In 2025, the artificial intelligence sector emerged as the most prominent leading theme in the A-share bull market, capturing intense capital focus and sustaining one of the most enduring market narratives. Against this backdrop, the ChinaAMC ChiNext AI ETF (159363) delivered a remarkable performance report: according to its annual fund data, the ETF generated a total profit of 1.724 billion yuan for the year, ranking first among comparable ETFs in terms of returns provided to investors.
This figure represents more than just profit—it reflects tangible gains for investors participating in the AI industry's growth. It also raises an important question for the market: how did the ChinaAMC ChiNext AI ETF (159363) achieve such substantial returns for investors during a highly polarized market environment in 2025? The answer likely lies in both the performance of its underlying index and the composition of its constituent stocks.
Strong Performance: Leading in Short Term, Sustaining in Long Term In 2025, driven by catalysts such as the explosive popularity of DeepSeek, increased capital expenditure by overseas cloud providers, and accelerated implementation of AI applications, the ChiNext AI Index recorded a cumulative gain of 106.35%. During the same period, the ChiNext Index rose by 49.57%, meaning the AI index outperformed it by more than 56 percentage points, demonstrating significant alpha generation.
Compared to other mainstream AI-themed indices, the ChiNext AI Index also showed notable outperformance. As of March 31, 2026, over a one-year period, the index surged 110.26%, ranking first among five peer indices including Sci-Tech Innovation AI (87.38%), Artificial Intelligence (63.05%), CS Artificial Intelligence (49.14%), and STAR AI (21.00%).
Over longer time horizons, the index's lead widened further. The ChiNext AI Index rose 189.20% over three years and 269.75% over five years, ranking first in both metrics among its peers. During the same periods, the second-best performing Sci-Tech Innovation AI Index gained 131.87% and 156.32%, respectively, indicating that the ChiNext AI Index's long-term excess returns continued to expand.
Note: Data as of March 31, 2026. The ChinaAMC ChiNext AI ETF passively tracks the ChiNext AI Index, which has a base date of December 28, 2018, and was launched on July 11, 2024. The index's annual returns from 2021 to 2025 were 17.57%, -34.52%, 47.83%, 38.44%, and 106.35%, respectively. Constituent stocks are adjusted according to the index methodology, and past performance does not guarantee future results.
Whether evaluated by short-term momentum or long-term endurance, the ChiNext AI Index has consistently demonstrated significant and sustained outperformance relative to peers. Leading peers over one year, expanding its advantage over three years, and delivering nearly 2.7 times returns over five years—this progression of performance metrics likely underpins the ETF's ability to generate over 1.7 billion yuan in profits for investors.
High Precision: Computing Power as Foundation, Applications as Catalyst If performance data reflect the outcome, the structure of constituent stocks reveals the cause. The ChiNext AI Index's ability to consistently outperform across short-, medium-, and long-term horizons is closely tied to its unique industry positioning.
In terms of industry chain distribution, the index exhibits a clear dual-engine drive characteristic, with computing power as the primary focus and applications as the secondary driver. As of March 31, 2026, the computing power segment accounted for 70.8% of the index weight, while AI applications made up 29.2%. This structure aligns precisely with the core logic of current AI industry development: computing power serves as the "infrastructure" of AI, the first-mover segment in commercialization, and the area with the most certain near-term earnings visibility.
Note: Data as of March 31, 2026. Constituent stocks are for illustrative purposes only; individual stock descriptions do not constitute investment advice and do not represent the fund manager's holdings or trading activities.
A deeper look into the computing power segment shows optical communications accounting for 45% of the weight, making it the largest sub-sector. Leading optical module companies—dubbed "Yi Zhong Tian" (referring to XYS, ZJXC, and TFTX)—serve as key index constituents. Data centers (IDC) represent 15%, chips account for 8.4%, and communication equipment makes up 2.4%. This composition allows the index to capture the high-growth optical module segment—which benefits most from surging AI computing demand—while maintaining balanced exposure across IDC, communication equipment, and chips, forming a complete computing power industry chain.
On the applications side, the index includes 26 carefully selected AI application companies with proven commercialization potential, covering representative firms such as THS, BlueFocus, ISoftStone, and EHualu. These span multiple verticals including fintech, marketing services, software solutions, and data infrastructure. Together, these AI application constituents represent a critical link in the transition of AI technology from "infrastructure" to "industrial enablement."
