11/20 2026
524
Preface
Every face unlock on a phone, every emergency brake as a car avoids a pedestrian on a rainy night, every millisecond a GPU cluster in a data center exchanges data—behind all of it shines the same kind of light.
On November 4, Vertilite, a Changzhou-based company, submitted an IPO counseling filing to the Jiangsu Securities Regulatory Bureau. The gap between volume and revenue has become the first question it poses to the capital market.
Author | Fang Wensan
Image Source | Internet

900 Million Chips Shipped Cannot Support the Appetite for a 10-Billion-Yuan Valuation
In 2017, Apple released the iPhone X, removing the Home button and switching to 3D facial recognition. The small hole above the screen instantly projects tens of thousands of infrared light dots onto the face, and the core device emitting these light dots is the VCSEL.
Before 2015, high-end VCSEL chips were almost monopolized by two major U.S. manufacturers, Lumentum and Coherent, with more than 80% of global capacity in the hands of overseas vendors. Every Apple stockpiling season, premium capacity was bought out first, and domestic manufacturers had to queue up a year in advance to get goods.
What Vertilite wants to tear open is exactly this gap. Founder Zhao Li was born in 1988. After graduating from Huazhong University of Science and Technology with a bachelor's degree, he went to Stanford for a PhD under James Harris, a member of the U.S. National Academy of Engineering. During his doctoral studies, he produced the world's first room-temperature continuous-wave long-wavelength VCSEL.
In November 2015, he returned to China with several fellow students to start the company. In 2019, they built China's first mass-production-grade VCSEL in Changzhou. In 2020, they entered Huawei's Mate 30 Pro supply chain. The shipment ramp-up was quite impressive.
In 2023, cumulative shipments surpassed 100 million units, reaching 600 million by 2025 and 900 million by September 2026, with a domestic consumer electronics VCSEL market share exceeding 70%.
But the financials tell another story. An investment announcement by shareholder Dongyangguang disclosed that in 2024 Vertilite's revenue was 125 million yuan and net loss was 58.2982 million yuan; in the first quarter of 2025, revenue was 41.2343 million yuan and net loss was 15.3154 million yuan.
Cumulative shipments of 900 million units against annual revenue of just over 100 million yuan—a simple division reveals the order of magnitude of the unit price in this business. A consumer-grade VCSEL often sells for only a few dimes to a few yuan.
No matter how fast matches sell, a matchbox is still a matchbox. More importantly, in 2025 consumer electronics still account for 45% of global VCSEL demand, but annual growth has already fallen to 3.5%, and this ceiling is visible to the naked eye.

A 60% Automotive Share Does Not Buy Pricing Power
Before consumer electronics peaked, Zhao Li bet on automotive LiDAR, a move with a much steeper threshold.
Automotive-grade chips must withstand temperature differences from minus 40 degrees Celsius to 125 degrees Celsius, ensure stable operation for 15 years, and require a yield above 95%. A single LiDAR integrates hundreds of VCSELs; if one fails, the entire radar unit worth thousands of yuan is scrapped.
Vertilite spent two and a half years obtaining AEC-Q102 automotive-grade certification, used an anti-reflection architecture to push detection range beyond 200 meters, and became a core light source supplier for Huawei, Hesai, RoboSense, and Waymo, with annual supporting installations exceeding one million units. It captured more than 60% of the domestic new energy vehicle LiDAR light source share, and its automotive VCSEL market share rose to second globally.
The share was won, but pricing power remained downstream. In the past two years, the price war among new energy vehicle makers has become brutal. The unit price of automotive LiDAR has been halved from nearly 10,000 yuan in the early years to around 2,000 yuan. Laser and detector chips account for 30% to 40% of radar material costs, and cost-reduction pressure is transmitted all the way down the supply chain to the chip end.
Shipments have surged exponentially while unit prices have declined year after year, forming a cycle of compensating price with volume. Peer financials confirm this harsh reality: Changguang Huaxin's 2025 VCSEL and optical communication chip revenue was 41.1938 million yuan, a year-on-year surge of 1,036.79%, yet gross margin fell by 37.5 percentage points to only 12.81%.

Indium Phosphide: The Unit Price in This Story Is Measured in Dollars
Optical communication is Vertilite's third pivot, and the only one that can lift unit prices from a few dimes to tens of dollars.
In 2026, the company's 100G VCSEL based on gallium arsenide materials has already entered volume supply to optical module giants such as Zhongji Innolight and Eoptolink. Its 100G and 1310-nanometer photodetectors have entered the volume shipment stage, and CW DFB lasers have entered trial production.
Large models are making computing clusters larger and larger. Copper wires inside server racks can no longer keep up, and the shift from copper to optics has become a rigid necessity.
Third-party institutions predict that the global VCSEL market will grow from $2.7 billion in 2026 to $12.6 billion in 2036, with a compound annual growth rate of 16.6%, among which data center optical interconnects are the thickest pipeline.
In 800G and 1.6T ultra-high-speed optical modules, the highest-priced and deepest-barrier components are indium phosphide-based EML and DFB laser chips. In this market, U.S. and Japanese manufacturers hold more than 95% of the share in high-speed optical communication chips above 25G.
In the frontier indium phosphide field, a mature foundry ecosystem has not yet formed globally. Material growth, structure design, and chip processes are highly bound together. Without building one's own line and repeatedly debugging, core process know-how will always remain in others' hands.
So this design-origin company decided to buy land and equipment. In 2025, the FabX production line, with an investment of 550 million yuan and planned annual capacity of 50 million units, was completed and put into operation. In August 2026, it signed a 5-billion-yuan total investment project for a high-end indium phosphide laser chip base in Wujin, Changzhou.
An asset-light design company turning into an asset-heavy IDM giant is the heaviest page in the prospectus.

