The TikTok influencer "Li Yien" has found his formula for viral success. Before daily market commentary, he always begins with the slogan: "As I always say, time will prove the value of optical modules and computing power." After repeating this for over a year, likes on his videos have surged from a few thousand to 40,000–50,000. In the comments, netizens all ask the same question: Isn’t it too late to now be "standing in the light"?
The three characters that tied together netizens’ anxiety are “Yi Zhongtian.” This isn’t referring to the scholar from百家讲坛, but rather a nickname given by A-share investors to three leading optical module companies: Yi Sheng, Zhongji Xuchuang, and Tianfu Communications—each contributing one character. From the April 2025 low, Yi Sheng rose 16-fold, Zhongji Xuchuang 17-fold, and Tianfu Communications 10-fold. Those who bought early made a fortune.
But the story took a turn in June 2026. On June 5, "Yi Zhongtian" collectively plunged, with Zhongji Shuangchuang falling nearly 8% in a single day. On June 11, Xinyisheng approached its daily trading limit during intraday trading, and the CPO concept began to retreat. Those fleeing the market and those rushing in to buy the dip completed their handoff amid record trading volumes.
The stories of getting rich have already been overdone. The real question no one has seriously answered is this: If you can only choose one out of three, which one is the most worthwhile? In this article, we won’t discuss whether it’s still possible to get on board—only one question: Among the three, which offers the highest value for money?
No one is looking at current P/E ratios anymore during this round of the optical module rally.
The reason is simple: when a company’s profits are still growing at triple-digit rates, using the past twelve months’ earnings to calculate the P/E ratio yields meaningless numbers. The market’s pricing anchor has shifted from “how much is earned this year” to “how much will be earned in 2026 and 2027.” By early 2026, the stock price increases since 2025 for the three companies were 428% for Zhiwei旭创, 410% for Xinyisheng, and 284% for Tianfu Communications, resulting in a combined market value increase of over one trillion yuan. This trillion yuan is not buying today’s performance—it’s buying expectations for the next two to three years.
So here, "value for money" isn't about which stock price is lower—it needs to be measured with three different metrics. The first is PEG, which is the forward P/E ratio divided by growth rate, assessing "how much you're paying for the same level of growth." The second is profit quality, evaluating how clean the earnings are and how high the gross margin is. The third is the premium or discount for certainty, measuring how much extra the market is willing to pay for stability—or how much it discounts for uncertainty.
After measuring with three different yardsticks, the three providers offered three entirely different answers: one is the king of digital value for money, one is expensive but reliable certainty, and one is the most expensive certainty.
New Light: The king of digital value for money, but there's a reason why it's cheap.
First, look at the one with the lowest price.
Based on PEG, New Hua Sheng is the most cost-effective among the three. Its year-over-year growth in net profit attributable to shareholders in 2025 is nearly 2.5 times, significantly higher than Zhongji旭创’s projected growth of 89.5% to 128% during the same period. Fourth-quarter net profit increased by 39% sequentially, and the 1.6T product has ramped up ahead of schedule. Despite such robust growth, its valuation remains the lowest. Calculated based on the consensus institutional forecast for 2026 net profit, its forward P/E ratio is only about 22.8x, yielding a PEG of approximately 0.30—the lowest among the three. For each unit of growth, you pay the least for New Hua Sheng.
It’s not just cheaper—it also earns the cleanest profits. In its 2025 financial results, New Light’s non-recurring gains and losses amounted to only RMB 33 million, while maintaining a gross margin above 47%, thanks to cost advantages from vertical integration. In terms of profit quality, it even outperforms the larger player, Zhongji旭创.
At this point, New Edge appears to be an undervalued黑马 in the market. But this is precisely where you cannot stop at the surface. Its low price is a discount, not a windfall.
The market doesn't arbitrarily discount a high-growth company. The undervaluation of Eoptolink stems from several real risks: high customer concentration, with performance heavily reliant on a few major clients; overseas revenue accounting for 78%, directly exposing the company to tail risks from tariffs and trade restrictions; and most critically, whether the "dark horse" can sustain its explosive growth. In terms of long-term technological narratives and forward-looking positioning, its story is not as compelling as InnoLight's. The low P/E ratio the market assigns essentially reflects a discount on sustainability.
