TLB said it is participating in KPCA Show 2026 at Songdo Convensia in Incheon, marking the first time the company has set up an independent booth at the exhibition. The company showcased printed circuit boards (PCBs) for Small Outline Compression Attached Memory Module (SOCAMM), Low Power Compression Attached Memory Module (LPCAMM), and Compute Express Link (CXL) memory modules, signaling a strategic expansion beyond its core memory module business into high-performance computing (HPC) and artificial intelligence (AI) server infrastructure.
8,600 MT/s signal integrity
A key technical highlight of the display was a PCB designed for high-speed memory modules supporting data rates of 8,600 MT/s, or 8.6 billion data transfers per second. As memory interface speeds climb, the physical constraints of signal integrity become a primary bottleneck for system performance. TLB emphasized that PCB design technologies capable of minimizing signal loss and interference are now critical to realizing the theoretical bandwidth of next-generation memory standards in dense AI server configurations.
SOCAMM and CXL module substrates
The simultaneous promotion of SOCAMM, LPCAMM, and CXL PCBs underscores the industry’s rapid shift toward higher-density, higher-speed computing architectures. These standards represent the evolution of memory integration, allowing for greater bandwidth and modularity within AI and HPC systems. By providing the physical substrate for these advanced modules, TLB is positioning itself to address the growing demand for high-bandwidth memory in modern AI infrastructure, where data movement efficiency is increasingly viewed as a core determinant of system throughput (SK hynix).
EDSFF SSD form factor support
In addition to memory solutions, TLB demonstrated its capability in data center storage by displaying PCBs based on the Enterprise and Data Center Standard Form Factor (EDSFF) for solid-state drives (SSDs). EDSFF is a next-generation SSD form factor designed specifically for data center applications, confirming the industry’s move toward standardized, high-capacity storage solutions optimized for enterprise environments. This expansion into storage PCBs aligns with broader trends where storage and memory tiers are increasingly integrated to support the massive data requirements of AI workloads.
Data movement as the AI bottleneck
The context for these advancements is a market where the bottleneck for AI performance is shifting from raw compute power to data movement. As AI workloads evolve toward inference and agentic AI, the frequency of data exchanges between memory and compute devices increases, making the physical architecture of memory and storage critical. TLB’s focus on high-speed interconnects and standardized form factors addresses this shift by enabling the physical realization of performance gains in memory and storage components, which is essential as system bottlenecks move toward data flow efficiency.
KPCA Show awards and expansion
TLB received the Best Booth Design Award, KPCA Education Award, and Rising Partner Award at the exhibition, recognizing its efforts to advance the PCB industry and strengthen future competitiveness. A TLB official stated that the exhibition allowed the company to showcase its next-generation products directly to customers and industry officials, with plans to expand new business opportunities in AI, data centers, and next-generation memory markets. This marks a significant step for TLB as it seeks to establish a broader footprint in the high-performance computing supply chain.