correction, Liu’s team developed 10 indicators covering market pricing, risk appetite, fund flows, positioning structure, industrial policy and corporate fundamentals. The indicators were designed to measure the current stage of the correction, potential downside risks and medium- to long-term return potential for AI hardware stocks.
Goldman said the decline had clearly absorbed four major risks that had built up earlier in the sector: speculative long positions had been largely closed, overall valuations had moved away from historical highs, onshore leverage had been substantially reduced, and crowded positioning in the sector had eased significantly.
Although those risks have been released in stages, Goldman warned that volatility in A-share and Hong Kong-listed technology stocks would remain elevated in the short term. The report advised against concentrating heavily in a single AI hardware segment. Instead, its core approach is to diversify across assets inside and outside the AI supply chain that have low correlations with one another, using a balanced portfolio to offset market swings.
The report identified four differentiated investment themes:
First, investors could add selected Hong Kong-listed internet and AI software companies in stages. After a prolonged correction, Hong Kong internet stocks offer more attractive valuations, while the rollout of AI applications could support an earnings recovery. Goldman said the group could provide balance outside AI hardware and help offset volatility in hardware stocks.
Second, investors should consider policy-supported themes while covering both domestic-demand and export-oriented AI companies. Goldman said investors should focus first on the Goldman Sachs 15-year investment portfolio as well as inward-focused and export-oriented artificial-intelligence companies.
Third, the report called for identifying nontechnology stocks that have been oversold and whose earnings expectations are being revised higher. In a high-volatility environment for technology stocks, some traditional-industry companies have undergone deep corrections and, combined with improved earnings expectations, may have room for valuation recoveries. Goldman described them as defensive core holdings that could smooth portfolio returns.
Fourth, investors could seek excess returns from initial-public-offering opportunities in the primary market. The new-stock market offers Alpha opportunities independent of the performance of AI stocks in the secondary market, the report said. Goldman highlighted listed companies with stable dividends and ample cash flow as the focus of that screening.
Liu’s team concluded that the current correction in AI hardware stocks represents a clearing of risks after a major rally, rather than the end of the industry’s long-term trend. In the short term, however, competition for capital is intensifying and the risk-reward appeal of a pure sector-concentration strategy has declined.
The report said investors should move away from a one-sided bet on AI hardware and instead combine software and hardware, balance technology and nontechnology holdings, and link primary- and secondary-market exposure. That approach could capture the AI industry’s long-term benefits while limiting the risk of short-term drawdowns, according to the report.
Goldman’s latest report points to a clear shift in allocation priorities for the second half of the year toward “correction, rotation and diversification.” The report said companies where valuation, earnings and policy support reinforce one another could have greater potential to generate excess returns.