Cloth-free construction represents a pivotal step in the evolution of high-speed copper interconnects, yet it does not entail removing all glass fiber fabric from an entire PCB. Instead, based on performance and reliability constraints, it involves reducing or eliminating fabric in specific dielectric layers. The cloth-free approach is mainly divided into two application scenarios: computing power PCBs and optical modules. The shared objective is to reduce loss and enhance signal integrity, while the difference lies in the entry point and material form. The Rubin Ultra (VR300) NVL576 rack switch board is the most likely scenario to adopt cloth-free or low-fabric solutions at scale first.
We expect the NVL576 rack switch board to form four combination schemes pairing PTFE/hydrocarbon cloth-free CCL with M9 Q-fabric PP or hydrocarbon cloth-free PP respectively, with the fully cloth-free scheme expected to land first. The Q-fabric-containing scheme would serve as a backup when processing or supply reliability proves insufficient. Driven by NVL576 rack shipments, cloth-free PCBs, CCL/PP, and upstream materials are expected to ramp up starting in 2028. The cloth-free trend concentrates industrial chain value toward the CCL/PP segment, warranting priority attention on CCL/PP leaders while also highlighting incremental opportunities in PCB processing and upstream materials.
Cloth-free construction is a critical step in the high-speed copper interconnect upgrade, but it does not mean removing all glass fiber fabric from an entire PCB. Rather, based on performance and reliability constraints, low-fabric or cloth-free configurations are implemented in specific dielectric layers. As AI rack interconnect speeds continue to rise, the differing dielectric constants of glass fiber fabric and resin give rise to the glass fiber effect, and traditional fabric-containing structures are gradually approaching bottlenecks in dielectric uniformity, loss, and signal integrity. Removing localized glass fiber fabric can reduce dielectric loss and support thinner dielectrics, laser microvias, and finer circuitry.
The material forms of cloth-free configurations can be categorized into cloth-free CCL, cloth-free PP/resin film, and RCC: cloth-free CCL is used for core boards, cloth-free PP for interlayer insulating bonding, and RCC/resin film for build-up layers outside the core. As high-speed CCL grades advance to M10 and above, the industry shift is not merely about being cloth-free or low-fabric, but rather about resin, spherical silica fillers, and high-grade copper foil jointly assuming dielectric, reinforcement, and processing functions. Incremental value concentrates toward the CCL/PP manufacturing segment, and Q fabric is expected to remain in low-fabric schemes at M10 and above.
Cloth-free applications fall into two scenarios: computing power PCBs and optical modules. The common goal is to reduce loss and improve signal integrity, while the difference lies in the entry point and material form. Computing power PCBs adopt cloth-free CCL/PP, penetrating from signal layers into the core board skeleton, targeting high-layer-count, large-size switch boards. We expect the Rubin Ultra (VR300) NVL576 to currently form four combinations pairing PTFE/hydrocarbon cloth-free CCL with M9 Q-fabric PP or hydrocarbon cloth-free PP respectively. The fully cloth-free scheme combines high-frequency performance with material economics and is expected to land first, while the Q-fabric-containing scheme serves as a backup when processing or supply reliability is insufficient. The Core layer has a greater impact on transmission performance than PP, so the Core is expected to upgrade first before gradually promoting cloth-free PP. Computing boards are subject to packaging and dimensional stability constraints, making adoption more difficult, so the Rubin Ultra (VR300) NVL576 switch board is the most likely scenario to adopt cloth-free or low-fabric solutions at scale first.
Optical modules, meanwhile, adopt RCC/cloth-free resin film focused on build-up layers outside the core, using thin dielectrics and mSAP to achieve microvias and fine circuitry. mSAP and carrier copper foil applications have already emerged at 1.6T. Zhongji Innolight's licensed patent explicitly specifies an RCC plus mSAP solution. Sihui Fuji's 800G RCC boards have entered mass production, and its 1.6T samples have obtained end-customer certification. Hongchang Electronics, Xunjiexing, Nuopei Core Materials, and Zecheng Electronics are continuously advancing material certification or R&D, with industrialization visibility gradually improving.
Driven by NVL576, cloth-free PCBs, CCL/PP, and upstream materials are expected to enter a ramp-up period starting in 2028. We estimate the cloth-free PCB market space for 2027E/2028E/2029E at approximately RMB 10.554 billion / RMB 112.456 billion / RMB 160 billion, and cloth-free CCL at approximately RMB 1.236 billion / RMB 11.626 billion / RMB 21 billion, respectively. Supporting PP is estimated at RMB 1.086 billion / RMB 13.115 billion for 2027E/2028E, with CCL and PP combined at RMB 2.322 billion / RMB 24.741 billion. The upstream materials market space for 2027E/2028E is estimated at RMB 584 million / RMB 5.833 billion, of which hydrocarbon resin accounts for RMB 2.741 billion in 2028, roughly 47%, making it the highest-value upstream material.
Investment recommendation: Prioritize CCL/PP leaders while also emphasizing incremental opportunities in PCB processing and upstream materials. As high-speed copper interconnects upgrade toward low-fabric/cloth-free configurations, CCL/PP are expected to become core beneficiaries. At the same time, increased PCB processing difficulty and growing demand for upstream materials such as resin and spherical silica are expected to bring new opportunities.
Risk of next-generation platform material路线 and cloth-free penetration falling short of expectations: The choice between cloth-free and fabric-containing schemes still requires balancing transmission performance, processing reliability, and cost. If the performance improvement of fabric-containing schemes such as Q fabric exceeds expectations, or if fully cloth-free schemes fail to meet requirements in dimensional stability, processability, and supply assurance, end customers may continue to adopt fabric-containing or low-fabric structures, delaying the cloth-free transition of Core and PP layers and causing the application layer count and penetration rate of cloth-free materials to fall below expectations.
Risk of processing yield and customer certification progress falling short of expectations: After reducing or removing glass fiber fabric, materials must meet dielectric performance, mechanical support, and processing reliability requirements through resin, filler, and structural design. If cloth-free CCL/PP underperforms in lamination, dimensional stability, and interlayer bonding in high-layer-count, large-size PCBs, or if optical module RCC/resin film encounters bottlenecks in microvia, fine circuitry, and long-term reliability verification, customer certification cycles may be extended, affecting volume introduction and delivery of related products.
Risk of downstream platform iteration and industrialization ramp-up falling short of expectations: Demand release for cloth-free solutions is closely tied to AI rack interconnect upgrades, switch board material selection, and high-speed optical module iteration节奏. If downstream capital expenditure slows, next-generation platform launches or mass production are delayed, or key material supply, equipment matching, and mass production yields fail to improve in tandem, scaled application of related products may occur later than expected, thereby affecting the ramp-up pace from 2028 projected in the report and the performance realization of industrial chain companies.
Comments