Computing power provides earnings certainty, while applications offer growth potential—this dual-engine structure is the core logic behind the ChiNext AI Index's sustained competitiveness across different industry cycles. At present, the AI industry is transitioning from "theme speculation" to "earnings realization," further strengthening the dual-engine investment thesis.
Timely Opportunity: From Narrative to Earnings At the current juncture, the AI industry is shifting from thematic speculation to a new phase of earnings delivery, with the dual-engine logic of computing power and applications gaining further traction.
On the computing power side, surging demand coupled with technological advancements continues to drive upward momentum. In Q1 2026, average daily token usage in China exceeded 140 trillion, representing a thousand-fold increase from early 2024. The 1.6T optical module segment is entering its first year of mass adoption, with supply constraints tightening and leading manufacturers strengthening their advantages. Recent guidance from the Ministry of Industry and Information Technology exploring innovative models such as "computing power banks" directly benefits the computing infrastructure industry chain.
On the applications side, commercialization is accelerating, pointing to significant earnings potential. 2026 is widely regarded as the first year of large-scale adoption of AI agents, with applications evolving from "tools" to "intelligent agents" capable of autonomous task execution. Breakthroughs have already occurred in verticals such as finance, marketing, and programming, with some companies expected to achieve profitability at scale, further expanding the commercial potential of AI applications.
As AI commercialization moves from narrative to earnings, the dual themes of computing power and applications are poised to create a dual resonance of profits and valuation expansion. To capture investment opportunities in both AI computing and applications, investors may consider the ChinaAMC ChiNext AI ETF (159363)—the largest and most liquid ETF in the dual-engine AI segment—along with its feeder funds (Class A: 023407, Class C: 023408), offering direct exposure to the growth dividends of AI technology commercialization.
Data source: Shanghai and Shenzhen Stock Exchanges. As of March 31, 2026, the ChinaAMC ChiNext AI ETF had a net asset value of 5.564 billion yuan, with an average daily trading volume exceeding 700 million yuan over the past six months. It ranks first in both size and trading volume among 26 ETFs tracking the ChiNext AI Index, STAR AI Index, and Sci-Tech Innovation AI Index.
ETF fee information: Subscription and redemption agents may charge a commission of up to 0.5% when investors subscribe or redeem fund units. Trading fees for on-market transactions are subject to securities firms' actual charges; no sales service fee is levied.
Feeder fund fee information: Class C of the ChinaAMC ChiNext AI ETF Feeder Fund does not charge a subscription fee. A redemption fee of 1.5% applies for holdings under 7 days, and 0% for 7 days or more. A sales service fee of 0.3% is charged annually. For Class A, subscription fees are 1% for amounts below 1 million yuan, 0.6% for 1-2 million yuan, and a flat 1,000 yuan for 2 million yuan or above. Redemption fees are 1.5% for holdings under 7 days and 0% for 7 days or more. No sales service fee is charged.
Risk disclosure: The ChinaAMC ChiNext AI ETF passively tracks the ChiNext AI Index, which has a base date of December 28, 2018, and was launched on July 11, 2024. The index's annual returns from 2021 to 2025 were 17.57%, -34.52%, 47.83%, 38.44%, and 106.35%, respectively. Constituent stocks are adjusted according to the index methodology, and past performance does not guarantee future results. Index constituents mentioned are for illustrative purposes only and do not constitute investment advice or reflect the fund manager's holdings or trading activities. The fund manager assesses this fund as R4—medium to high risk—suitable for aggressive (C4) or higher risk-rated investors. Suitability assessments are subject to sales institutions' evaluations. All information presented (including but not limited to stocks, commentary, forecasts, charts, indicators, theories, and any form of expression) is for reference only. Investors are solely responsible for their investment decisions. Furthermore, no views, analysis, or predictions herein constitute investment advice of any kind, and no liability is accepted for direct or indirect losses resulting from the use of this content. Fund investments carry risks; past performance does not indicate future results. The performance of other funds managed by the fund manager does not guarantee this fund's future performance. Invest with caution.
Comments