Lead-in: AI computing power demand is still growing, but optical module companies can no longer rely solely on order expectations to support their stock prices.
On November 11, the optical module sector saw a significant rebound.
The Wind Optical Module (CPO) Concept Index rose by 8.25%, with leading stocks such as Zhongji InnoLight, Eoptolink, and Tianfu Communication gaining strength collectively. Prior to these trading days, these companies had just undergone a notable correction.
The market repeatedly traded on the same question: How many optical modules will overseas cloud vendors continue to purchase?
Rumors regarding delays in North American major cloud service providers' data center projects and cuts in optical module orders have circulated recently. Although relevant companies have responded one after another, investor concerns about orders have not completely disappeared.
On November 3, Zhongji InnoLight stated in an investor exchange that the vast majority of customers had not lowered demand or orders, some customers were adjusting product structures, and demand for 1.6T products from certain customers had actually increased.
Cambridge Industries also stated at its earnings briefing on November 9 that the company's high-speed optical module order demand was robust, capacity utilization remained at a high level, and market rumors did not align with the company's actual operating situation.
On November 12, discussions about order cuts in the optical communication industry resurfaced, with multiple listed companies responding that there had been no adverse changes in recent operations.
The reason rumors keep recurring is not merely due to fragile market sentiment.
Over the past year, the stock prices and valuations of optical module companies have largely been built on the expectation of sustained growth in overseas AI capital expenditure. When investors begin to doubt whether these investments can be converted into orders as planned, every change in the supply chain gets amplified.
Orders No Longer Only Have One Possibility: Growth
In the past two years, optical modules have been one of the few directions in the AI computing power supply chain capable of consistently delivering performance.
NVIDIA continues to launch new generations of GPUs, while cloud service providers like Microsoft, Google, and Amazon expand their data centers, leading to increased data transmission needs between servers. As a critical component for transmitting data within data centers and between different devices, optical modules directly benefit from network bandwidth upgrades.
From 400G to 800G, and then to 1.6T, product iterations continuously push up the optical communication demand per single data center.
This growth logic has not completely changed yet.
Amazon's Q3 AWS revenue reached USD 46 billion, a year-on-year increase of 39.3%. Alphabet's Q3 Google Cloud revenue reached USD 30.7 billion, a year-on-year increase of 103%. Microsoft's capital expenditure in the first quarter of fiscal year 2027 reached USD 45.2 billion, with expectations to exceed USD 50 billion in the next quarter.
The scale of investment by these companies remains massive, and their cloud businesses are growing. At least based on currently disclosed data, overseas tech giants have not collectively contracted their AI infrastructure construction.
However, the focus of market attention has shifted.
Data center construction is not an investment that can be completed in one go. Land, power, servers, switches, optical modules, and network architectures need to be deployed step-by-step according to project progress. A delay in a specific project may affect the delivery time of a batch of equipment; changes in network architecture may also alter the procurement ratio of different product models.
This means that continued growth in cloud vendors' capital expenditure does not imply that orders for every supplier will grow at the original pace.
Zhongji InnoLight's response precisely reflects this change.
The company stated that while most customers have not lowered demand or orders, some indeed have product structure adjustments. Meanwhile, demand for 1.6T products from certain customers has increased.
This is quite different from what the market refers to as comprehensive order cuts.
If customers reduce 800G procurement while increasing 1.6T procurement, orders for a specific category may decline; however, looking at the entire optical communication system, total demand may not decrease.
For optical module companies, what truly matters is total order volume, product mix, and supplier share, rather than procurement changes for a single model.
In its Q3 earnings call, Zhongji InnoLight stated that orders for 800G and 1.6T products both achieved sequential growth compared to Q2, with revenue from 1.6T products accelerating further. On November 6, the company also noted that orders from major customers have basically covered the full year of 2027, and some key customers have already provided demand guidance for 2028.
Eoptolink also stated in its production and operation announcement on October 23 that 800G and 1.6T are the main products delivered this year, and 1.6T optical modules have entered a phase of continuous mass production ramp-up.
This information indicates that leading companies currently still enjoy strong order support.
However, having orders cover the full year does not mean that revenue and profits for the next year are guaranteed. Customer demand, product certification, delivery progress, and procurement prices will still impact final operating results.
1.6T Has Become the New Competitive Focus
The optical module industry is undergoing a shift in product structure.
800G was once the main product for high-speed interconnection in AI data centers. As GPU cluster sizes expand, the amount of data exchanged between servers increases, putting pressure on traditional network architectures. Consequently, 1.6T optical modules have begun entering large-scale adoption.
For optical module manufacturers, this is not just about doubling product speed; it involves re-adjustments in chips, optical components, packaging processes, and production yield rates.
Different technology routes are being advanced simultaneously.
The traditional EML solution uses indium phosphide lasers for optoelectronic conversion, while silicon photonics solutions integrate some optical functions onto silicon-based chips. The two routes have distinct characteristics in terms of cost, power consumption, performance, and manufacturing processes.
As 1.6T products ramp up, the application of silicon photonics technology is also increasing.
This means competition among optical module companies is no longer just about who has more orders, but also about who can complete new product validation faster, who can stably provide high-yield products, and who can enter the next-generation supply systems of core customers.
Although Zhongji InnoLight, Eoptolink, and Tianfu Communication are all positioned importantly in the optical communication supply chain, their business structures differ.
