“Yi Zhongtian” still has to endure

The pressure for this round of adjustment mainly comes from two aspects:Changes in technology routes brought about by CPO and slowdown in AI capital expenditure growth。

Let’s look at CPO first.

On October 21, NVIDIA announced the expansion of its supply chain production capacity. Its partner Silicon Precision Industries plans to increase its CPO light engine production capacity to 15 times the current level by mid-2027.

This allows the market to see a possibility: CPO technology, which was still being gradually introduced, may achieve large-scale application faster than expected.

What does this mean for optical module leaders such as Zhongji InnoLight and Xinyi Sheng?

In the past few years, these companies have mainly benefited from the explosive demand for 800G and 1.6T pluggable optical modules, and their revenue and profits have continued to grow rapidly. CPO may change the product form of optical modules and even redistribute industry chain profits.

The biggest difference between CPO and traditional pluggable optical modules is that the optical engine is moved from the switch panel to the vicinity of the switching chip. This shortens the electrical signal transmission distance, reduces power consumption and loss, and increases bandwidth density.

Especially in the context of the continuous expansion of AI computing power clusters, the power consumption of optical interconnections is becoming an important technical constraint. The advantages of CPO in this area are an important reason why companies such as NVIDIA and Broadcom promote their commercialization.

But the fact that CPO is starting to increase its volume does not mean that traditional pluggable optical modules will withdraw from the market.

On the one hand, although CPO has advantages in power consumption and bandwidth density, more complex packaging, fiber coupling, testing and maintenance also bring new cost and reliability challenges. Even if related products have entered the production stage, the cost and economics of large-scale deployment still need to be further verified.

On the other hand, traditional pluggable technology has not stopped upgrading.

On October 12, Nvidia and Cisco announced joint promotionNew high-density pluggable optical module form USFP, supports 400G/lane, and the future goal is to increase the bandwidth of a single module to 12.8Tb/s(The current single-module bandwidth of mainstream 800G and 1.6T products is 0.8Tb/s and 1.6Tb/s respectively)。

At the same time, NPO(near package optics)It is also becoming another important route. It places the optical engine close to the switching chip, but does not co-package it with the chip. The power consumption reduction is between pluggable and CPO, and maintenance is more flexible.

This shows that the industry is not betting on a single technology route.In the next few years, CPO, NPO and pluggable optical modules are more likely to develop in parallel in different application scenarios, rather than one technology rapidly replacing other technologies.

• 800G and 1.6T are speeds, which answer how much data can be transmitted per second and solve the problem of how fast it can run;

• Pluggable, CPO, and NPO are packaging architectures, which refer to how the optical engine and switching chips are placed and how they are wired. All three routes will be upgraded in speed in the future.

Therefore, we believe that CPO has indeed increased the uncertainty of the future competitive landscape of optical modules, but the accelerated development of CPO alone is not enough to judge that the growth logic of existing leaders has fundamentally changed.

Compared with technology substitution, another variable that deserves more vigilance is the change in the growth rate of AI capital expenditure.Because CPO affects the competitive landscape, capital expenditure directly determines demand itself.

Optical modules are a growth cycle business. Cyclicity comes from cloud vendors’ capital expenditures, data center construction and customer inventory changes; growth comes from the continuous upgrades from 100G, 200G, 400G to 800G and 1.6T. The former determines whether demand fluctuates, while the latter continuously creates new product demand and value.

In the past two years, the construction of AI computing power has promoted both growth.The expansion of GPU clusters has increased the demand for data transmission between servers and cabinets; at the same time, the optical module speed has been continuously upgraded from 400G to 800G and 1.6T, which has also brought about improvements in product structure and value.

However, recent predictions of a slowdown in the growth of AI investment by cloud vendors have triggered market concerns about the sustainability of growth.

Morgan Stanley calculated in relevant reports released from the end of September to mid-October(Including leasing investment), the related investment amount of the world’s major cloud service providers and emerging computing power service providers will be approximately US$980 billion in 2026, approximately US$1.58 trillion in 2027, and approximately US$1.81 trillion in 2028;The year-on-year growth rate will drop from approximately 100% to 60%, and then to approximately 14%.

This is a slowdown in the slope and does not mean demand is peaking.

But for the optical module industry, the growth rate of cloud factory investment is the most noteworthy variable. It determines the expansion pace and networking density of the GPU cluster, which in turn determines the demand elasticity of optical modules.

This is why the high growth in the third quarter report did not directly lead to a rebound in stock prices.

The third quarterly report reflects past operating results and cannot change the market’s concerns about the growth slope after two years. andThe market is trading growth expectations for 2027, 2028 and beyond.

This is the pricing characteristic of growth cyclical stocks. When the industry booms, performance growth and valuation expansion may occur at the same time;When the market begins to worry about the next round of capital expenditure cycles, or changes in the competitive landscape brought about by technological iterations, valuations may shrink in advance even if current performance is still growing.

Therefore, this round of adjustment does not mean that industry demand has peaked, but that the market has begun to reassess the sustainability of future growth. CPO affects the long-term competitive landscape, while the decline in capital expenditure affects the recent demand slope. The two pressures are superimposed, and the premium the market is willing to pay for the long-term growth of optical modules will naturally narrow.

