The differentiation and competition of optical modules under the rigid demand for computing power

Ask AI · Why does the differentiation of optical modules highlight a company’s core competitiveness?

After four consecutive declines in the optical module sector on April 29, a recovery ensued. Senior meetings increased investment in 6G construction, coupled with orders from Google NPO and TSMC’s silicon photonics mass production, boosting expectations for CPO/silicon photonics; Yuanjie Technology’s first-quarter net profit surged by 1153%, Cambridge Technology hit the daily limit, Zhongji Xuchuang rose over 2%, with performance serving as a core support.

On that day, the CPO index closed at 14,035.81 points, up 0.22%, with a trading volume of 124.23B yuan leading the market. The main net inflow of funds into the sector was 1.08B yuan; constituent stocks showed mixed gains and losses, with Zhongji Xuchuang’s main net inflow of 1.648 billion yuan ranking first, Dongshan Precision’s net outflow leading, while overseas LUMENTUM and Broadcom rose simultaneously.

Performance** Realization Period**

Wind data shows that by the end of the first quarter of 2026, among actively managed equity-heavy public funds, Zhongji Xuchuang ranks first, and Xinyisheng ranks third in their top holdings. Yuanjie Technology, Tanfeng Communications, and Hengtong Optoelectronics have entered the top 20 holdings, highlighting institutional recognition of core targets in the optical module and CPO industry chain.

However, on April 24, the A-share optical module sector experienced a divergence, reflecting the ongoing tug-of-war between AI computing power sentiment and fundamentals. That day, Wind’s optical module index plummeted 1.91%. Xinyisheng’s stock fell 11.67% in a single day due to a 13.25% decline in quarterly net profit compared to the previous quarter, falling short of market expectations, with a trading volume of 1.65B yuan, setting a record for daily turnover among A-share tech stocks; Tanfeng Communications also dropped nearly 7.5%, while industry leader Zhongji Xuchuang resisted the decline, only falling 1.12%. The next day, the sector continued to fluctuate at high levels, with funds accelerating their concentration into leading companies with more certain performance.

This sharp divergence broke the market’s previous perception of a “general rally” in the optical module track, marking the industry’s transition from a “prosperity dividend period” to a “performance realization period.” The core barriers and potential risks of the hard technology track were thoroughly exposed under the “demon-slaying mirror” of the first quarter report.

Dissecting the fundamentals of the “Three Major Leaders” of Yi Zhongtian, the essence of performance divergence lies in differences in core competitiveness.

Zhongji Xuchuang, by deeply partnering with North American cloud giants like Nvidia and Google, maintains a leading global position. Its 800G optical modules hold over 40% global market share, and its 1.6T modules account for 50%-70% market share.

In Q1 2026, Zhongji Xuchuang achieved revenue of 50.1B yuan, up 192.12% year-over-year, with a net profit of 5.735 billion yuan, up 262.28% year-over-year. Gross margin further increased to 46.06%. Orders from key clients are projected to extend to 2028, highlighting long-term certainty.

Xinyisheng, relying on urgent overseas AI orders for short-term rise, reported revenue of 19.5B yuan in Q1, up 105.76% year-over-year. However, with overseas revenue accounting for as high as 96.16% in 2025, it faces exchange rate risks, with RMB appreciation causing significant exchange losses. Coupled with shortages of high-end materials leading to delivery constraints for high-margin products, profits ultimately declined quarter-over-quarter.

Tanfeng Communications focuses on upstream optical devices, with high gross margins, but its customer concentration is high. Due to shortages of core materials, its performance growth is notably weaker than peers.

**Three Core Barriers

From the product perspective, 800G optical modules remain the main demand in the industry. 2026 is expected to be the year of large-scale deployment of 1.6T modules. The three leading companies have completed full product line layouts and are simultaneously advancing R&D of cutting-edge technologies such as CPO, silicon photonics, and thin-film lithium niobate, with product structures continuously moving toward high-end.

At the industry level, “copper replacing fiber” is a long-term core logic. As AI computing clusters expand to hundreds of thousands of cards, the ratio of GPU to optical modules increases to 1:3-1:5, making high-speed optical modules a necessity for computing infrastructure.

The global optical module market is projected to reach $26 billion by 2026, with an annual growth rate exceeding 60%. Domestic companies, leveraging cost and delivery advantages, hold about 70% of the midstream market share. Domestic substitution continues to deepen, while technological iteration, customer binding, and global capacity deployment form the three core barriers enabling companies to cycle through market fluctuations.

Beneath the high prosperity of the industry, multiple structural risks lurk, becoming real bottlenecks hindering long-term growth.

Major domestic optical module companies’ revenues are highly tied to overseas markets, with over 80% of income concentrated among four North American cloud giants. Geopolitical changes and adjustments in major clients’ capital expenditure rhythms directly impact orders and revenue.

Meanwhile, key segments of the industry chain still depend heavily on foreign suppliers. High-end EML chips, DSP chips, and indium phosphide substrates are predominantly supplied by overseas firms. Capacity expansion cycles for these materials can take up to 18 months, and persistent shortages can cause mismatched capacity, representing industry bottlenecks.

Profitability pressures are also mounting. The 800G market has long been price-competitive, with overall gross margins falling from over 40% to between 25%-30%. When 1.6T capacity is fully released by 2026, industry prices are expected to further decline.

Additionally, fluctuations in RMB exchange rates lead to significant exchange gains and losses. Companies stockpile upstream resources, with inventories and prepayments rising, further increasing financial burdens. Profitability and cash flow stability face considerable challenges.

Comprehensive** Competition**

Technological iteration and industry competition also harbor uncertainties.

Currently, CPO accounts for only 3% of the replacement ratio for traditional pluggable optical modules. While the short-term impact is limited, long-term potential could reshape the entire industry structure. Changes in technological routes also introduce transformation uncertainties for existing manufacturers.

Meanwhile, domestic peers are expanding capacity collectively, and overseas giants are increasing their investments. Industry overcapacity risks are emerging. Moving production to Southeast Asia and other regions has become a necessary step for leading firms to secure North American clients. Competition has shifted from product comparison to a comprehensive contest of global capacity and supply chains.

In the long run, the 1.6T volume, CPO commercialization, and global capacity battles will reshape the industry landscape.

In the short term, the supply-demand gap for 1.6T is significant. Leading companies, with their supply chain and client barriers, will benefit first, but overcapacity of 800G may lead to profit declines.

Long-term, from 2027 to 2028, CPO may enter large-scale commercial use, and photonic integration will become a main industry direction. Guojin Securities states that AI development drives the expansion of training clusters and increases inference-side computing power, greatly boosting demand for computing and networking. CPO can effectively reduce cluster power consumption, enhance interconnect density, and ensure high-speed transmission stability. Industry demand is urgent, and supply is gradually maturing, promising accelerated penetration. Investors should focus on core optical components, manufacturing, and solution providers in the CPO supply chain.

Therefore, for investors, it is crucial to abandon the mindset of sector-wide rallies, strictly distinguish between cyclical dividends and core enterprise value, and focus on leading companies with 1.6T mass production capacity, forward-looking CPO deployment, autonomous supply chains, and diversified customer structures.

It can be said that the wave of computing power is far from over. The fundamental logic of optical modules as “selling shovels” remains valid, but industry reshuffling has begun. Only companies that truly master technological lifelines, bind to core ecosystems, and build global competitive moats can navigate volatility in the differentiation phase and continue to stand at the forefront of the computing revolution.

Editor | Wang Peng