A significant new policy directive has been issued by eight government departments, including the Ministry of Industry and Information Technology and the National Development and Reform Commission.

The "Implementation Opinions on Promoting the High-Quality Development of the Industrial Internet" calls for enhancing computing power support and integrating the planning and simultaneous construction of Industrial Internet infrastructure with computing facilities like intelligent and supercomputing centers.

The document also proposes exploring the establishment of an industrial computing network system to improve the dynamic coordination of computing power across devices, edge, and cloud, meeting the diverse operational needs for computing, networking, storage, and application.

Leveraging an integrated computing network, it aims to strengthen interconnectivity, boost the supply of intelligent and edge computing power, and enhance the capacity for high-speed processing and deep analysis of massive heterogeneous data, empowering scenarios such as large industrial model training and real-time interaction in the industrial metaverse.

What's New in the "Industrial Internet + Computing Power Revolution"?

Industry insiders illustrate the shift with an example: previously, weld quality on an automotive production line relied on manual sampling by experienced workers. Later, visual inspection systems were introduced, sending images to the cloud for analysis.

However, the transmission delay of several hundred milliseconds meant defective parts could already have moved to the next station, limiting the system to remote monitoring rather than real-time control.

Now, the requirement is to deploy edge computing nodes (edge AI servers/GPU boxes) directly on the production line, enabling millisecond-level local image analysis and immediate feedback to adjust robotic arm parameters. This embodies the real meaning of "bringing computing power into the factory."

How Will the Computing Network and Industrial Internet Integrate in Factories?

Firstly, at the device level, future industrial robots, AGVs, and visual sensors will be equipped with built-in NPUs or inference chips, capable of running lightweight AI models locally for defect recognition and obstacle avoidance without relying on external networks.

Secondly, for edge computing, each workshop or factory zone will deploy edge computing nodes (industrial edge servers) to handle data aggregation, real-time analysis, and model inference from multiple devices, with latency controlled within a few milliseconds, forming the core layer of the "industrial computing network."

Thirdly, at the cloud/regional intelligent computing level, heavy-duty tasks such as pre-training large industrial models, plant-wide process optimization simulation, and cross-facility data lake analysis will utilize intelligent or supercomputing centers.

This three-tier architecture aims to create an integrated system for computing, networking, storage, and application. The policy specifically emphasizes strengthening the capacity for high-speed processing and deep analysis of massive heterogeneous data to empower industrial large model training.

The Implementation Opinions also call for initially building a relatively complete foundational system for industrial data and strengthening the national Industrial Internet big data center system.

Market analysts point out that the true value of industrial data elements will only be unlocked when data can be clearly attributed, circulated, and traded. In the coming years, companies involved in industrial data governance, industrial dataset construction, and industrial large-model platforms may see sustained policy support and market demand.

Investment Vehicles for This Trend

The Big Data ETF (516700) passively tracks the CSI Data Index, which is deeply tied to domestic computing power (IDC, servers) and AI application fields, covering the entire data technology process including storage, production, analysis, operation platforms, and application, reflecting the overall development of China's big data industry.

As of the end of May, the index's constituent stocks include concepts such as cloud computing (88.41% weight), IDC/computing power leasing (44.80%), computing power (43.97%), and AI applications (27.84%).

The Information Technology Innovation ETF (562030) and its feeder funds (Class A: 024050, Class C: 024051) passively track the CSI Information Technology Innovation Index, focusing on the field of independent and controllable information technology, covering core segments of the IT innovation industrial chain like basic hardware, basic software, application software, information security, and peripheral equipment.

Its top holdings include leaders in storage chips, domestic computing power, and AI applications. Influenced by relevant policies, orders for related software and hardware are expected to accelerate.

The Cloud Computing ETF (Subscription Code: 159099) passively tracks the CSI Cloud Computing 50 Index, comprising stocks of companies involved in providing cloud computing services (IaaS, PaaS, SaaS) and hardware for cloud computing.

Its top ten holdings cover leading companies in core areas such as optical modules, servers/AI computing power, data centers, and software platforms, aiming to capture the investment opportunities from AI computing power expansion and the upturn in cloud infrastructure.

ETF Fee Information

The Big Data ETF and the Information Technology Innovation ETF do not charge sales service fees. Subscription and redemption agents may charge a commission of up to 0.5%, which includes fees charged by exchanges and registration institutions. On-market trading fees are subject to the rates set by securities firms.

For the Cloud Computing 50 ETF, the subscription fee is 0.3% for amounts below 1 million yuan, and 1,000 yuan per transaction for amounts of 1 million yuan or above. On-market trading fees are subject to the rates set by securities firms. When subscribing or redeeming fund units, agents may charge a commission of up to 0.3%. Please refer to the fund's legal documents for detailed fee structures.

Risk Disclosure

The Big Data ETF passively tracks the CSI Big Data Industry Index (base date: Dec 31, 2012; release date: Oct 18, 2016). The Information Technology Innovation ETF passively tracks the CSI Information Technology Innovation Index (base date: Dec 29, 2017; release date: Dec 21, 2012). The Cloud Computing 50 ETF passively tracks the CSI Cloud Computing 50 Index (base date: Dec 31, 2014; release date: June 12, 2020). The composition of index constituents is adjusted according to the index methodology, and its past performance does not predict future results.

Constituent stocks mentioned are for illustrative purposes only; individual stock descriptions are not investment advice and do not represent the holdings or trading activities of the fund manager. The fund manager assesses the risk rating of these ETFs as R3-Medium Risk, suitable for Balanced (C3) and above investors. Suitability matching opinions are subject to the selling institutions.

Any information presented (including but not limited to stocks, commentary, forecasts, charts, indicators, theories, etc.) is for reference only. Investors are responsible for their own investment decisions. Furthermore, any views, analysis, or forecasts do not constitute investment advice, and no responsibility is taken for any direct or indirect losses resulting from the use of this content. Fund investment carries risk. Past fund performance does not guarantee future results, and the performance of other funds managed by the manager does not constitute a guarantee for this fund's performance. Invest with caution.