Focus on Hardware Computing Power: Investment Value Analysis of STAR AI ETF Fullgoal (589380)

Keywords: STAR AI ETF; SSE STAR AI Index; hardware computing power; AI investment; STAR Market

Introduction

Artificial intelligence is reshaping the global economic landscape at an unprecedented pace. From large model training to edge inference, computing power, as the "water, electricity, and coal" of the AI industry, is increasingly strategically important. Against this backdrop, the STAR Market, as a gathering place for China's hard-tech enterprises, has produced a number of outstanding companies deeply involved in hardware computing power areas such as chips, optical modules, and servers. The STAR AI ETF Fullgoal (589380) tracks the SSE STAR AI Index, with its constituent stocks highly focused on the hardware computing power segment, providing investors with a tool product that precisely captures growth opportunities in AI underlying infrastructure.

SSE STAR AI Index: Based on Hardware, Focused on Computing Power Core

The SSE STAR AI Index (hereinafter referred to as the "STAR AI Index") selects listed companies from the STAR Market in AI-related fields. Its construction approach differs significantly from other broad AI indices – it does not aim to cover the entire AI industry chain but instead concentrates weight on the hardware computing power direction. According to the latest data, the top-weighted constituent stocks are concentrated in areas such as AI chips, optical modules, server manufacturing, and semiconductor equipment, including companies with hard-tech attributes like Sugon, Montage Technology, and Cambricon. This "heavy hardware, light software" allocation structure gives the STAR AI Index different elasticity characteristics compared to the overall market during AI rallies.

The above figure shows the industry distribution of the STAR AI Index, where the three major sectors – electronics (chip design, packaging and testing), communications (optical modules), and computers (servers) – together account for over 80%, fully confirming the index's tilt toward hardware computing power.

Hardware Computing Power: Foundation and Bottleneck of the AI Industry

Currently, the core contradiction in AI development lies in "supply-demand imbalance of computing power." For example, training a 100-billion-parameter large model like GPT-4 requires tens of thousands of GPUs working continuously for months, while inference deployment at the edge also imposes high computing demands. According to IDC, China's AI computing power scale will exceed 3,000 EFLOPS by 2027, with a compound annual growth rate of over 30% from 2022 to 2027. In this process, domestic substitution of hardware is particularly important – due to external technology controls, the self-sufficiency rate of domestic AI chips is still below 20%. A group of innovative companies on the STAR Market in areas such as GPU, ASIC, and DPU are accelerating breakthroughs.

The hardware segments covered by the STAR AI Index are exactly the directions with the most urgent domestic substitution needs and the highest prosperity. For example, optical modules, as core devices for data center interconnection, benefit from the volume ramp of 800G/1.6T high-speed products; AI servers face structural opportunities due to tight supply of Nvidia GPUs. Leading companies in these sub-sectors are mostly listed on the STAR Market and have R&D investment intensity far exceeding similar companies on the main board.

Investment Advantages and Risks

Advantages: Focused Track, High Elasticity

As a passively managed product, the primary advantage of the STAR AI ETF Fullgoal (589380) lies in its "purity" – it does not include non-AI constituent stocks and its exposure to hardware computing power far exceeds other STAR Market ETFs. With a highly concentrated portfolio (the top ten weights account for about 50%), the fund is extremely aggressive when computing power rallies erupt. Between 2023 and 2025, the STAR AI Index showed excess returns over the STAR 50 Index under every large model release or policy stimulus.

Additionally, the liquidity advantage and low fee structure (management fee 0.50% + custody fee 0.10% per year) make it one of the lowest-barrier tools for individual investors to participate in STAR Market computing power investment. Investors can allocate in one click through a fund account without needing to open a STAR Market stock trading account.

Risks: Volatility and Industry Concentration

However, the highly concentrated investment style also implies higher risks. The hardware computing power industry is a strongly cyclical area with high R&D investment. Individual company performance is easily affected by product iteration and downstream capital expenditure cycles. In the second half of 2024, due to an oversupply of Nvidia H100, the A-share computing power sector corrected by over 20%, and the STAR AI Index also experienced significant drawdown. Moreover, the characteristic "high volatility" of the STAR Market – individual stocks can fluctuate by up to 20% daily – makes ETF net value movements more volatile than main board products. Investors need strong risk tolerance.

Future Outlook: AI Hardware Still in Upward Cycle

Looking ahead to 2026-2027, three major drivers are expected to support the hardware computing power investment theme: first, large models shift from "parameter competition" to "application landing," triggering demand for on-device AI chips; second, domestic substitution moves from "verification period" to "volume period," with domestic chip supply chains maturing (e.g., Huawei Ascend, Haiguang DCU); third, supportive policies continue, with cities like Shanghai and Beijing issuing computing infrastructure construction plans, directly driving procurement of servers and optical modules. Under these combined factors, the hardware computing power track represented by the STAR AI ETF Fullgoal (589380) has medium- to long-term allocation value.

Conclusion

The STAR AI ETF Fullgoal (589380), by tracking the SSE STAR AI Index, provides investors with an extremely pure hardware computing power investment tool. It is suitable for long-term investors who believe "computing power is national power" and are optimistic about the logic of domestic AI chip and infrastructure substitution. However, it also requires holders to tolerate the high volatility inherent in the STAR Market. For rational investors who wish to share in the underlying technology dividends of AI rather than chasing short-term hotspots, this ETF deserves a place in their asset allocation framework.