This discount is partially being repaired. Since 2026, E-Commerce's stock price has risen over 79%, and the company has begun planning a Hong Kong listing. Capital is voting with its feet, pulling it from a "distrusted underdog" toward a "revalued leader." It's cheap, but the discount is narrowing.
And where exactly is the more expensive one stable?
InnoLight: The Certainty of Value
Zhongji Xuchuang's value lies not in low cost, but in certainty.
Just look at the numbers—it’s clear. In the first quarter of 2026, Zhongji Xuguang generated a quarterly revenue of RMB 19.496 billion and a net profit of RMB 5.735 billion. That single quarter’s profit exceeded its entire annual net profit for 2024. During the same period, its gross margin for optical communication transceiver modules rose from 34.65% in 2024 to 42.61%, an increase of nearly 8 percentage points. Both scale and profitability are growing—this is the hallmark of a market leader.
The confidence behind this is rooted in market share and technological superiority. Zhongji Xuchuang has secured more than half of NVIDIA’s 800G optical module procurement. For the 1.6T generation, leveraging its first-mover advantage by being among the first to complete NVIDIA’s validation, market expectations anticipate it capturing 50% to 60% of the share. During last quarter’s earnings call, company executives clearly outlined the timeline: “This quarter, key customers are beginning to deploy 1.6T and steadily increasing orders... We expect other major customers to大规模 deploy 1.6T between 2026 and 2027.” To meet this demand, the company is securing chips and expanding production capacity, with initiatives underway both domestically and internationally.
The cost is that it’s the most expensive. Infinera’s forward P/E ratio once reached 73 to 74 times, over 40% higher than Eoptolink. What you’re paying for is a premium for being a market leader with technological superiority. This premium is suited for those who prioritize certainty and can afford the higher price.
But certainty does not equate to the absence of risk—in fact, its risks are more of a “black swan” nature. On June 8, 2026 (U.S. time), InnoLight was added to the U.S. Department of Defense’s “1260H List.” The company responded urgently, stating that this designation contradicts objective facts, as it is neither a defense contractor nor a military-civil fusion enterprise, and that the listing has not materially impacted its operations—orders, production, and supply chains remain unaffected. Yet, despite such reassurances, for a company with 86.8% of its revenue coming from overseas, geopolitical risk remains the true sword hanging over its head. It may not affect fundamentals, but it can slash valuations on any given trading day.
After dismantling the two module factories, there’s still Tiantfu, which isn’t even at the same table.
Tianfu Communications: The most expensive certainty bets on the next-generation architecture.
What makes Tiantfu Communications unique? It doesn’t sell modules—it sells “water.”
The most intuitive analogy is to think of it as a supply chain. If Zhongji Xuchuang and XinYisheng are restaurants directly serving customers, TFCOM is the supplier to those restaurants. It sells core components like optical engines and photonic devices to downstream optical module manufacturers, who then assemble them into complete modules for shipment. TFCOM doesn’t take orders directly from cloud providers, but every high-end optical module contains its components.
Positioned upstream, it boasts the highest gross margin among the three, consistently maintaining over 50%, with the clearest competitive landscape. More importantly, it has bet on a highly certain long-term trend: the CPO/NPO architecture. According to institutional estimates, in the value chain of a high-end 51.2T switch, the combined potential value of Compal Communications in the FAU, precision lens, and optical engine packaging segments could reach $7,000 to $10,000.
Compared to the single-machine component values of just tens of dollars in the traditional plug-and-play era, this represents a complete surge in both volume and price. Regardless of which module solution downstream cloud providers ultimately choose, as long as data centers continue evolving toward more efficient and energy-saving architectures, the position of the “water sellers” remains secure.
Sounds great. But Tianfu’s issue is precisely hidden in the three words “water seller”—it has the least flexibility, the highest valuation, and is most prone to disappointing expectations.
