A piece of glass less than 0.2 mm thick is now rattling the nerves of the global semiconductor industry.
Historically, glass has mostly been used in the screens in front of chips, responsible for presenting the digital world to our eyes. As AI (Artificial Intelligence) computing power drives the accelerated iteration of packaging systems, glass materials are moving from the front of the chip to the bottom, stepping from display materials into advanced packaging, and becoming a critical substrate for carrying chiplets and connecting computing power.
What exactly gives this piece of glass the power to disrupt the global semiconductor landscape? For Hubei, is this merely a chance to follow in the footsteps, or a new anchor point for overtaking on a different track?
In 2026, from Intel unveiling the world's first mass-produced processor featuring a glass core substrate at CES, to Apple reportedly directly introducing its self-developed AI server chips into glass substrate testing, and then to TSMC (Taiwan Semiconductor Manufacturing Company), Samsung, and SK Hynix intensively announcing their mass production timelines—this material, once only used for windows and screens, has overnight found itself at the forefront of the AI computing power race.
In the A-share market, glass substrate concept stocks have hit the daily limit up multiple times this year. In less than half a year, the market capitalization of Woge Optoelectronics surged from 7.9 billion RMB at the beginning of the year to a peak of approximately 34 billion RMB (as of June 16). The frenzy in the capital market reflects a clear industrial signal: in the post-Moore era, the "foundation" of chip packaging is being recast.
01. Why Are Giants Targeting Glass Substrates?
The answer lies not in the word "glass," but in the physical limits of AI computing chips.
To understand this "glass revolution," one must first clarify a pain point: the organic substrates traditionally used for chip packaging can no longer "bear the load."
Traditional Substrates Have Reached Their Physical Limits.
Under the trend of AI chiplet integration, the packaging area and interconnect density of a single package are approaching physical limits. Traditional organic substrates are prone to warpage when heated and face a ceiling in precision; although silicon interposers can achieve micron-level fine interconnects, they are limited by the 12-inch wafer size, leading to a sharp increase in cost and complexity for large-scale mass production. The industry urgently needs a new type of base that can simultaneously meet the requirements of ultra-large size, ultra-high density, and ultra-low deformation.
Glass Substrates Find the Optimal Balance Among "Large Size, High Density, High Stability, and High Speed."
Glass not only possesses excellent dimensional stability, but its coefficient of thermal expansion (CTE) can also be adjusted through formulation to highly match silicon and build-up dielectrics, effectively solving the warpage pain point in large-size packaging. Meanwhile, the high insulation and low dielectric loss of glass enable faster and lower-loss high-speed signal transmission. Coupled with its naturally transparent characteristics, it provides a unique and ideal carrier for the next-generation Co-Packaged Optics (CPO).
Glass Substrates Are Catalyzing a High-Prosperity Blue Ocean Market.
According to SEMI forecasts, the compound annual growth rate (CAGR) of global glass core substrates from 2028 to 2040 will reach as high as 67.2%; Yole predicts that by 2030, the total global market for advanced IC (Integrated Circuit) substrates will reach $31 billion. Faced with this scarce increment that far exceeds the growth rate of traditional organic substrates, whoever masters the core processes and capacity of glass substrates first will hold the key to future high-end computing power packaging.
Simply put, the rush of giants to deploy glass substrates is essentially driven by "rigid demand" in the computing power era: old materials cannot support new computing power, and glass provides a feasible solution. Whoever achieves technological and capacity breakthroughs first will hold the critical voice in the semiconductor industry in the post-Moore era.
02. How Far Are Glass Substrates from Industrialization?
Amidst the bustling news and capital frenzy, a sober question must be answered: exactly how far are glass substrates from true large-scale industrialization?
The industry has formed a basic consensus on the timeline: 2026 will be the window for industrial validation, 2027 will see the gradual release of orders, and 2028 will enter the volume ramp-up stage.
The judgment of Yu Zongguang, the rotating chairman of the Packaging and Testing Branch of the China Semiconductor Industry Association (CSIA), is quite representative: "This year is the 'critical validation period,' not the 'first year of comprehensive mass production.'"
The Timelines of Overseas Giants Are Very Clear:
From the perspective of the global industrial timeline, the industry has formed a relatively clear expectation of the pace: Intel is advancing the fastest, conducting small-batch trial production in 2026 and seeking scaled-up volume in 2027; TSMC's CoPoS panel-level packaging pilot line has been operational, targeting mass production in 2028; Samsung aims to achieve production in 2027.
Domestic Enterprises Are in the Final Sprint of the Pilot Production Stage:
BOE's 510×515mm large-panel pilot line has been fully automated and operational; Lens Technology has built a 30,000-square-meter dedicated factory and sent 22-layer glass core substrate samples to customers such as Intel; the Wuhan base of Woge Optoelectronics has completed the entire process from thinning, laser drilling to copper filling and electroplating, and its 1.6T optical module glass-based substrate has entered validation by leading customers.
Three Core Hurdles Lie on the Path to Industrialization:
First, TGV (Through Glass Via) process yield. TGV is the core process for glass substrates to achieve vertical electrical interconnection, and it is also the key differentiator from organic substrates. A single substrate requires the processing of millions of vias, and the single-via yield must be increased to over 99.999999% for the whole-panel yield to cross the commercialization threshold. Via metallization is particularly tricky: copper leads need to be filled in the glass vias, but the expansion coefficient of copper is much larger than that of glass. When the temperature rises, the expanding copper will crush the glass or damage the circuits. The industry is validating multiple solutions, including improving glass strength at the material level and setting up flexible buffer layers at the process level.
Second, shortfalls in the supporting industrial chain. In terms of high-end inspection equipment and special chemicals, domestic enterprises still need time to catch up. Equipment such as lasers, via inspection, and plating solution analysis used for glass substrate packaging mainly relies on imports, and raw materials are dominated by products from manufacturers such as Corning and SCHOTT. Localization is gradually advancing.