5 Billion Yuan Lands in the Factory: Incremental Expectations for the New Business
What was signed and settled in Wujin High-tech Zone in August 2026 is a high-end optical communication chip and device R&D and manufacturing base, with continuous investment totaling about 5 billion yuan, building R&D and mass production lines for high-end indium phosphide laser chips.
Vertilite's rationale is quite practical. The epitaxy and processes for gallium arsenide VCSELs are already mature, and foundries can be found. On the indium phosphide path, there is no established foundry division of labor anywhere in the world. Material growth, structure design, and chip processes are tangled together, and process know-how can only be accumulated through repeated debugging on one's own production line.
The company has concluded that the old gallium arsenide craft cannot support the next eight years. The most expensive light sources in 800G and 1.6T optical modules are indium phosphide-based EML, DFB, and CW lasers. Short-reach multimode VCSELs are confined to the short distance inside the rack, and as speeds increase, silicon photonics and CPO increasingly tend to start anew.
The product progress on the company's official website already spells out the pivot: 100G VCSELs and 100G, 1310-nanometer photodetectors are in volume shipment, while CW DFB is still in trial production. Changing the material system is equivalent to admitting that the ceiling of the old base is right overhead.
In Dongyangguang's capital increase announcement dated June 17, 2025, besides the pre-investment valuation of 3.285 billion yuan, two less-circulated figures were also included: Vertilite's total assets were 895 million yuan at the end of 2024 and 878 million yuan at the end of March 2025.
The previously implemented FabX project involved an investment of 550 million yuan and was completed and put into operation in 2025. The 3-inch high-speed communication optical chip production line is planned to produce 50 million units annually, a step sufficient only for pilot testing and small-scale mass production.
The 5-billion-yuan step must be lifted jointly by the IPO, local industrial funds, banks, and equipment financing. In 2025, the output value of above-designated-size optoelectronics enterprises in Changzhou was 59.4 billion yuan; from January to September 2026 it was 48.3 billion yuan, up 15% year on year. In the past two years, 36 projects were signed and under construction, with total investment exceeding 20 billion yuan, and the 5-billion-yuan project alone accounts for one quarter.

Conclusion: Policy Opens the Door, and the Timer Starts at the Same Time
In March 2026, Nvidia invested $2 billion each in Coherent and Lumentum, totaling $4 billion, accompanied by multi-billion-dollar procurement commitments and capacity priority rights. Procurement is scheduled from early 2027 through 2030, with the money mainly going to indium phosphide capacity expansion and building factories in the United States.
Coherent's rare 6-inch InP production line globally is planned to double internal capacity by the end of 2026, and Lumentum has publicly said that capacity through the end of 2028 has already been booked out.
According to domestic third-party research institutions, the domestic localization rate of high-speed EML above 100G is less than 5%, and about 4% overall for 25G and above. The InP substrate capacity expansion cycle is 18 to 24 months, and after equipment is in place, yield ramp-up requires another 8 to 12 months.
The window left for domestic substitution is roughly pinned to the two certification cycles of 2027 and 2028. Optical module manufacturers changing light sources must rerun reliability tests and re-pass customer validation, and 18 to 24 months is the industry norm.
Domestic peers are all on the same starting line. Huaxin Semiconductor's 25G and 56G high-speed VCSELs are already in mass production, with 112G in small-batch production. Ruikang Guanglian and Eagle Semiconductor are each moving toward single-wavelength 100G and 200G. Changguang Huaxin and Yuanjie Technology have each made progress in sampling 100G EML and 200G.
Vertilite's advantage is that it has run through the entire discipline of automotive-grade mass production. The yield and delivery experience accumulated over two and a half years of AEC-Q102 certification remain valuable even after changing the material system.
Some materials referenced: Star Market Daily: "Vertilite Initiates IPO Counseling, VCSEL Cumulative Shipments Exceed 900 Million Units, with Guokai Manufacturing Fund, Xiaomi, BYD and Others Betting"; The Paper: "Vertilite Submits IPO Counseling Filing, Initiates STAR Market Listing Counseling"; China Securities Journal: "Deepening Bottom-Layer Innovation in Optical Chips, Vertilite Builds Hard Strength to Ride Through Cycles"