Xizhi Technology Research Report: The hype is only second to Alibaba and Kuaishou, how to view the market today?
At first, I just wanted to see if this IPO was worth participating in, but the more I looked into it, the more I realized it wasn’t just about the IPO issue—it led to discussions about the new track of Hong Kong stocks 18C, the industrial pattern of domestic AI computing power, the valuation arbitrage logic between A-shares and Hong Kong stocks, and how retail investors can use information advantages to play in the Hong Kong stock market with larger propositions.
1. What is this company doing
Xizhi Technology is doing two things: optical interconnection and optical computing. Additionally, there’s a concept of optical modules in the market; these three things are often mixed together, but they serve completely different positions in the industry chain.
Optical Interconnection: Replacing copper cables with optical fibers
When an AI training cluster expands from a thousand cards to tens of thousands or even hundreds of thousands of cards, copper cables hit a physical wall—signal attenuation, rising power consumption, and uncontrollable heat dissipation. This isn’t an engineering problem, but a hard constraint of electromagnetism. Xizhi’s solution is to connect GPUs with optics, reducing transport latency by over 90% and improving the compute power utilization (MFU) of GPUs in the cluster by 50%.
This part accounts for 71% of the company's revenue in 2025 and is currently the cash flow business. The specific products are OCS (Optical Circuit Switching) + oNET (Inter-chip Optical Network).
Optical Computing: Directly using light for matrix multiplication
This is a completely different concept. It does not transmit data; it performs calculations using light itself.
How is this achieved? Light is naturally suitable for linear operations—its propagation in a medium is a mathematical transformation. By letting two beams of light pass through a modulator at specific intensities, the total intensity after passing through directly represents the result of matrix multiplication. This is not a metaphor; it is literally "light doing computation."
What does this mean? Delays can reach the nanosecond level (GPUs are in the tens to hundreds of microseconds), and energy consumption is an order of magnitude lower.
The optical computing business is expected to account for 19% of Xizhi’s revenue in 2025, though its absolute value is quite small (20.1 million RMB), but it shows a year-on-year growth of 579%.
This is the "second curve" it presents to the secondary market.
Optical Modules: Translators between light and electricity
Optical module = photoelectric signal converter. It converts electricity into light and light back into electricity, purely a "translator."
However, the optical computing that Xizhi Technology is doing uses light itself for mathematical operations—this is a completely different product category. Optical modules are the "veins" of data centers, while optical computing represents the "brain cells" of data centers. These are two completely different markets, with different customers and players.
The AI optical communication sector in the U.S. stock market, Lumentum (LITE), has increased nearly 20 times in the past year—but its story involves lasers + OCS, not traditional optical modules.
The global leader in optical modules is China’s A-share company Zhongji Xuchuang, expecting revenues of over 20 billion RMB in 2024 to supply Google and Meta, with a PE ratio of 30-40 times. However, Zhongji Xuchuang and Xizhi are not direct competitors but players in completely different positions in the industry chain. Overseas, there’s Coherent, which has seen a sevenfold increase in one year.
Currently, aside from Xizhi, there are no listed pure plays in the optical computing track—this is the source of Xizhi's rarity as the "first global AI silicon photonics stock" listed on the Hong Kong Stock Exchange. Overseas players Lightmatter and Ayar Labs are still in the private placement stage.
For the full text, move to the official account for better formatting
https://mp.weixin.qq.com/s/TPILjvGR65qDJvX9LyWXDg
免责声明:本文章仅代表作者个人观点,不代表本平台的立场和观点。本文章仅供信息分享,不构成对任何人的任何投资建议。用户与作者之间的任何争议,与本平台无关。如网页中刊载的文章或图片涉及侵权,请提供相关的权利证明和身份证明发送邮件到support@aicoin.com,本平台相关工作人员将会进行核查。