Introduction: The AI optical communication supply chain, previously the most sought-after by investors over the past year, collectively plummeted.

On October 8, the optical chip sector suddenly faced a stress test.

Yuanjie Technology and Changguang Huaxin both plummeted by 20%, hitting their daily limit, while Dongshan Precision also hit its limit. A batch of optical communication supply chain companies, including Shijia Guangzi, Tengjing Technology, and Tecolet, saw significant declines. The Wind Optical Chip Concept Index fell by 8.18% on the same day.

Among these, Yuanjie Technology remained the most closely watched.

Once priced at nearly 2,000 yuan and ranked among the most expensive stocks in the A-share market, this leading optical chip company closed at 1,290.40 yuan on October 8, shedding 20% in a single day.

More intriguingly, this plunge occurred at a time when the company's fundamentals appeared to be at their strongest.

In the first half of 2026, Yuanjie Technology achieved revenue of 925 million yuan, a year-on-year increase of 351.40%; net profit attributable to shareholders reached 607 million yuan, up 1,212.20% year-on-year; and net profit after excluding non-recurring items was 514 million yuan, a year-on-year increase of 1,033.70%. Revenue from its data center business reached 774 million yuan, a year-on-year increase of 640.29%, while the company's overall gross margin rose to 80.42%.

Why did a company with a six-month profit growth exceeding 12 times and a gross margin exceeding 80% suddenly experience a 20% daily limit decline?

On the surface, the answer points to the FCC.

During the National Day holiday, overseas institutions discussed potential regulatory pathways by which the U.S. Federal Communications Commission (FCC) might further restrict Chinese-made optical modules, sparking market concerns about the overseas business of China's optical communication supply chain.

However, a closer look at policy texts and market rumors reveals that the situation is not as straightforward as the market's reaction suggests.

The FCC does have a formal rule set to take effect.

Its Third Report and Order, adopted in July, attempts to close the so-called 'component loophole' in the equipment authorization system: if equipment contains hardware components with logical functions produced by entities on the FCC Covered List, the FCC authorization for new equipment may be restricted under certain conditions. This rule will officially take effect on October 13.

But this is not a '3.2T ban' targeting Chinese optical modules.

The widely circulated market notion of 'restrictions starting from 3.2T products' while 'existing 800G and 1.6T products remain unaffected for now' primarily stems from deductions by overseas institutions, not from officially released FCC rule texts.

Overseas institutions speculate that if further restrictions on Chinese-made high-speed optical modules are implemented in the future, they are more likely to be phased in starting with the next-generation 3.2T optical transceivers, with current mainstream 800G and 1.6T products not directly affected for now.

In other words, what the market traded on October 8 was not actual business losses but a policy risk that might materialize years later.

This is the most noteworthy aspect of this plunge.

The real 'exam results' for AI optical chips may not be their performance but their valuations.

01 Risks Shift from Complete Systems to Upstream Components

One interesting aspect of the market is that if the so-called FCC risks target China's optical communication supply chain, the most direct exposure in the U.S. data center market should theoretically be with complete optical module manufacturers first.

However, on October 8, the most intense market reaction occurred in upstream optical chips.

Yuanjie Technology and Changguang Huaxin hit their daily limits, while leading optical module companies reacted relatively mildly overall.

One reason is that this is not the first time optical module companies have faced such policy expectations.

In early August this year, rumors emerged about potential U.S. restrictions on importing new Chinese optical modules.

At the time, leading companies like Zhongji Xunce, Xin Yisheng, and Tianfu Communication saw their stock prices drop noticeably during intraday trading before quickly narrowing their losses.

On August 5, the A-share optical module sector even ended up rising by 3.98%. Companies like Zhongji Xunce publicly responded that the U.S. FCC had not yet issued restrictive documents in relevant areas and that their operations remained normal.

Back then, it amounted to a round of policy risk pricing for leading optical module companies.

By October, the risk narrative had changed.

Overseas institutions began to further speculate that even if Chinese companies are not directly banned from producing optical modules in the future, the proportion of locally sourced components (BOM) in the U.S. might be increased to boost the share of U.S.-made chips and devices in next-generation high-speed optical modules.

The direct impact of this assumption shifts from 'who produces the optical modules' to 'whose chips are inside the modules.'

Risks began migrating from downstream complete systems to upstream value components.

This is precisely why optical chips were heavily sold off by the market.

Some institutions speculate that potential regulatory frameworks might drive greater use of U.S.-supplied products for high-value devices like DSPs, PDs, EMLs, and CW Lasers. While still speculative, this strikes at the core valuation logic of domestic optical chips—shifting from domestic substitution to global substitution.

Over the past two years, capital markets have imagined domestic optical chips not just as replacements for overseas products in the domestic market but also as opportunities for Chinese optical chip manufacturers to enter the supply chains of North American cloud giants alongside domestic optical module leaders.

The former determines domestic market space, while the latter sets the global market ceiling. The high valuations given to companies like Yuanjie Technology are partly based on this second possibility.

