DeepSeek has changed everything! Goldman Sachs proposes a new framework for investing in the Chinese stock market

DeepSeek has triggered a revaluation of Chinese assets, and Goldman Sachs has proposed a brand new AI investment framework for the Chinese stock market in its latest research report.

On February 17, Goldman Sachs analysts Kinger Lau and others stated in the report that the emergence of DeepSeek-R1, along with other recently launched globally competitive and cost-effective Chinese AI models, has changed the narrative around Chinese technology, enhancing investors' optimistic expectations regarding AI growth and economic benefits.

Goldman Sachs estimates that the widespread adoption of AI will increase the earnings per share (EPS) of Chinese companies by 2.5% annually over the next decade. Improved growth prospects and potential confidence boosts could raise the fair value of the Chinese stock market by 15-20%, potentially attracting over $200 billion in portfolio inflows.

Goldman Sachs has raised the target prices for MSCI China and CSI 300 to 85 and 4700, implying potential upside of 16% and 19% over the next 12 months, respectively.

Despite the immense potential of AI for China's growth trajectory, Goldman Sachs believes that strong policy stimulus is still needed to address macroeconomic challenges and drive sustainable stock market returns.

Goldman Sachs has also constructed a Chinese AI stock investment framework that includes six thematic sectors, including semiconductors, infrastructure, data and cloud computing, software and applications, revenue growth, and productivity enhancement.

Regarding investment opportunities for DeepSeek in the Chinese AI ecosystem, Goldman Sachs pointed out that offshore stocks have a higher exposure to the AI theme compared to A-shares, which is consistent with the former's better performance over the past month.

Goldman Sachs believes that among major benchmark indices, Hang Seng Tech, CSI 1000, ChiNext, and STAR Market are relatively more technology/AI-oriented. Thematically, Goldman Sachs is currently more optimistic about data and cloud as well as software and applications.

Chinese AI Investment Framework

Goldman Sachs has introduced a systematic framework to help investors understand the dynamics within the Chinese stock market's AI ecosystem:

First, Goldman Sachs divides the $14 trillion total market capitalization of Chinese stocks into two main categories: AI technology and non-technology;

Second, within the AI technology sector, the $6 trillion market capitalization is further segmented based on industry classification and business nature in the technology supply chain into semiconductors (including software design), infrastructure (hardware, data storage, cooling systems), data and cloud (e.g., internet platform companies), and software and applications (autonomous driving, biotechnology, humanoid robots, internet service providers, etc.);

Third, in the non-technology sector, the $7 trillion market capitalization is categorized into: 1) revenue enhancers, which may increase incremental revenue due to their relatively high future growth investments (i.e., capital expenditures and R&D); 2) productivity enhancers, which may improve operational efficiency and profitability through technology given their labor-intensive cost structure.

Finally, Goldman Sachs ranks various sub-industries and thematic groups based on their relative price sensitivity to NVIDIA and META, allowing for a better understanding of the trading extent of Chinese industries and AI agents in the two dominant trends in the tech world—capital expenditures and applications—and assisting investors in gaining reasonable exposure within the ecosystem as dynamics are changingBased on the above framework, Goldman Sachs found:

Influenced in part by the U.S. AI cycle, China's technology infrastructure and semiconductors have performed outstandingly over the past two years, thanks to significant capital expenditures in global computing capacity construction and upgrades. Both are currently trading above mid-cycle valuations, indicating that the optimism surrounding AI and the theme of China's technological self-reliance seem to be quite priced in;

As new capital expenditures in foundational technologies grow at a more moderate pace in both China and the U.S., DeepSeek may accelerate the adoption of AI in China and shift investors' focus from semiconductors and infrastructure to data and cloud as well as software and applications, which are currently trading below mid-cycle ranges and have relatively better profit growth potential;

In the tech cycles of the U.S. and China, revenue enhancers and productivity boosters have lagged so far, which may reflect the uncertainties of the cost-benefit trade-offs and socio-economic considerations (such as workforce reduction and deflation) that AI may bring. Investors will continue to focus on fundamental and policy factors to approach these potential late-stage AI beneficiaries until these benefits are better quantified and demonstrated