The AI trade continued to adjust on Thursday (July 23rd), with divergences remaining in the CPO (co-packaged optics) optical module segment, leading to a mixed performance among individual stocks. Liantek Technology led gains, rising 5.86%, while Zhongji Innolight rose over 1%. Conversely, Xinyisheng and Tianfu Communication continued their decline, falling over 2%.
Among popular ETFs, the ChiNext AI ETF (159363), which is heavily weighted in leading CPO optical module stocks, saw an initial rise followed by a pullback, closing down 1.99% for the day. It continues to fluctuate around its 200-day moving average, with a net daily subscription of 34 million units.
The optical module sector has been experiencing sustained volatility recently, reigniting market debate. What stage is the current adjustment in, and how should the outlook be assessed? This article will explore these questions from two perspectives: a historical review and an industry outlook.
Learning from History: Past Pullbacks Consistently Disproven by Industry Trends
A research report from Guosheng Securities reviewed four typical pullbacks in the optical communication sector in recent years. The key takeaway is that market fears and divergences based on short-term linear extrapolation were ultimately disproven by the long-term certainty of AI computing power expansion.
In early 2024, fears of "copper in, optics out" emerged. However, copper cables are limited to within server racks, actually reinforcing the core status of optical modules.
In the second half of 2024, concerns arose about a slowdown in North American capital expenditure (Capex). Yet, starting from Q2, the year-on-year growth rate of capital expenditure from the four major cloud providers exceeded 50% and continued to rise.
In Q1 2025, the impact of DeepSeek and trade tariffs were concerns. Jevons' paradox proved true, as efficiency gains instead stimulated a surge in demand.
In mid-2026 (present), rising memory costs squeezing budgets became a worry. This underestimated the visibility of order books and earnings certainty for optical module companies.
Looking at history, the report suggests that divergence points represent buying opportunities. The fundamental upward trajectory of the optical communication industry remains unchanged, with leading companies' innovation cycles and corporate governance building solid moats. Short-term logical disruptions are ultimately absorbed by long-term earnings delivery, making each pullback a window for strategic industry positioning. The report maintains a positive outlook on the mid-year reporting season for the communications sector, recommending a focus on leading optical module companies.
Future Opportunities: Heavyweight Bets Validate Computing Power Demand
Reports indicate that OpenAI, Anthropic, Microsoft, SpaceX AI, and Google have all unveiled AI investment plans within less than two days, with intensity not diminishing but increasing. Google has raised its 2026 capital expenditure forecast to $195-205 billion and explicitly plans continued expansion into 2027. OpenAI has increased its 2030 computing power expenditure plan from $600 billion to $750 billion. Anthropic has signed a chip supply deal with AMD and secured a $5 billion investment.
From an industry trend perspective, this "heavyweight capital injection" validates the long-term certainty of AI computing infrastructure expansion, which will directly drive demand in the computing power supply chain, including optical modules. The potential for long-term investment continues to be revised upward. This sustained commitment from tech giants provides strong underlying demand support for the optical communication sector. The market divergence triggered by short-term pressures may once again constitute a window for positioning in optical modules.
The ChiNext AI ETF (159363) and its off-exchange feeder funds focus on leading CPO optical module companies. Its underlying index has an approximate 40% combined weighting in Zhongji Innolight, Xinyisheng, and Tianfu Communication, making it a core vehicle for AI computing power exposure. Furthermore, the ChiNext AI ETF (159363) has a latest AUM exceeding 7.4 billion yuan, with an average daily turnover over the past six months exceeding 1 billion yuan, leading its peer group of eight ETFs tracking the same index in both size and liquidity.
ETF Fee Information: When subscribing for or redeeming fund shares, subscription/redemption agents may charge a commission not exceeding 0.5%. On-exchange trading fees are subject to the actual charges by securities firms, with no sales service fee.
Feeder Fund Fee Information: The ChiNext AI ETF Feeder Fund C Class 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% applies. For the ChiNext AI ETF Feeder Fund A Class, subscription fees are 1% for amounts below 1 million yuan, 0.6% for 1-2 million yuan, and a flat 1000 yuan per transaction for 2 million yuan or more. A redemption fee of 1.5% applies for holdings under 7 days, and 0% for 7 days or more. No sales service fee is charged.
Risk Disclosure: The ChiNext AI ETF passively tracks the ChiNext Artificial Intelligence Index. The base date for this index is December 28, 2018, and its release date is July 11, 2024. The Sci-Tech Leaders ETF passively tracks the CSI Sci-Tech Leaders Index. The base date for this index is June 29, 2012, and its release date is March 20, 2019. The composition of index constituents is adjusted according to the index methodology, and its back-tested historical performance does not indicate future index performance. Constituent stocks mentioned herein are for illustrative purposes only. Descriptions of individual stocks do not constitute investment advice in any form, nor do they represent the holdings or trading动向 of any fund managed by the manager. According to the fund manager's assessment, the ChiNext AI ETF carries a risk rating of R4 (Medium-High Risk) and is suitable for investors with an Aggressive (C4) or higher risk profile. Suitability matching opinions are subject to the sales institution. Any information appearing in this article is for reference only. Investors are solely responsible for their independent investment decisions. Furthermore, any views, analysis, or forecasts herein do not constitute investment advice of any kind to the reader, and no liability is accepted for any direct or indirect losses arising from the use of this content. Fund investment involves risks. The past performance of a fund is not indicative of its future results. The performance of other funds managed by the fund manager does not guarantee the performance of this fund. Invest with caution in funds.