The key is, has this round of adjustments fully digested these concerns? When will it rebound?

For the optical module to rebound, three signals are missing.

After this round of adjustments, a lot of bad news has indeed been released.

As of October 30, 2026, Zhongji InnoLight PE-TTM(rolling P/E ratio) 25.15 times, with a ranking of 18.99% in the past five years; Xinyi Sheng PE-TTM is 18.75 times, with a ranking of 14.78% in the past five years; Tianfu Communications PE-TTM is 61.26 times, with a ranking of 60.28% in the past five years.

From the perspective of forward valuation, the expected PE of Zhongji InnoLight and Xinyi Sheng in 2027 and 2028 have both fallen below 10 times, reaching the valuation range of cyclical stocks; Tianfu Communications has fallen back to 27 times in 2027 and further dropped to 18 times in 2028. The safety margin at the valuation level is accumulating.(The expected PE for forward valuation is calculated by Wind’s consensus expectation)

Therefore, from the perspective of valuation, some leading companies have entered relatively attractive ranges, but the bottom still needs to stabilize profit expectations.

Also, cheap does not mean safety pad. If the price of the current mainstream 800G/1.6T products drops more than expected, CPO replacement is faster than expected, or if cloud vendors slow down procurement, lengthen the procurement cycle, or even cut orders for 2027 and 2028 because the return on AI investment is lower than expected, any of these may turn the order guidance into a cut, triggering a downward revision of profit forecasts, and the current low valuation will continue to be revised downwards.

At the same time, when it rebounds, we can’t just look at valuation. Next, there are three signals that are really worth tracking.

First, whether orders and demand guidance can remain stable.

The slowdown in AI investment growth does not mean that demand for optical modules will inevitably decline.

The cloud factory’s investment not only includes optical modules, but also includes expenses such as building computer rooms, power supply, and refrigeration. Therefore, as long as the total investment by cloud vendors continues to increase, GPU clusters continue to expand, and 800G and 1.6T products continue to be upgraded, there will still be new demand for optical modules. Then, even if the total plate growth slows down, the proportion of optical modules in total capital expenditures may increase, and orders will have the micro-foundation to maintain strong.

On the other hand, if customers begin to reduce purchase plans, or order guidance is significantly lowered, it means that the demand side has substantially weakened, and the market may further revise profit expectations accordingly.

Therefore, whether optical module manufacturers’ order guidance for 2027 and 2028 is stable, and whether 800G and 1.6T shipments can continue to grow, is the first signal worth tracking.

Second, next-generation products such as CPO and NPO must move from technical layout to batch delivery.

For companies such as Zhongji InnoLight and Xinyi Sheng, next-generation products such as CPO are both a threat and may also bring new growth opportunities.

The key is whether it can continue to obtain customer orders and maintain customer share and profitability in the next-generation optical interconnection solution.

In particular, it is necessary to distinguish between samples, customer verification, order instructions and bulk delivery. Only when the latter two stages are truly realized will they have a more direct contribution to the company’s performance.

Third, institutional profit forecasts have begun to be revised upward.

This is the key to whether the stock price can move from the bottom to a sustained rebound.

For growth cyclical stocks, stock prices often reflect profit expectations for the next one or two years in advance. If orders, product prices or gross profit margins fall short of expectations and institutions lower their profit forecasts for 2027, the current seemingly low forward price-to-earnings ratio may also rise passively.

On the other hand, if subsequent orders remain stable, product volume and gross profit margins do not deteriorate significantly, and institutions do not lower profit forecasts or even start to revise them upward, then the market will have reason to reassess the company’s reasonable valuation.

Among the three signals, orders determine demand, product delivery determines whether the company can obtain new revenue, and profit forecasts determine whether the valuation can be repaired.What the sector really needs is no longer a grand industrial story, but solid evidence to support future performance growth.

“Yi Zhongtian”, who will miss the opportunity first?

Based on the above, the optical module industry does not need to be overly pessimistic. There may still be room for growth in the future. However, it does not mean that all companies can maintain the growth rate of the past, nor does it mean that all companies should enjoy the same valuation.

During the bottoming period, flexibility cannot be realized, so when selecting stocks, the weight of certainty is higher than flexibility.

Under the above circumstances, among the three “Yizhongtian” companies, if ranked according to the current performance certainty, valuation level and new product progress, I would prefer Zhongji Innolight to be observed first, followed by Xinyi Sheng, and then Tianfu Communications. Next, the reasons are given one by one in this order.

(1) Zhongji InnoLight: Performance certainty is relatively stronger.

In the first three quarters of 2026, Zhongji InnoLight achieved revenue of 69.547 billion yuan, a year-on-year increase of 178.13%; net profit attributable to the parent company was 22.238 billion yuan, a year-on-year increase of 211.81%. 800G/1.6T pluggable products continue to increase in volume, and 1.6T enjoys a pricing premium for new products due to tight supply and demand, which is an important reason for good revenue and profit growth.