Third, cost and market validation. Currently, the cost of glass substrates is 30% to 50% higher than that of organic substrates. The core task from 2026 to 2027 is to transform laboratory parameters into stable mass-production process recipes, gradually approaching the commercialization cost line during the yield ramp-up.
In other words, the issue with glass substrates is not "whether to use them," but "who can first run through yield, cost, and capacity." 2026 is the year of validation, 2027 is the year of orders, and 2028 is the year of volume ramp-up—the endgame of this race depends on who can cross the chasm from "laboratory samples" to "factory mass production" first in the next two to three years.
Layout of the Glass Substrate Industrial Chain at Home and Abroad
03. Where Are the Opportunities for Hubei?
The largest application scenario for glass substrates is advanced packaging. For Hubei, deploying glass substrates is not an optional choice to add icing on the cake, but a mandatory question to build on the local advanced packaging industrial highland and connect the entire chain of "materials—packaging—computing power." Many regions in China first develop glass substrates and then search everywhere for downstream packaging customers; Hubei, on the contrary, already holds a domestically leading advanced packaging industrial foundation and naturally possesses the most core landing market for glass substrates. This is Hubei's most core differentiated advantage compared to other provinces.
First, Hubei Possesses a Domestically Scarce Vertically Integrated Industrial Foundation in the Glass Substrate Track
The "materials—equipment—manufacturing" TGV hard-tech foundation has taken shape, achieving multi-point breakthroughs on the supply side.
Upstream Materials: Gebijia has developed dedicated glass carrier boards for semiconductor packaging. Feilihua's technological accumulation in high-purity quartz glass and ultra-low expansion quartz glass provides a solid reserve for the extension of semiconductor-grade glass materials.
Core Equipment: HGTECH's TGV equipment has achieved a minimum hole diameter of 5 microns, a maximum aspect ratio of 1:100, and a through-hole rate of 99.9%; DR Laser has formed a one-stop solution of "laser modification + chemical etching + AOI (Automated Optical Inspection)." The complementary technological routes of these two Optics Valley enterprises constitute Hubei's global comparative advantage in the field of TGV hole-making equipment.
Manufacturing and Forming: Tonggewei (a subsidiary of Woge Optoelectronics) focuses on massive TGV hole-making and high-precision wiring, with processing capabilities advancing towards a minimum hole diameter of 3 microns; Xinchuangyuan, relying on its accumulation in BT and ABF substrates, is extending towards glass core substrates and fine metallization. From micro-hole processing to the shipment of complete carrier boards, Hubei has run through a small-batch closed loop.
Layout of Hubei's Glass Substrate Industrial Chain
Second, Sitting on a First-Tier Domestic Advanced Packaging Industrial Highland, Providing a "Testing Ground" at the Doorstep for Glass Substrates
Glass substrates cannot develop in isolation; their value can only be realized in advanced packaging scenarios. Hubei has built one of the few complete advanced packaging systems in China, covering scientific research—pilot production—industrialization, which is the most precious downstream resource in this track.
Jiangcheng Laboratory has built a nationally leading 12-inch advanced packaging comprehensive experimental platform, possessing full-stack capabilities in wafer-level vertical packaging, chiplet heterogeneous and hybrid packaging, and large-size carrier board packaging. It has completed the tape-out of 27 advanced packaging chips and can naturally carry out packaging validation and reliability testing for glass substrates, directly introducing glass raw sheets and TGV processed parts into the packaging platform for joint debugging. Its incubated Xingchen Technology has completed an over 4 billion RMB Series A financing, setting a record for the largest financing in the domestic advanced packaging field this year. The second-phase pilot line is about to be fully operational, capable of undertaking risk production of AI computing chips and optoelectronic integrated chips. In the future, it can directly introduce domestic glass substrates for landing validation, realizing the closed loop of "producing glass substrates locally and completing packaging landing locally."
Coupled with the storage and logic chip packaging demands brought by YMTC (Yangtze Memory Technologies Co.) and XMC (Wuhan Xinxin Semiconductor Manufacturing), Hubei possesses a large number of real chip customer scenarios. In many places, after glass substrate samples are made, they still have to travel far to find packaging and testing plants for certification; Hubei can achieve close coordination of materials, equipment, substrate processing, and packaging validation, significantly shortening the cycle of sample validation and iterative optimization. This is an extremely precious advantage in the industrial pilot production stage.
Third, the "Optics, Chips, Displays, Terminals, and Internet" Industrial Cluster Opens the Earliest Commercialization Landing Window for Glass Substrates
The ultimate landing of glass substrates requires massive application scenarios to force iteration. Optics Valley possesses cross-border synergy from panel giants such as BOE on glass substrates, the rigid demand for storage and logic chip packaging and testing brought by YMTC and XMC, and the frontier exploration in high-speed optical modules (CPO) by optical communication leaders such as Accelink Technologies. This natural market soil of "promoting packaging with optics, driving chips with displays, and validating materials with applications" is difficult for other regions in China to replicate.
04. Conclusion
Glass substrates are not meant to replace silicon, but to allow silicon's computing power to be released to a greater extent at the system level.
In many places, the deployment of glass substrates starts with making materials and then finding downstream markets. Hubei's unique endowment is that it already possesses a domestic advanced packaging industrial highland and holds the most core downstream application scenarios. Hubei's way out lies not in simple following, but in breaking through on a different track.
Based on Hubei's unique endowment in optoelectronic integration, making good use of advanced packaging as the biggest "industrial trump card," and connecting the local closed loop of "glass materials—TGV processing—advanced packaging—computing power applications" is the precious ticket in Hubei's hand.