Thus, even if potential rules may only truly impact the 3.2T era, and even if existing 800G and 1.6T orders remain unconstrained in the short term, stock prices still adjusted first.

The stock market never trades on this year's profits; it trades on the discounted value of future profits, especially for a stock that has already surged above 1,000 yuan.

The Shanghai Securities News cited investor communications from Zhongji Xunce, stating that significant volume production of 3.2T optical modules is not expected until around 2028-2029, with 800G and 1.6T remaining the industry's mainstream products from 2026 to 2028.

In other words, even under the market's speculated policy pathway, the impact pertains to a distant market.

But for high-valuation growth stocks, three years are never distant because today's market capitalization relies on profits three years from now.

02 A Single Message Pierces Sector Sentiment

If only the FCC rumors existed, Yuanjie Technology might not have needed to hit its daily limit.

On October 8, another message simultaneously circulated: potential price drops for optical chips.

Markets briefly speculated that prices for 1.6T- matching optical chips were under pressure, with high premiums on some premium optical chips potentially loosening due to silicon photonics advancements, improved supply, and changes in customer procurement strategies.

However, like the FCC's 3.2T rumors, this message currently lacks confirmation from listed companies.

Changguang Huaxin explicitly stated it had not received any notices of optical chip price reductions; Shijia Guangzi and Yongding Co., Ltd. also said they were unaware of such developments. Shijia Guangzi further speculated that this stock price adjustment might be more related to market discussions about FCC policies during the National Day holiday.

Thus, as of October 8, interpreting this plunge as '1.6T optical chips have fully dropped in price' lacks sufficient evidence.

But capital markets remain sensitive.

Because the price drop rumors touch on another critical variable in Yuanjie Technology's valuation: how long can its 80% gross margin be sustained?

In the first half of 2026, Yuanjie Technology's overall gross margin reached 80.42%. Its data center business gross margin even approached 84%.

For a manufacturing company, this is an almost lavish profit level.

It indicates that, at least at this stage, high-end optical chips remain in a clear seller's market.

Rapidly growing demand coexists with a limited number of suppliers capable of stably providing high-power, high-speed laser chips, granting upstream players strong bargaining power.

This is also why optical chips became one of the most sharply rising segments in the entire AI hardware supply chain over the past year.

GPUs are expensive, optical modules are scarce, and optical chips are even scarcer.

The scarcer a link in the supply chain, the higher its profit margins.

Capital thus assigned the highest valuations to the scarcest links.

But all high-margin manufacturing industries face an inevitable issue: high profits attract competition.

When a product boasts an 80% gross margin, industrial capital cannot remain indifferent indefinitely.

Existing players will expand production.

Competitors will develop new products.

Downstream customers will actively cultivate secondary and tertiary suppliers.

Manufacturing yields will improve.

Unit costs will decline.

Ultimately, supply-demand dynamics will gradually shift from 'severe shortages' to 'tight balance' and then to mature supply.

Last year's scarcest products may not remain scarce forever.

This is why even without genuine price reduction notices, a single price rumor can swiftly pierce sector sentiment.

Investors worry not about 'how much cheaper it is today' but begin to question:

Is an 80% gross margin a long-term moat or merely a temporary dividend from the early stages of AI infrastructure expansion?

The two answers have vastly different implications for valuations.

If Yuanjie Technology's profitability primarily stems from technological barriers, it could still maintain high margins through continuous product performance improvements as 1.6T and 3.2T upgrades occur.

However, if the current ultra-high margins include significant price premiums driven by supply-demand imbalances, margins will eventually normalize as production expands and competition intensifies.

Yuanjie Technology is expanding production, as are others in the industry.

This is precisely the most challenging question for capital markets: continued growth in AI optical communication demand does not imply perpetually rising optical chip prices.

Market expansion and price declines can occur simultaneously.

This has been the history of the semiconductor industry for decades.

Demand grows continuously, technology advances steadily, and unit prices decline persistently. Only companies capable of reducing costs through process improvements, yield enhancements, product iterations, and customer stickiness can truly retain profits.

After October 8, the market may no longer settle for asking: 'How many 1.6T units are shipped?'

Instead, it will further inquire: 'After significant 1.6T shipments, how much profit will remain?'

03 A Divergence Begins with a Daily Limit

Interpreting Yuanjie Technology's plunge as the end of the AI optical communication industry's logic would be overly simplistic.

At least currently, no data supports this judgment.

In the first half of 2026, Yuanjie Technology's data center business revenue reached 774 million yuan, a year-on-year increase of 640.29%; its 100G PAM4 EML products achieved bulk deliveries, 200G PAM4 EML products are undergoing customer verification, and high-power CW products continue to ramp up.

Changguang Huaxin reported first-half revenue of 391 million yuan, a year-on-year increase of 82.67%, and net profit attributable to shareholders of 30.3378 million yuan, a year-on-year increase of 238.04%. The company attributed a significant portion of its growth to rapid expansion in its previously established optical communication business.

