The first wave of ChiNext-listed computing power infrastructure exchange-traded funds has been established concurrently by three major public fund management companies: Tianhong Fund, China Asset Management, and E Fund Management.

According to the latest announcements, the Tianhong ChiNext Computing Power ETF and the ChinaAMC ChiNext Computing Power ETF are scheduled to begin trading on September 16th and September 17th, respectively. In addition, seven other asset management firms, including Guotai Fund, Dacheng Fund, and GF Fund, currently have related products either in the issuance process or in the preparatory stage.

Focus on Infrastructure Hardware

While other existing AI-focused ETFs in the market have varied investment strategies, the ChiNext Computing Power ETF specifically tracks the ChiNext Computing Power Infrastructure Index. This index carefully selects 50 ChiNext-listed components that span the hardware spectrum, covering crucial areas such as computing units, networking gear, data storage, and operational maintenance systems.

Li Haowei, the fund manager for the Dacheng ChiNext Computing Power ETF, explained that this index stands out in the market as one of the few with a pure hardware and infrastructure focus, deliberately excluding software, algorithms, and application-oriented companies. Wu Lei, a quantitative investment fund manager at Guotai Fund, provided further insights into the index's composition, noting that the communication and electronics sectors dominate with a combined weighting of nearly 70 percent, while the computer sector accounts for approximately 18 percent. In terms of market capitalization, the index primarily targets specialized and innovative companies listed on the ChiNext board with valuations ranging from 10 billion to 100 billion yuan.

Hong Minghua, who manages the Tianhong ChiNext Computing Power ETF, highlighted the index's significant concentration in domestic computing power, which carries a weight exceeding 70 percent. Within this allocation, internet data center operators represent over 40 percent of the index, positioning them to directly benefit from the surging demand for computational resources. The index is characterized by high growth and high elasticity; the ChiNext board's 20 percent daily price fluctuation limit grants it sharp market responsiveness, while its annualized volatility of over 47 percent signals robust offensive characteristics. This makes it a suitable option for investors who maintain a long-term positive outlook on AI trends. Looking at the fundraising results, Tianhong, ChinaAMC, and E Fund raised 414 million yuan, 332 million yuan, and 1.477 billion yuan respectively, attracting 4,976, 4,261, and 18,200 valid subscription accounts.

Enhancing Investor Portfolio Choices

With continuous advancements in computing infrastructure construction, the future investment opportunities in this sector and the value of these new ETFs have become key focal points for investors. Wu Lei emphasized that computing power serves as the fundamental foundation for the entire AI industry in the current era, asserting that no matter how powerful the models become—whether for video generation, image creation, or text processing—they are unsustainable without it. Global efforts are now accelerating to build out AI infrastructure. Industry projections indicate that global AI capital expenditure could reach 874.5 billion US dollars by 2026, representing a 60 percent year-over-year increase, and potentially climb to 1.5 trillion US dollars by 2027, marking a further surge of over 70 percent. As AI capital spending continues its upward trajectory, the business prospects across various sub-sectors of computing infrastructure remain highly promising.

Hong Minghua believes that a closed-loop business model for AI has been established, evidenced by the explosive growth in token call volumes and the accelerating adoption rate of AI agents. Domestic large language models, homegrown chips, and local applications have together formed a complete domestic AI ecosystem. The confluence of policy support, application deployment, and capital expenditure expansion presents historic development opportunities for domestic computing power. The launch of the ChiNext Computing Power ETF effectively resolves the stock-picking challenge within this sector, providing investors with a standardized and cost-effective tool for computing power allocation.

Li Haowei pointed out that the computing power supply chain is extensive and complex, spanning multiple segments including chips, optical modules, printed circuit boards, storage, data centers, and liquid cooling technologies. These segments each have their own distinct technical pathways and cyclical patterns, making individual stock research quite demanding. The ChiNext Computing Power ETF packages a basket of leading computing infrastructure companies into a single, diversified instrument with high purity and a balanced structure, offering retail investors a low-threshold and easy-to-participate vehicle for sharing in industry growth. Furthermore, it distinguishes itself from existing AI and semiconductor ETFs, positioning it as an appropriate satellite allocation within technology-focused growth portfolios. Additionally, as the index covers a substantial number of small and mid-cap constituents, it can complement broad-based indices and dividend-yielding assets, serving effectively as the growth component in a barbell-like investment strategy.