Goldman Sachs bullish on China AI: Global funds with 4 trillion market cap only allocate 1.2%

Goldman Sachs recommends going long on China's AI value chain, pointing to a significant gap between a $4 trillion market cap and 16% global revenue contribution, with only about 1.2% allocation by global funds.

(Previous Summary: Temasek discloses holdings in Anthropic & OpenAI! AI investment target 15%) (Background: Nvidia officially launches global "compute power for revenue share" plan! Startups can skip buying GPUs, use future profits for compute power)

Table of Contents

Toggle

• Goldman Sachs recommends long China AI value chain
• Funding gap between market cap and revenue contribution
• Hardware and infrastructure prioritized
• Orders, policy, and capital market triple drivers
• Trading appeal and risks coexist

Goldman Sachs' thematic research team is pushing the "China AI value chain" to the center of trading focus.

According to its report titled "Trading Strategy: Long China Artificial Intelligence Value Chain," Goldman Sachs recommends going long on a China AI basket covering power, semiconductors, AI infrastructure, models, and applications. Over the past two years, global AI trading has been dominated by large U.S. tech stocks, Nvidia's supply chain, and cloud capital expenditure; Goldman Sachs is now eyeing the mismatch between the market cap, revenue contribution, and global fund holdings of China's AI assets.

According to Goldman Sachs' assessment, Chinese AI-related companies already have a market cap of about $4 trillion, contributing about 16% of global AI-related revenue, but as of January 2026, global mutual fund managers had only about 1.2% allocation to China within their global tech exposure.

These figures form the core trading logic of the entire report: if China's AI industry already accounts for a double-digit share of revenue, while global fund allocation remains significantly low, then there is room for China's AI value chain to be repriced.

Goldman Sachs' breakdown of global AI assets provides a direct comparison.

Goldman Sachs recommends long China AI value chain

Since the end of 2022, global AI-related stocks have created about $34 trillion in market cap, with China's AI-related market cap at about $4 trillion, accounting for about 10% of the global AI market cap. In terms of revenue, China contributes about 16% of global AI-related revenue.

Fund allocation is far below this ratio. Goldman Sachs estimates that as of January 2026, global mutual fund managers had only about 1.2% allocation to China within their global tech exposure.

This is also the core reason Goldman Sachs proposes going long on China's AI value chain. U.S. AI assets have been repeatedly bought by global funds, with Nvidia, cloud vendors, semiconductor equipment, and power infrastructure all included in the main AI trading line. In contrast, although China's AI assets have formed a certain scale of revenue, they remain underweighted in global fund positions.

In other words, Goldman Sachs is betting not on a simple "China AI narrative," but on a more specific funding allocation gap: revenue contribution has already emerged, but global holdings have not yet caught up.

Goldman Sachs specifically emphasizes that this trade is different from the traditional KWEB trade.

Funding gap between market cap and revenue contribution

KWEB typically corresponds to China's internet and platform economy exposure, with investors thinking of e-commerce, advertising, online entertainment, and local life. But this time, Goldman Sachs has constructed a GS China AI Value Chain (GSXACART) basket, covering from power, semiconductors, and AI infrastructure to models and applications, more akin to a complete China AI supply chain.

In this framework, hardware and infrastructure are positioned more prominently.

China's push for technological self-reliance and advanced computing capacity construction has brought AI hardware, data centers, power support, and semiconductor sectors under simultaneous policy, industry, and capital attention. Goldman Sachs believes that the value of these sectors has not yet been fully reflected in the stock market.

Its research estimates that the potential economic benefits from AI through efficiency improvements and creating new profits could be 50% to 100% higher than the level already reflected in current AI stock prices. This is also why power, AI infrastructure, and semiconductors are placed at the core of the basket.

Hardware and infrastructure prioritized

Whether models and applications can explode ultimately depends on the supply of compute power, storage, electricity, and equipment. And these sectors are precisely where China has capabilities in large-scale manufacturing, engineering construction, and industrial supporting chains.

Changes in China's AI hardware chain are moving from concept to more specific orders, exports, and financing milestones.

On the demand side, customs data cited by multiple media outlets shows that China's exports in May grew 19.4% year-over-year, the strongest increase in three months; among them, integrated circuit export value surged about 111% year-over-year, while export volume increased only slightly. Behind the price and structural changes, AI hardware demand is seen as one of the important drivers. For storage, semiconductor equipment, and upstream materials, such data points to the possibility of improved orders and capacity utilization.

On the policy investment side, according to a Bloomberg report cited by Reuters, China is preparing a five-year plan worth about 2 trillion yuan, approximately $295 billion, to build a national AI data center network. The plan has not yet been officially announced, but if implemented, it would directly drive demand for domestic storage chips, semiconductor equipment, power support, and data center infrastructure.

On the capital market side, public reports show that A-shares, Hong Kong stocks, and some global indices have increased the weight of AI and semiconductors during the 2026 adjustments. This will increase the passive fund visibility of related companies and also channel more domestic and foreign funds toward advanced computing and semiconductors.

Orders, policy, and capital market triple drivers

Individual stock and industry cases are also reinforcing this thread. Yangtze Memory Technologies (YMTC) saw first-quarter 2026 revenue surge about 445% year-over-year, its global NAND flash market share rising from 8% a year ago to 13%, jumping to a tie for fourth place, and advancing its domestic IPO plan to support expansion.

ChangXin Memory Technologies (CXMT) is considered an important company in China's DRAM industry. Third-party research estimates its 2026 revenue may exceed $50 billion; according to the company's prospectus, first-quarter revenue was 50.8 billion yuan, with a first-half revenue guidance of 110 billion to 120 billion yuan.

These cases do not mean that Chinese storage companies have fully caught up with overseas giants, but they show that China's AI hardware chain is moving from a "policy concept" to more observable milestones in revenue, market share, financing, and expansion.

Goldman Sachs also notes that China's AI sector has outperformed other China-related assets and shows signs of fund allocation shifts. However, compared to U.S. AI, Chinese AI asset performance still lags significantly.

Trading appeal and risks coexist

This is also where the trade's appeal and risk boundaries coexist.

The appeal lies in the fact that if global investors continue to seek growth lines beyond U.S. AI, China's underweighted status may leave room for fund rotation. Especially after U.S. AI leaders have already reached high valuations and capital expenditure expectations have been fully discussed, the market will naturally look for supply chain and application assets that are not yet fully held.

The risk is that this is still a trading recommendation, not an already realized industry conclusion. The 2 trillion yuan AI data center plan depends on policy details and actual execution; the listing, expansion, and profit improvement of companies like CXMT and YMTC also take time; whether chip export and sales data can be sustained also depends on the global AI hardware cycle and trade environment.

U.S. AI remains the main benchmark for global funds. Whether in model capabilities, cloud vendor capital expenditure, GPU ecosystem, or enterprise application revenue, the U.S. market still has more mature benchmarks. For China's AI to attract more global funds, it must not only prove "cheap valuation, low holdings" but also consistently deliver revenue, profits, and technological progress.

The highlight of Goldman Sachs' recommendation to go long on China's AI value chain this time is not to declare that China's AI has caught up with the U.S., but to present a market mismatch: about $4 trillion in market cap, about 16% of global revenue contribution, yet corresponding to only about 1.2% China allocation in global mutual fund tech exposure.

Whether funds can fill this gap will depend on whether policy investment, hardware demand, and corporate earnings can continue to materialize.