Its earnings elasticity is low because its growth is a steady stream, not a pulse. Zhongji旭创 and Xinyisheng directly benefit from the pulse-like surge in AI capital expenditures, resulting in massive earnings elasticity. Tianfu’s growth is stable but gradual. Its valuation is high because the market has priced in this certainty far too optimistically. As of February 10, 2026, its trailing P/E ratio stood at approximately 122x, significantly higher than the other two companies. The most recent and painful example of an earnings surprise occurred in Q1 2026: institutional consensus forecasted net profit between RMB 780 million and RMB 820 million, but the actual figure was only RMB 490 million. This massive shortfall resulted from institutions applying the pulse-driven logic of module manufacturers to a passive optical component company.
This precisely reminds everyone trying to rank Yi Zhongtian: Tianfu and the other two are not in the same league. Pricing an engine seller using the logic for whole machine sellers is itself a misunderstanding.
That’s all three split up. But there’s one hidden variable in the question of “value for money” that everyone has missed.
The profit pool is not in their hands at all.
Back at the table, ask a tougher question: Is the money Yi Zhongtian earned truly "good money"?
The essence of an optical module is system integration: purchasing optical chips, electrical chips, and optical components, then assembling them into a complete module through packaging processes. The barrier does not lie in the assembly itself. The true profit pools and competitive advantages are concentrated at the two ends of the supply chain: upstream laser chips and switching chips. Chinese manufacturers dominate the middle-stage assembly process.
Therefore, the claim that “InnoLight is crushing Lumentum and Coherent” should be viewed on two levels. In terms of module market share, it holds true—InnoLight has indeed surpassed these two established U.S. manufacturers. However, when it comes to profit quality, it’s a different story.
Lumentum and Coherent are precisely positioned upstream. They hedge against supply shortages through vertically integrated laser chip production, and the advantages of III-V platforms such as indium phosphide and gallium arsenide in high-power applications remain real today. Moreover, these two companies are not defeated rivals—they are upstream players rapidly recovering. Lumentum’s revenue increased by 58% year-over-year in the first quarter of fiscal year 2026, with gross margin rising from 28% to 34%.
Coherent reported a single-quarter revenue of $1.81 billion, a 21% year-over-year growth. Data center and communications businesses now account for 75% of total revenue, with growth exceeding 40%. Non-GAAP gross margin increased to 39.6%.
An even sharper point lies ahead. Yi Zhongtian’s trillion-dollar valuation bet hinges on the transition to CPO architecture. And CPO cannot do without CW light sources and indium phosphide substrates—both of which lie squarely in the domain of U.S. manufacturers. Coherent is doubling its indium phosphide capacity; its facility in Sherman, Texas, is the world’s most advanced indium phosphide production line, ramping up CW lasers specifically for solutions including NVIDIA’s CPO.
The more Yi Zhongtian bets on architecture upgrades, the more he is expanding the market share for upstream U.S. chip manufacturers. Thus, Yi Zhongtian earns money from assembly and components, while Coherent and Lumentum earn money from chips. The latter are thinner, slower, yet more durable.
This is also why people often mention "Yuanjie High-Power Laser Diodes" alongside "Yi Zhongtian." Yuanjie Technology represents China’s effort to climb up the laser chip supply chain. Its 100G EML has already passed customer validation in 2025 and is set to enter mass production in 2026, while its CW 100mW high-power light sources have achieved bulk deliveries, with first-quarter revenue more than tripling year-over-year. If China can truly break through at this critical bottleneck—EML and high-power laser chips—then Yi Zhongtian’s moat will extend from assembly to chip-level integration, making the partnership truly robust. If it fails to climb this high, no matter how cost-effective it is, it will still only be earning hard-earned profits.
This is the true hidden variable for measuring the long-term value for money among the three—not who has a lower PEG, but whether China’s optical module industry can seize the profit pool from upstream players.
Whether time will prove the value of optical modules and computing power is unknown. But at least, those standing in the light should first clarify which beam they are standing in.