The AI wave is fundamentally transforming the structure and investment logic of China's stock market.

According to the latest research report from Goldman Sachs (NYSE: GS), since the DeepSeek moment in early 2025, China's AI-related listed companies have collectively created nearly US$5 trillion in market value through share price gains and IPOs. AI stocks now account for 42% of the total market capitalization of A-shares and H-shares, and 57% of trading volume, making them the dominant narrative in China's capital markets.

The report clearly divides China's AI investment opportunities into two main paths: first, export-oriented AI companies that serve as proxies for the Western AI capital expenditure boom; second, domestic demand-driven AI leaders that benefit from China's local AI ecosystem development and strong policy support. Goldman Sachs believes that despite lingering geopolitical and trade policy risks, both paths offer considerable investment value.

On the market impact front, Goldman Sachs data shows that Chinese AI exporters delivered a cumulative return of 126% from early 2025 to June 2026, but have fallen 23% from their peak since the global AI stock correction at the end of June. Domestic AI companies have shown greater resilience during this adjustment, outperforming by approximately 10 percentage points. Goldman Sachs also updated its "China Select AI Portfolio" of 50 stocks, which has gained 33% cumulatively since its inception in February 2025, with a Sharpe ratio of 2.23, significantly outperforming the MSCI China Index's 0.74.

AI Reshapes China's Stock Market: Nearly US$5 Trillion in Market Value Created

The emergence of DeepSeek marked a historic turning point in China's AI investment narrative. According to Goldman Sachs statistics, from early 2025 to now, the market value increment of China's AI-related listed companies has approached US$5 trillion, with share price gains contributing approximately US$2.8 trillion and IPOs contributing about US$2 trillion.

From a market structure perspective, the penetration of the AI theme into China's stock market has risen substantially. AI stocks' share of total A-share and H-share market capitalization has climbed from about 20% in 2024 to the current 42%, while their share of trading volume has reached 57%. Their share of IPO fundraising has also jumped from 4% in 2024 to 13% year-to-date in 2026.

On earnings fundamentals, the divergence between Chinese AI companies and non-AI companies is equally striking. Goldman Sachs data shows that Chinese AI companies are expected to deliver a compound EPS growth rate of 16% from 2023 to 2028, versus only 4% for non-AI companies. Within the global AI equity universe, Chinese AI companies represent approximately US$7.6 trillion in market value (weighted by AI business exposure), accounting for about 10% of global AI market capitalization, 24% of global AI revenue, and 14% of capital expenditure.

US-China AI Bipolar Landscape Takes Shape: China Strengthens Domestic Ecosystem

The Goldman Sachs report notes that the narrative of a global AI bipolar landscape is gaining broad recognition among investors—an AI ecosystem centered on the United States developing in parallel with one centered on China, driven by multiple dimensions including geopolitics, national security, and economic growth.

China already occupies a pivotal position in the global AI supply chain. In segments including rare earth refining, power equipment, PCBs, optical modules, and mature-node semiconductors, China's global market share exceeds 50%. At the same time, China is gradually building a domestic-circulation, increasingly self-sufficient AI value chain, achieving breakthroughs in advanced packaging, open-source large language models (LLMs), computing power production and consumption, and physical AI applications.

From a financial data perspective, Chinese AI companies have a combined market value of US$7.6 trillion, with an expected net profit margin of 12% and ROE of 16% for 2026. By comparison, AI companies in the rest of the world show corresponding figures of US$55.1 trillion, 24%, and 29%, respectively, leaving a still-wide gap.

Two Paths Diverge: Exporters and Domestic Leaders Each Have Their Own Logic

Goldman Sachs divides China's AI equity universe (over 1,700 names with market capitalization exceeding US$100 million) into two camps based on revenue source geography, with the two showing clear divergence in drivers, risk exposure, and market performance.

Export-oriented AI companies: Goldman Sachs identified more than 360 listed companies with a combined market value of approximately US$1.7 trillion, whose overseas revenue share exceeds 40%. These companies are mainly concentrated in AI infrastructure and power equipment, and are viewed by investors as proxies for the Western AI capital expenditure boom. Historically, this group delivered a cumulative return of 126% from early 2025 to June 2026, leading domestic AI companies by 84 percentage points, mainly due to stronger earnings delivery and growth expectations. However, since the global AI stock correction at the end of June, this group has fallen 23% from its peak, with its relative advantage narrowing by about 10 percentage points. Goldman Sachs analysis shows that export-oriented AI companies are more sensitive to global AI equity beta and the US-China relations barometer.

