CITIC SEC: Copper Interconnect Approaching Physical Limits; NPO/CPO Drives Revaluation of Passive Components

CITIC SEC released a research report stating that Scale-up opens up an incremental pool for passive components. Passive components are leaping from low-value accessories in the pluggable era to core BOM items that affect system insertion loss and maintainability. As a key link in optical engine coupling, FAU stands out with prominent elasticity in both volume and price. Domestic manufacturers with large-scale manufacturing capabilities for high-precision products and certification barriers with leading customers are expected to be the first to realize performance elasticity. CITIC SEC's main views are as follows: Copper interconnect is approaching physical limits, and NPO/CPO is driving massive inflation in passive components. Copper interconnect is approaching physical boundaries in the 400G/lane era. Optical interconnect is gradually penetrating from Scale-out scenarios to Scale-up scenarios. NVIDIA's Rubin Ultra NVL576 is expected to adopt a hybrid architecture of "copper within the rack + optics across racks," marking the transition of Scale-up optical interconnect from concept to implementation. Pluggable solutions are difficult to support rack-level high-channel full interconnect, and NPO/CPO will undertake the massive bandwidth for inter-rack interconnect, with NPO as a pragmatic transitional solution expected to ramp up first. NPO/CPO share optical engines, external light sources, and passive components such as FAU (Fiber Array Unit), MPO/MMC (fiber connectors), and Shuffle (fiber flexible boards). Passive component usage scales synchronously with the number of engines, and the importance of passive components in the overall system is gradually increasing. For passive component manufacturers, the Scale-out pluggable upgrade trend continues, while Scale-up ushers in entirely new incremental demand from density upgrades, with promising performance elasticity. According to Frost & Sullivan forecasts (as cited in the Hong Kong IPO prospectus of Aidete), the total market size of fiber connectors, fiber array units, and fiber flexible boards in CPO alone will exceed 100 billion yuan by 2030. FAU stands out with prominent elasticity in this round of passive component inflation: usage has a rigid ratio with optical engines, and high-spec premiums drive value leaps. On the usage side, the Scale-up level injects new incremental demand into the FAU market. On the switch chip side, FAU forms a 1:1 rigid ratio with optical engines, while the ELS side simultaneously drives demand for polarization-maintaining MT-FA. On the Scale-out level, it also benefits from the high growth in optical module shipments and the density upgrade of NPO/CPO. On the price increase side, in the NPO/CPO era, FAU channel counts are leaping from 8/12/16 cores to 32/36/64 cores and higher. Higher channel counts increase fiber arrangement and polishing difficulty, polarization-maintaining FAU adds process and testing steps, fiber length and category upgrades, and pluggable FAU raises design and process thresholds. These fourfold high-spec premiums significantly boost the value per FAU unit. Domestic passive component manufacturers participate in the midstream manufacturing of FAU through OEM or cooperation models. The current core bottleneck in the industry lies in channel consistency and yield ramp-up under high channel counts, with overall capacity being tight. Meanwhile, North American leading cabling vendors and optical communication manufacturers have long certification cycles and stringent reliability requirements. Passive component manufacturers that are bound to core customers and possess large-scale, high-precision product delivery capabilities are expected to benefit deeply from this round of industrial upgrading. Risk factors: AI development falling short of expectations; cloud capital expenditure falling short of expectations; Scale-up optical penetration falling short of expectations; NPO/CPO commercialization progress falling short of expectations; yield ramp-up falling short of expectations; technology path switching risk; intensifying competition and price decline risk; geopolitical and supply chain disruption risks.