For stock investors, authoritative, professional, timely and comprehensive analyst research reports help uncover potential thematic opportunities.
"Someone please save me!" retail investor Xiaodian cried out repeatedly. "I not only hold Eoptolink, but also Sungrow." Facing the sharp decline in Eoptolink Technology Inc.,Ltd. (ASX: 300502), Xiaodian felt somewhat desperate.
On October 22, the optical module sector suffered a steep selloff. Among the "Yi Zhong Tian" trio — Eoptolink, Zhongji Innolight and T&S Communications — Eoptolink led the decline with a 12.28% drop on heavy volume, while Cambridge Technology hit the daily limit down, T&S Communications fell 8.36%, Zhongji Innolight dropped 8.76%, and Dongshan Precision declined 7.84%, with all hitting recent lows.
Why did optical modules tumble? Has the moment to buy the dip arrived?
In my view, looking across all available information, the trigger for the optical module selloff was primarily a concentrated repricing of whether mainland Chinese manufacturers will still hold a core position in the "CPO era," compounded by the exposure of top fund Q3 position reductions, financing balances sitting at historically extreme percentiles, and a simultaneous contraction in external risk appetite.
The decline in optical modules is driven by four layers of pressure.
The first pressure: Nvidia's CPO expansion list "bypasses the mainland" — the most direct loosening of the valuation anchor. At the GTC Berlin conference on October 21, Nvidia announced capacity expansion with six supply chain partners, the most closely watched being its CPO collaboration with Taiwan's Siliconware Precision Industries (SPIL): optical engine capacity will be increased to 15 times the current level by mid-2027, and doubled again by the end of 2028. The market noticed that no mainland manufacturers appeared among the CPO expansion entities officially announced, sparking concerns over whether the next-generation technology roadmap is leaving mainland Chinese companies behind. In fact, this does not mean mainland manufacturers have been completely excluded. When Nvidia officially announced full-scale mass production of CPO switches in August, T&S Communications appeared on the supply chain list, responsible for laser module sub-component packaging and assembly; major domestic players are also actively engaged in custom development or sample validation. But the market's pricing logic is forward-looking: CPO is regarded as the next-generation core technology route for optical interconnection, and current high valuations — T&S Communications trades at a dynamic PE of roughly 91 times — largely price in Chinese manufacturers' continued participation in the CPO era. When the core expansion list lacks mainland entities, this layer of valuation narrative has developed cracks.
The second pressure: top fund Q3 position reductions exposed — "loosening of the herd" signals fermenting. The Q3 report for Zhongou Digital Economy, managed by Zhongou Fund's Feng Ludan, showed: Zhongji Innolight holdings fell from 889,300 shares to 575,000 shares, a reduction of 35.34%; Eoptolink went from 1,781,400 shares to 1,286,200 shares, a cut of 27.8%; the two optical module leaders together gave up more than 4 percentage points of the fund's net asset value weight, while new positions were initiated in names such as Yuanjie Semiconductor and SG Micro. This is lagged disclosure as of the end of September, but amid fragile sentiment, "top fund managers already made their move in Q3" was amplified into a "herd collapse" narrative.
The third pressure: chip structure is extremely crowded — financing balances are above the historical 90th percentile. As of October 21, Zhongji Innolight's margin financing balance stood at 25.658 billion yuan, representing 3.57% of its free-float market capitalization, exceeding the historical 90th percentile level; on October 20, the financing balance was 25.696 billion yuan, also at that percentile. The optical module index has gained about 84% year-to-date, and public funds' position concentration in electronics and telecommunications is at historically high levels. Under such a structure, the sector exhibits a classic pattern of "good news muted, bad news amplified": the four major North American cloud providers' combined 2026 capital expenditure guidance of $720–745 billion, and Goldman Sachs raising its 2027 1.6T optical module shipment forecast to approximately 71 million units (about 33 million units in 2026) — these industry prosperity signals failed to provide effective offset.
The fourth pressure: external risk appetite weakening in tandem. The U.S. optical communications sector fell broadly in pre-market trading, with Lumentum, Corning, Marvell and Coherent declining in a range of 1.5% to 3%. These declines are not extreme in themselves, but the direction is uniform, and A-share optical modules, as a high-beta sector, saw sensitivity further amplified during a window of converging global tech stock risk appetite.
So, has the moment to buy the dip arrived? Eoptolink responded that day: "The company's current production and operations are normal and orderly, and order delivery work is proceeding normally." 800G and 1.6T are the main delivery drivers this year, with 1.6T having entered a sustained volume ramp-up phase and sufficient orders in hand. Zhongji Innolight's orders in hand cover the full year of 2026, with some scheduled into 2027; some customers have already provided quarterly shipment guidance for 2027, and 800G and 1.6T demand in 2027 is expected to maintain rapid growth compared to 2026. Eoptolink's revenue for the first half of 2026 reached 20.910 billion yuan, up 100.34% year-on-year, with net profit attributable to shareholders of 7.529 billion yuan, up 90.98% year-on-year. From the perspective of industry prosperity and corporate profitability, current fundamentals have not reversed.
However, financing balances above the historical 90th percentile mean leveraged positions have not been fully flushed out; the heavy-volume decline — Eoptolink's single-day turnover approached 19.7 billion yuan — indicates enormous divergence between bulls and bears at the current position, rather than stabilization on shrinking volume. Some institutions note that the magnitude and speed of this optical module correction have reached historically extreme levels, and the full release of short-term selling pressure provides room for a technical rebound, but this is more of a post-oversold repair than a trend reversal. Before funding positions, macroeconomic and geopolitical factors substantially ease, the technology sector's medium-term trajectory will remain primarily range-bound bottoming.
The key variable is: what exactly the participation landscape in the CPO route will look like — there is currently no definitive answer. Nvidia's expansion list "bypassing the mainland" this time is a signal worth watching closely. When uncertainty has not been eliminated, the market's moment to buy the dip may not yet have arrived. Stock price stabilization often requires confirmation through shrinking volume to distinguish whether it is a bottom or merely halfway down the mountain.
Comments