Domestic AI leaders: Goldman Sachs identified more than 880 listed companies primarily serving the domestic market (with overseas revenue share below 10%), with a combined market value of approximately US$3.2 trillion. This group benefits from the rapid growth of AI capital expenditure by China's hyperscale cloud service providers, the surge in AI application demand, and strong policy support for technological self-reliance. Compared with export-oriented companies, domestic AI leaders are more sensitive to China's AI capital expenditure trends and domestic AI policy developments, and have shown stronger defensiveness during the recent global AI stock correction, offering certain diversification value for international investors.

Path One: China AI Export Leaders—Balancing Growth and Valuation

While acknowledging geopolitical and trade policy risks, Goldman Sachs remains strategically optimistic on Chinese AI exporters, believing they are poised to further expand global market share through cost competitiveness, manufacturing capability, R&D intensity, and global execution.

In terms of screening criteria, Goldman Sachs focuses on the following characteristics: already-high global market share (low risk of displacement), accelerating overseas revenue growth (continuous market share expansion), deep technological moats (higher R&D intensity than peers), strict capital expenditure discipline (positive free cash flow), and relatively low sensitivity to geopolitical uncertainty.

The resulting "China AI Export Leaders" list contains 20 stocks rated Buy/Neutral by Goldman Sachs, covering 12 AI-related industries including optical modules, humanoid robots, autonomous driving, healthcare, consumer hardware, data centers, storage, and PCB/CCL. The portfolio derives an average of 66% of its revenue from overseas markets in the first half of 2026, currently trades at 21 times 2027 expected earnings, and has a sell-side consensus compound EPS growth rate of 38% over the next two years. Goldman Sachs notes that the portfolio's valuation has fallen back to the lower end of its recent range, with a current 12-month forward P/E of approximately 16.5 times.

Path Two: Domestic AI Champions—Policy Dividends and Scarcity Value

Developing AI has been elevated to a national policy goal. The Goldman Sachs report shows that policy support covers three dimensions: supply side (government-led investment), demand side (consumption promotion measures), and capital markets (IPO green channels, improved market liquidity). The breadth of this support is considered to exceed previous industry-wide easing cycles. Specific measures include: approximately RMB 800 billion from ultra-long special government bonds in 2026 directed at "Two Major" areas (with AI explicitly included in the second batch); RMB 344 billion for National Big Fund Phase III, focused on advanced packaging, equipment, and AI chips; RMB 60 billion for the National AI Industry Investment Fund; and the 15th Five-Year Plan explicitly requiring R&D spending growth of over 7% annually, reaching about 3% of GDP by 2030.

Goldman Sachs forecasts that the total addressable market for China's 21 AI-related industries could reach US$3 trillion to US$4 trillion by 2030, corresponding to a compound growth rate of approximately 28% from 2026 to 2030.

The "Domestic AI Champions" list also contains 20 names, covering 11 AI-related industries, with hard tech (semiconductor equipment, IC design, storage, wafer foundry) and soft tech (entertainment media, e-commerce, large models) each accounting for roughly half. The portfolio derives an average of over 90% of its revenue from the domestic market, currently trades at 23 times forward earnings, corresponding to 33% compound EPS growth and a PEG of approximately 0.7 times. Goldman Sachs cautions that some domestic semiconductor localization names carry high policy premiums, with valuations already significantly above global peers, and that over-chasing policy certainty is the main risk for this path.

Portfolio Strategy: 50 Stocks Covering the Entire AI Value Chain

Beyond the two main paths, Goldman Sachs also updated its "China Select AI Portfolio," adding and removing 15 stocks each, now containing 50 names (38 A-shares, 27 H-shares, and 3 ADRs, including dual-listed stocks), covering 23 specific AI industries, aimed at providing investors with balanced exposure across the entire AI value chain.

Structurally, the portfolio retains a relatively high allocation to upstream segments (power, semiconductors, infrastructure) while incorporating recently listed AI large model companies. By industry tier, semiconductors account for 35.9%, AI applications 33.6%, infrastructure 21.2%, power 8.2%, and AI large models 1.2%.

In terms of financial metrics, the portfolio (equal-weighted) currently trades at 31 times 12-month forward P/E, with consensus earnings growth of 46% over the next two years, and 58% of revenue directly tied to AI-related businesses. Since its inception in February 2025, the portfolio has gained 33% cumulatively, while the MSCI China Index rose 6% over the same period, with a Sharpe ratio of 2.23, far above the benchmark's 0.74.

Goldman Sachs also notes that international investors remain significantly underweight on Chinese AI stocks—Chinese AI companies account for about 10% of global AI market capitalization, but mutual funds' actual allocation to Chinese tech stocks is only about 5.8%, representing a notable underweight that may provide potential room for subsequent capital inflows.