From an industry sentiment perspective, at the China International Optoelectronic Exposition held in September this year, multiple supply chain companies told Securities Times that high-speed optical module orders remain robust, with 800G becoming the mainstream product and 1.6T in the sampling and bulk delivery stages. A key factor currently constraining industry expansion remains insufficient capacity for some core devices.

Thus, October 8 did not prove that AI demand has vanished, that 800G or 1.6T products have suddenly become unsellable, or even that the most feared 3.2T policy risks have materialized beyond overseas institutional speculation.

This is the most intriguing aspect of this plunge: fundamentals did not change dramatically in a single day, but valuations did.

From this perspective, Yuanjie Technology's plunge resembles a valuation shift.

Previously, the market priced optical chip companies based on three key words: shortages, upgrades, and domestic substitution.

As long as AI capital expenditures continued expanding, 800G upgraded to 1.6T, and Localization rate (domestic substitution rates) kept rising, investors were willing to continually push back profit forecasts.

With vast market space projected for the next three years, today's market capitalization could also remain lofty.

This trading approach centers on calculating market space.

But when a company's profits actually begin materializing—and at a pace as rapid as a 12-fold increase in six months—investors shift to the next stage: calculating profits.

How long can profit growth last?

What gross margin can be sustained?

Can it enter North American supply chains in the 3.2T era?

Will customers demand price reductions?

Will new capacity trigger price competition?

How will chip value quantities change after silicon photonics penetration?

Every question directly alters final valuations.

This indicates that the optical chip sector is undergoing a transition from 'industry Beta' to 'company Alpha.'

Previously, belonging to the AI optical communication supply chain was enough to lift valuations collectively.

In the future, this may no longer suffice.

Yuanjie Technology and Changguang Huaxin are both optical chip companies, but their profit stages now differ entirely.

Yuanjie Technology reported first-half net profit attributable to shareholders of 607 million yuan, non-recurring net profit of 514 million yuan, and net cash flow from operating activities of 488 million yuan, indicating massive profit and cash flow realization.

Changguang Huaxin, meanwhile, reported first-half net profit attributable to shareholders of 30.3378 million yuan but a non-recurring net loss of 10.3844 million yuan.

In the past, the market could lump both companies under the umbrella of 'domestic optical chips.'

In the future, this may no longer be possible.

Even among companies entering the 1.6T era, some sell EMLs while others sell CW Lasers;

even in the shift toward silicon photonics, some segments gain value while others lose it; even facing 3.2T, some enterprises have entered overseas customer systems while others remain in verification stages; even with high revenue growth, some stems from genuine technological barriers while others may derive from short-term industry shortages.

The more mature the industry becomes, the greater these disparities grow.

Thus, the truly critical questions for optical chip investments are shifting: from 'how large is the market?' to 'whose profits will it be?'

This is also key to understanding the October 8 market action.

04 Capital No Longer Buys Just 'Optical'

The FCC risks circulating in the market currently remain more of a long-term policy gamble.

The notion of restrictions starting from 3.2T does not appear in the FCC's existing formal rule texts; 800G and 1.6T products have not suddenly lost their markets due to these policy speculations.

Industrial demand has not suddenly reversed.

What has truly changed is capital's risk appetite.

Optical module companies already weathered a round of policy rumor shocks in August. This time, the market directed its most intense reaction toward optical chips, which had previously seen greater gains, higher valuations, and stood at the center of the domestic substitution narrative.

This appears to be policy-driven trading.

Essentially, however, it resembles a valuation stress test.

During the previous rally, the market willingly believed three things simultaneously: AI investments would keep growing, domestic optical chip market share would keep rising, and the scarcity of high-end optical chips would persist.

After October 8, these three beliefs remain unfalsified.

But for the first time, investors are concentration (intensively) questioning their boundaries.

If 3.2T supply chains truly require localization in the future, how many overseas orders can domestic chips secure?

If capacity releases, how long can an 80% gross margin be sustained?

If silicon photonics and CPO technologies iterate rapidly, will today's scarcest products remain scarce three years from now?

These questions could be left for the future in the past.

Not anymore.

Because Yuanjie Technology is now a company with a market capitalization of hundreds of billions of yuan.

When a company's half-year profit surges from tens of millions of yuan to 600 million yuan, the capital market will no longer simply reward it for 'entering the AI track'.

In the next phase, it needs to prove that: the 600 million yuan profit is not a fluke, the 80% gross profit margin is not due to a temporary shortage, there will be 3.2T after 1.6T, there is an overseas market beyond the domestic market, and there will be growth next year after this year's growth.

This is where the real challenge lies for stocks priced at a thousand yuan.

When an industry is evolving from 0 to 1, the capital market buys into imagination.

From 1 to 10, it buys into penetration rate.

After reaching 10, it ultimately comes down to profits, cash flow, and competitive landscape.

On October 8th, the AI optical communication industry did not come to a halt.

It's just that the capital market has, for the first time, taken out its calculator so seriously.

Over the past year, funds have been buying into 'optics'.

From now on, it may start to be selective about the companies within the optics sector.

The valuation of AI optical chips is now being put to the test.