The market's current focus is often divided into two camps: those "standing in the light" and those "where the light stands." This "light" refers to optical modules—a segment within AI computing that boasts the most solid performance and concentrated capital flows. However, a widely overlooked fact is that there is no pure "optical module ETF" available in the market. To capture the beta returns from optical modules, investors must select and allocate funds among related ETFs with higher exposure to this segment.
In this context, one investment vehicle stands out as unavoidable—the ChiNext Artificial Intelligence ETF (159363), which has surged 200% over the past year. Some may question: What about the Communications ETF (515880)? Doesn’t it also include optical modules? Let’s examine this in detail. Below, we break down the data across three dimensions—historical performance, weighting analysis, and portfolio structure—to explain why the ChiNext Artificial Intelligence ETF (159363) is the true "optical module ETF."
A summary comparison is provided at the end for reference.
**1. Historical Performance Review: Excess Returns of the ChiNext AI Index Since the AI Model Boom**
In November 2022, the launch of ChatGPT sparked global interest in generative AI. By 2023, tech giants worldwide—from OpenAI, Google, and Meta to Baidu, Alibaba, and Tencent—had entered the large-model race, initiating an intense competition around computing power. Optical modules, as a core component of computing infrastructure, emerged as the segment with the highest degree of realized performance within the AI industry chain. Why? Because training and inference of large models require massive computational support, and high-speed optical modules are indispensable for interconnecting that computing power.
Optical modules serve as the "data channels" for computing interconnectivity—converting electrical signals into optical signals and transmitting them at high speeds via optical fibers, effectively linking thousands of GPUs into a supercomputer. Without them, computing resources would remain isolated. Since 2023, orders for high-speed optical modules have been consistently full, with related companies reporting doubled earnings.
During this historic market trend, the returns of optical module-related assets have become a key measure of "AI content." The chart below compares the performance of the ChiNext Artificial Intelligence Index (970070.CNI) and the Communication Equipment Index (931160.CSI) since the beginning of 2023:
Visually, the two indices show a degree of correlation, especially during the broad computing rally phase in 2025, where both benefited from the optical module trend. However, in cumulative terms, the ChiNext AI Index gained 440% from 2023 onward, while the Communication Equipment Index rose about 377% over the same period—meaning the AI index outperformed by more than 63 percentage points.
**2. Component Weighting Analysis: Comparing Optical Module Exposure via the "Yi Zhong Tian" Concentration**
In the optical module sector, three stocks are collectively referred to as "Yi Zhong Tian"—Xinyisheng, Zhongji Innolight, and Tianfu Communication. Why are they so important? Because these three companies account for nearly half of the AI computing optical module market:
- Xinyisheng: A leader in 800G optical modules, with rapid volume growth from overseas clients and high earnings elasticity. - Zhongji Innolight: Holds the top global market share in optical modules, leads in 1.6T development, and offers strong certainty. - Tianfu Communication: A core supplier of optical components, deeply integrated into overseas computing supply chains.
These companies share common traits: deep ties to overseas AI computing chains, rapid volume expansion in 800G and 1.6T high-speed optical modules, and top-tier earnings realization. In a sense, the weighting of "Yi Zhong Tian" directly determines an ETF’s purity in optical module exposure.
So, which ETF has a higher "Yi Zhong Tian" content—the ChiNext AI ETF or the Communications ETF? As shown in the chart below, the combined weighting of "Yi Zhong Tian" in the ChiNext AI ETF (159363) is 41.31% (the highest in the market), compared to 37.37% in the Communications ETF (515880)—a nearly 4-percentage-point advantage for the AI ETF.
For every 1 million yuan invested, the ChiNext AI ETF allocates nearly 40,000 yuan more directly to the three optical module giants than the Communications ETF. This 4% difference should not be underestimated; if optical module stocks rise 100%, it translates to additional returns. Higher optical module concentration means purer AI exposure—a key reason for the ChiNext AI ETF’s sustained outperformance.
**3. Portfolio Structure and Driving Factors: Single-Factor vs. Multi-Factor Model Comparison**
We have established that the ChiNext AI ETF (159363) holds a clear advantage in "Yi Zhong Tian" content. But a follow-up question arises: If optical modules rise, will the ETF definitely follow? Could other assets drag it down? This leads to the second criterion for a "true optical module ETF"—not only must it have "sufficient optical modules," but it must also have "few impurities." Let’s further analyze the two ETFs’ portfolio structures.
First, the ChiNext AI ETF (159363) is 100% focused on the AI industry chain, spanning upstream infrastructure and downstream applications. As of the end of March, optical modules and CPO (co-packaged optics) accounted for about 45% of the portfolio, serving as the primary drivers. The remaining allocation covers domestic computing (e.g., IDC data centers, computing chips, computing leasing) and AI applications:
- Approximately 45% in optical modules/CPO: Mainly "Yi Zhong Tian" and other overseas computing leaders, capturing the most certain dividends of AI infrastructure. - About 25% in domestic computing: Benefiting from computing shortages and asset revaluation in China. - Roughly 30% in AI applications: Including companies like Inspur Software, Sangfor Technologies, and Kunlun Tech, poised to gain from model deployment and AI agent growth.
In contrast, the Communications ETF focuses heavily on communication equipment as a single asset class. Beyond optical modules, it holds significant traditional communication assets—base stations, routers, optical fibers, and cables—which have a more indirect relationship with AI.
The ChiNext AI ETF (159363) is driven by three core factors, all high-elasticity assets within the AI chain that are likely to move in synergy with optical modules. The Communications ETF’s returns, however, are highly concentrated in a single factor—communication assets—where non-AI holdings may act as a drag during AI rallies.
In summary, data across these three dimensions confirm that the ChiNext Artificial Intelligence ETF (159363) is the true "optical module ETF": - Historically, it has outperformed the Communications ETF by over 63 percentage points since 2023. - In "Yi Zhong Tian" content, it leads with 41.31% versus 37.37% for the Communications ETF. - In portfolio structure, it allocates 100% to the AI industry chain, with optical modules exceeding 45% as the dominant driver.
Risk Disclosure: All information presented is for reference only. Investors are responsible for their own investment decisions. The views, analyses, and forecasts herein do not constitute investment advice, and no liability is accepted for any direct or indirect losses resulting from the use of this content. Market risks exist; invest with caution.
MACD golden cross signals have formed, and certain stocks are performing well.
For extensive information and precise analysis, visit the Sina Finance app.
Comments