Goldman Sachs: China AI Stocks Near $5 Trillion Market Cap, Two Investment Themes Identified — Exports vs. Domestic Demand

nashnova research
今天发布阅读约 13 分钟

Goldman Sachs reports that since DeepSeek's launch, China's AI stocks have generated nearly $5 trillion in market cap, now accounting for 42% of all A-share and Hong Kong-listed equities; two diverging tracks — export proxies and domestic champions — will shape the next phase of capital flows.

01

Where did the $5 trillion come from?

Since DeepSeek's early-2025 release, Chinese AI-related listed companies have added nearly $5 trillion in market cap through share-price gains and IPOs.
AI stocks now represent 42% of total A-share and Hong Kong market cap, up from ~20% in 2024, and account for 57% of trading volume. This means → more than half of every dollar traded in Chinese equities is an AI bet.
On earnings, Chinese AI companies carry a 2023–2028 expected EPS CAGR of 16%, versus just 4% for non-AI peers. In plain terms = a 4× growth gap is pulling capital decisively toward one side.
02

Track one: export-oriented AI companies — whose money are they making?

Goldman identifies roughly 360 export-oriented AI companies with a combined market cap of ~$1.7 trillion and overseas revenue exceeding 40%, concentrated in AI infrastructure and power equipment.
This means → these firms are essentially China-based proxies for Western AI capex — as global hyperscalers build out data centers, orders flow into the Chinese supply chain.
Cumulative returns from early 2025 to June 2026 reached 126%, but the group has fallen 23% from its peak since the late-June global AI selloff. It now trades at 21× 2027E P/E, near the bottom of its recent range, with consensus two-year EPS CAGR of 38%.
03

Track two: domestic AI champions — what supports the valuation?

Roughly 880 companies primarily serving the Chinese market (overseas revenue below 10%) carry a combined market cap of ~$3.2 trillion.
Three drivers underpin them: rising AI capex from China's hyperscale cloud providers, surging AI application demand, and direct policy support. During the recent global selloff, domestics outperformed exporters by ~10 percentage points.
The group trades at 23× forward P/E against 33% EPS CAGR, yielding a PEG of ~0.7×. Goldman flags a risk → some semiconductor-localization names carry heavy policy premiums, pushing valuations well above global peers. Overpaying for policy certainty is the key risk on this track.
04

How much is Beijing actually spending?

Ultra-long special treasury bonds: ~RMB 800 billion in 2026, earmarked for "dual priorities" — major strategic and major security initiatives, with the second tranche explicitly including AI.
National IC Fund Phase III: RMB 344 billion, targeting advanced packaging, equipment, and AI chips. National AI Industry Investment Fund: RMB 60 billion.
The 15th Five-Year Plan mandates R&D spending growth above 7% per year, reaching ~3% of GDP by 2030. This reflects a policy framework that treats AI as a national strategic priority on par with semiconductors — the funding scale implies a long-cycle commitment, not a short-term stimulus.
05

How far ahead is the "curated AI portfolio"?

Goldman's updated "China Curated AI Portfolio" holds 50 names (38 A-shares, 27 HK-listed, 3 ADRs), spanning 23 AI sub-industries. By sector: semiconductors 35.9%, AI applications 33.6%, infrastructure 21.2%, power 8.2%, large language models 1.2%.
The equal-weighted portfolio trades at 31× 12-month forward P/E, with consensus two-year earnings growth of 46% and 58% of revenue directly tied to AI operations.
Since its February 2025 inception it has returned 33%, versus 6% for the MSCI China Index, with a Sharpe ratio of 2.23 against 0.74 for the benchmark. In plain terms = for the same level of volatility, this portfolio delivered roughly 3× the benchmark return.
06

Why haven't foreign investors caught up?

Chinese AI companies account for ~10% of global AI market cap, yet mutual-fund allocation to Chinese tech stocks sits at only ~5.8% — a notable underweight.
This means → there is a clear gap between international capital's awareness of China AI and its actual positioning.
Whether this gap narrows as the export-versus-domestic divergence sharpens is, in Goldman's view, the key variable for future capital flows.

市场有风险,内容仅供研究参考,不构成投资建议。