That is to say, Zhongji InnoLight’s core advantage is still its customer base and large-scale delivery capabilities of high-speed pluggable optical modules.

In the next stage, we need to observe whether 1.6T products can continue to increase in volume, and whether next-generation products such as NPO and CPO can enter customer supply chains as planned. Management stated in an exchange on October 29 that the next generation of products is expected to be gradually released in the second half of 2027, and is expected to increase significantly in 2028.

In addition, Zhongji InnoLight’s PE-TTM was 25.15 times on October 30, which is at the historical level of about 19% in the past five years; if the 2027 profit forecast can be realized, the PE is expected to be reduced to 9.16 times, and the valuation is already somewhat attractive.

Zhongji InnoLight’s advantage is not that it is betting on a certain new technology, but that its existing business still has a strong foundation for growth and it has the ability to extend to the next generation of products. Therefore, if the sector experiences a fundamentals-driven rebound, it is more worthy of priority tracking.

(2) Xinyi Sheng: The forward valuation is lower, and the rebound elasticity deserves attention.

In the first three quarters of 2026, Xinyi Sheng achieved revenue of 31.850 billion yuan, a year-on-year increase of 92.97%; net profit attributable to the parent company was 11.372 billion yuan, a year-on-year increase of 79.74%, and its performance was also good.

However, although it is the same as Zhongji InnoLight, it also focuses on high-speed pluggable optical modules as its core business. However, its 1.6T pluggable products have a slower pace of ramping up volume. They started in Q2 of 2026 and only started to gradually ramp up in the third quarter, which is about a quarter slower than Zhongji InnoLight.

In the next stage, we need to observe whether the volume of 1.6T pluggable products can be accelerated, and whether NPO orders can be transferred from guidance to batch delivery. Management said on October 29 that the NPO is progressing smoothly and has seen clear order guidance for 2027.

Calculated based on the current market value and institutional profit forecast, Xinyi Sheng’s PE-TTM on October 30 was 18.75 times, and the historical valuation metric is about 15%; in 2027, the institution expects the net profit attributable to the parent company to increase by about 84%, and the expected price-to-earnings ratio is about 7.44 times, which is lower than Zhongji InnoLight.If the 1.6T volume accelerates and NPO orders are fulfilled in the future, it may have greater flexibility to repair its valuation.

However, low valuation does not mean that there are no risks. The pace of volume expansion, price and gross profit margin of 1.6T pluggable and NPO products are still the key to determining whether the profit forecast can be fulfilled.

(3) Tianfu Communications: The CPO theme is relatively flexible, but the margin of safety in valuation is relatively insufficient.

In the first three quarters of 2026, Tianfu Communications achieved revenue of 4.751 billion yuan, a year-on-year increase of 21.25%; net profit attributable to the parent company was 2.083 billion yuan, a year-on-year increase of 42.18%. The revenue and profit growth rates are the lowest among the three companies of “Yizhongtian”, mainly due to different business models.

Different from the first two companies, Tianfu Communications mainly provides optical components and does not produce complete optical modules.

As CPOs and NPOs increase the integration of optical engines, devices such as FAU optical fiber arrays and ELS external light sources are expected to gain new demand.The company’s management stated that CPO supporting products will gradually increase volume starting from 2027. If NPO supporting products are successfully verified and introduced by customers, they are also expected to achieve large-scale mass production in 2027.

But from a valuation perspective, Tianfu Communications’ expected price-to-earnings ratio in 2027 is about 27.28 times, the current PE-TTM is 61.26 times, and the historical valuation metric is about 60%. Even after experiencing a significant correction, its valuation is still not as fully digested as the first two.

If the market re-trades CPO industrialization, Tianfu Communications may gain better theme flexibility; but if it wants to form a sustained rebound supported by performance, it needs to see CPO and NPO device orders truly converted into revenue and profits. Current valuations still require a high level of future growth.

Taken together, the three companies each have their own highlights and the conditions for rebound are not the same. Zhongji InnoLight’s existing business growth is relatively more certain and is suitable for priority observation; Xinyi Sheng’s forward valuation is lower, and if the product volume increases, the flexibility of valuation recovery deserves attention; Tianfu Communications CPO and NPO have great potential, but more performance is needed to digest the valuation.

因此，Although the market has digested a considerable part of the concerns after this round of adjustments, the market’s concerns about future growth have not completely disappeared. In the short term, it will be difficult for the sector to return to its previous upward trend.。

Next, whether orders can maintain growth, whether next-generation products such as CPO can be delivered in batches, and whether institutional profit forecasts can be stabilized will determine whether the sector can rebound from an L-shaped bottom.。

Related Reading

• Nvidia live updates: Jensen Huang ousted, CFO Colette Kress to serve as interim CEO2026-11-13
• Are mangoes not afraid of red fruits?2026-11-12
• Original: After Huang Renxun, who is most likely to take over Nvidia?2026-11-12
• Samsung’s roll-screen mobile phone is exposed, creating a new form ahead of Huawei and Apple, and folding screens are expected to be eliminated2026-11-12
• Musk’s favorite car is going to be discontinued.2026-11-12