TL;DR

• Goldman Sachs recommends buying into a basket of China's AI value chain, covering power, semiconductors, AI infrastructure, models, and applications.
• Goldman Sachs estimates that the market value related to China's AI is approximately $4 trillion, contributing about 16% of global AI-related revenue, but the allocation to China in the global mutual fund tech exposure is only about 1.2%.
• The core of this trade is not a single AI application explosion, but rather the revaluation opportunity brought about by under-allocated funds, policy investments, and hardware demand.
• The risk lies in whether data center investment, storage expansion, IPO financing, and AI hardware exports can continue to deliver.

The Goldman Sachs thematic research team is putting "China's AI value chain" in the center of trading vision.

According to their report titled "Trading Strategy: Go Long on the China AI Value Chain," Goldman Sachs recommends going long on a Chinese 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, the Nvidia supply chain, and cloud capital expenditures; what Goldman Sachs sees now is the misalignment between the market value, revenue contribution, and global fund holdings of Chinese AI assets.

According to Goldman Sachs’ estimates, Chinese AI-related companies have a combined market value of about $4 trillion, contributing about 16% of global AI-related revenue, but as of January 2026, global mutual fund managers have only allocated about 1.2% to China in their global tech exposure.

This set of numbers constitutes the most important trading logic of the entire report: if China's AI industry already holds a double-digit share of revenue, while global fund allocations remain significantly low, then there is room for the revaluation of China's AI value chain.

Maximum Contrast: Revenue Contribution Not Low, Global Fund Allocation Very Low

Goldman Sachs’ breakdown of global AI assets offers a very direct comparison.

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

However, the fund allocation is far below this ratio. Goldman Sachs estimates that as of January 2026, global mutual fund managers have allocated only about 1.2% to China in their global tech exposure.

This is also the core reason Goldman Sachs advocates going long on China's AI value chain. U.S. AI assets have already been repeatedly purchased by global funds, with Nvidia, cloud providers, semiconductor equipment, and power infrastructure included in the main line of AI trading. In contrast, while Chinese AI assets have already formed a certain revenue scale, they still remain under-allocated in the global fund positions.

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

This is Not Traditional KWEB Trading; Hardware and Infrastructure are More Prominent

Goldman Sachs particularly emphasizes that this trade is different from traditional KWEB trading.

KWEB typically corresponds to Chinese internet and platform economy exposure, where investors think of e-commerce, advertising, online entertainment, and local living. However, Goldman Sachs is constructing the GS China AI Value Chain (GSXACART) basket, which ranges from power, semiconductors, AI infrastructure, to models and applications, closely resembling a complete Chinese AI supply chain.

Within this framework, hardware and infrastructure take precedence.

China's push for technological self-sufficiency and advanced computing capabilities ensures that AI hardware, data centers, power support, and semiconductors simultaneously receive attention from policies, industries, and capital. Goldman Sachs believes that the value of these segments has not been fully reflected in the stock market.

Their research estimates that the potential economic benefits brought by AI through efficiency improvements and the creation of new profits may exceed the levels already reflected in current AI stock prices by 50% to 100%. This is also why power, AI infrastructure, and semiconductors are placed at the core of the basket.

The potential explosion of models and applications ultimately relies on computing power, storage, electricity, and supply of equipment. And these segments are precisely where China has scale manufacturing, engineering construction, and industrial support capabilities.

Exports, Policies, and IPOs are Strengthening AI Hardware Clues

The changes in China's AI hardware chain are transitioning from concepts to more concrete orders, exports, and financing nodes.

On the demand side, customs data, as cited by several media outlets, shows that China's exports in May grew by 19.4% year-on-year, the strongest increase in three months; among them, the export value of integrated circuits increased by about 111%, with only slight growth in export volume. Behind the changes in prices and structures, AI hardware demand is viewed as one of the important driving factors. For storage, semiconductor equipment, and upstream materials, this type of data points to the potential improvement of orders and capacity utilization.

On the policy investment front, according to Reuters citing Bloomberg, China is preparing a five-year plan of about 2 trillion yuan, or approximately $295 billion, to establish a nationwide AI data center network. This plan has not been officially announced, but if implemented, it will directly stimulate domestic demand for storage chips, semiconductor equipment, power support, and data center infrastructure.

On the capital market front, public reports indicate that A-shares, Hong Kong stocks, and some global indices have increased their weights in AI and semiconductors during the adjustments in 2026. This will enhance the visibility of passive funds towards relevant companies and will attract more domestic and foreign funds towards advanced computing and semiconductors.

Individual stocks and industry cases are also reinforcing this clue. Yangtze Memory Technologies Company reported a year-on-year revenue increase of about 445% in the first quarter of 2026, with its global NAND flash market share rising from 8% a year ago to 13%, jumping to a tie for fourth place, and pushing forward with its domestic IPO plan to support capacity expansion.

Changxin Memory Technologies is also seen as an important player in China's DRAM industry. Third-party research estimates that its revenue in 2026 could exceed $50 billion; the company’s prospectus indicates first quarter revenue of 50.8 billion yuan, with a half-year revenue guidance of 110 billion to 120 billion yuan.

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

Funds are Starting to Switch; U.S. AI Remains the Main Reference

Goldman Sachs also noted that the Chinese AI sector has outperformed other China-related assets and there are signs of fund allocation shifts. However, compared to U.S. AI, the performance of Chinese AI assets still lags significantly.

This is where the attractiveness of the trade and the boundaries of risk coexist.

The attractiveness lies in the fact that if global investors continue to seek growth lines outside of U.S. AI, the low allocation state of Chinese AI may leave space for fund switching. Especially when valuations in leading U.S. AI stocks are already high and capital expenditure expectations have been thoroughly discussed, the market naturally will look for supply chain and application assets that have not been sufficiently held.

The risk lies in the fact that this remains a trading recommendation, not an already realized industrial conclusion. The 2 trillion yuan AI data center plan depends on policy details and actual execution; the listings, capacity expansions, and profitability improvements of companies like Changxin Memory Technologies and Yangtze Memory Technologies will also require time; whether chip exports and sales data can continue will depend on the global AI hardware cycle and trade environment.

The U.S. AI remains the primary reference for global funds. Whether it's model capabilities, cloud providers’ capital expenditures, GPU ecology, or enterprise application revenues, the U.S. market still has more mature benchmarks. For Chinese AI to attract more global funds, it cannot merely prove "valuation is cheap, holdings are low," but must continuously deliver on revenue, profit, and technological progress.

The highlight of Goldman Sachs' going long on China's AI value chain is not to announce that Chinese AI has caught up with the U.S., but to place a market misalignment in front of us: about $4 trillion in market value, about 16% contribution to global revenue, yet corresponding to only about 1.2% allocation to China in global mutual fund tech exposure.

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

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。