Zhongji InnoLight primarily engages in the R&D, production, and sales of high-speed optical modules and holds an important position in the supply chains of large overseas cloud service providers.
Eoptolink similarly benefits from demand for 800G and 1.6T products, with its business growth closely tied to overseas data center construction.
Tianfu Communication is more involved in the optical components and optical engine segments, providing core components for optical modules.
Opportunities brought by product upgrades will not be evenly distributed among all enterprises in the supply chain.
Whoever completes technical validation first and secures more orders from core customers has the chance to gain higher market share. Conversely, if new product introduction falls short of expectations, companies may face a growth gap even if overall industry demand grows.
Market concerns about the optical module industry have extended from simple demand changes to technology routes and competitive landscapes.
The Other Side of Order Growth
A change easier to observe in financial reports than cut-order rumors is that optical module companies are increasing inventory and prepayments.
As of the end of Q3 2026, Zhongji InnoLight's prepayments reached CNY 1.567 billion, a 1068.43% increase from the end of 2025; inventory reached CNY 23.878 billion, an 88.30% year-on-year increase.
Eoptolink's changes are even more pronounced.
As of the end of Q3, the company's prepayments reached CNY 1.322 billion, a 7695.01% increase from the beginning of the year; inventory reached CNY 17.413 billion, a 140.70% increase from the beginning of the year.
Both companies linked these changes to business growth, order stocking, and upstream material procurement.
Amid growing demand for high-speed optical modules and tight supply of core materials, paying suppliers in advance and increasing raw material reserves is a way for companies to ensure delivery.
However, this also means companies need to invest more capital to maintain production.
Optical modules cannot be delivered immediately upon receiving customer orders. From optical chips and lasers to DSPs, circuit boards, and optical components, any shortage in a critical link may affect final shipment.
Eoptolink's operational performance in Q3 has shown some changes.
The company's revenue for the first three quarters was CNY 31.850 billion, a year-on-year increase of 92.97%; net profit attributable to parent company shareholders was CNY 11.372 billion, a year-on-year increase of 79.74%.
However, single-quarter revenue in Q3 was CNY 10.940 billion, and net profit attributable to parent company shareholders was CNY 3.843 billion, representing sequential declines of 12.98% and 19.07%, respectively.
Company operational explanations and related reports mentioned that floods in Thailand affected factory shipments, and tight raw material supply also impacted output and delivery.
This does not indicate that customer demand has decreased, but rather shows that order growth and revenue recognition are not always synchronized.
Cash flow is also worth noting.
Eoptolink's net cash flow from operating activities for the first three quarters was CNY 1.464 billion, a year-on-year decrease of 68.42%. The company explained that this was mainly due to increased payments for materials and various taxes.
Zhongji InnoLight's net cash flow from operating activities for the first three quarters was CNY 8.008 billion, a year-on-year increase of 46.81%, but accounts receivable at the end of Q3 reached CNY 17.309 billion, a 175.76% increase from the beginning of the year. The company attributed this to increased receivables resulting from sales growth.
For rapidly growing companies, increases in inventory, prepayments, and accounts receivable are not necessarily bad. The key lies in whether these capital occupations can ultimately be converted into shipments and collections.
If orders are executed smoothly, early stocking can help companies expand revenue; if customers delay project schedules, companies must bear longer inventory turnover cycles and capital occupation costs.
This is why the market has started paying more attention to the cash flow of optical module companies, rather than just profit growth rates.
More Than Just an Order Issue
Another change facing the optical module industry comes from the return-on-investment pressure faced by overseas tech giants themselves.
Amazon's Q3 AWS revenue grew by 39.3%, but free cash flow over the past 12 months turned into a net outflow of USD 42.4 billion, and the company expects cash capital expenditure to be approximately USD 220 billion in 2026.
Alphabet expects capital expenditure to reach USD 195–205 billion in 2026 and stated that capital expenditure will increase significantly again in 2027.
These investments remain staggering, but the capital market is increasingly focusing on a question: Can the revenue generated by AI businesses gradually cover the rising infrastructure costs?
If cloud vendors continue to expand capital expenditure, optical module companies still have opportunities to secure new orders; if the return on investment for some projects falls short of expectations, customers may rearrange construction sequences and procurement rhythms.
This change may not signify the end of the AI infrastructure investment cycle, but it could make performance differentiation among suppliers more pronounced.
For companies like Zhongji InnoLight and Eoptolink, which have already entered the supply chains of overseas core customers, future challenges lie in maintaining market share and converting product upgrades into profits.
For other companies in the supply chain, competition may become more direct. Customer certification, product performance, delivery capability, and cost control will all influence whether they can enter the next round of procurement lists.
The optical module industry still has clear demand for technological upgrades, but industry growth does not mean every company can grow synchronously.
The sharp rebound of the sector on November 11 indicates that the market has not given up on the AI computing power supply chain. However, the previous correction also shows that investors are no longer satisfied with hearing news about cloud vendors increasing capital expenditure and optical module companies having full order books.
What the market needs to see next is actual shipment volumes, the proportion of 1.6T products, gross margins, and operating cash flow.
For optical module companies, orders remain important, but whether orders can be delivered on time and how much profit ultimately remains are becoming more critical issues than the orders themselves.