ADLINKジャパン Product Lineup

The core of the DLAP-701 is equipped with the cutting-edge NVIDIA Blackwell architecture. With the NVIDIA Jetson T5000/Jets T4000 modules, it achieves an astonishing maximum of 2070 FP4 TFLOPS of extreme AI performance. By combining a powerful ARM Neoverse-V3AE CPU (up to 14 cores) with 128GB of LPDDR5X memory, it processes complex AI workloads at unprecedented speed and low latency. Connectivity is extremely important in modern autonomous systems, and the DLAP-701 excels with its rich high-speed I/O. In addition to large data throughput supported by QSFP ports with 4x 25GbE, it features dual Gigabit Ethernet, multiple USB 3.2 Gen 2 ports, and comprehensive M.2 expansion slots for Wi-Fi 6 and 5G connectivity. The inclusion of an isolated CAN-FD interface further enhances reliability in vehicle and robot control.

DLAP-IGX is a high-performance edge AI platform specifically designed for demanding applications such as humanoid robots, autonomous mobile robots (AMR), and advanced vision sensing systems. It features the NVIDIA IGX Thor architecture, achieving an astonishing 4,293 TFLOPS of AI performance. This industry-grade system is equipped with a 14-core Arm CPU and NVIDIA Blackwell GPU architecture, along with 128 GB of LPDDR5X memory. With dual 200GbE networking and integrated safety support, it provides the overwhelming throughput and reliability required for complex multi-model generative AI workloads.

ADLINK is a global leader in the fields of industrial PCs and edge computing. This guide introduces the latest edge AI platforms, industrial PCs, servers, COM modules, vision systems, and display products that support digital transformation (DX) across a wide range of industries, including manufacturing, transportation, healthcare, and retail. By collaborating with major technology partners such as Intel, NVIDIA, AMD, and NXP, we provide high-performance and highly reliable solutions. Through case studies in AI visual inspection, railway communications, medical devices, and public infrastructure, you can explore our comprehensive product lineup that enables practical applications of edge AI.

Compact AI Acceleration for Embedded Systems with Space Constraints The EGX-MXM-BW500 is an ultra-compact MXM 3.1 Type A module that delivers up to 294 AI TOPS of AI performance. It is optimized for embedded platforms with limited space, providing efficient AI inference, computer vision, and graphics acceleration in intelligent edge devices where compact size, thermal efficiency, and reliable performance are essential.

The highest computational density in MXM Type A modules The EGX-MXM-BW2000, achieving up to 798 AI TOPS, maximizes AI computational density within the compact MXM 3.1 Type A form factor. It is designed for embedded AI systems that require significantly higher inference performance without increasing the size of the module, making it ideal for robotics, machine vision, and medical imaging applications.

High-performance AI and graphics for advanced edge systems The EGX-MXM-BW4000 combines powerful AI acceleration and advanced graphics performance in an MXM 3.1 Type B module, achieving up to 1,334 AI TOPS. This product, enhanced with GPU resources and memory bandwidth, is designed for demanding applications in computer vision, industrial AI, medical imaging, and real-time visualization.

Ultimate Embedded GPU for High-Performance Edge AI The EGX-MXM-BW5000, achieving up to 1,824 AI TOPS, is the flagship model of this series. Equipped with up to 24GB of high-speed GDDR7 memory, it delivers exceptional AI inference and graphics performance for the most demanding embedded AI workloads. It is optimized for large language models (LLM), generative AI, multi-stream AI inference, and other compute-intensive edge AI applications.

The ADLINK VPX3-XMXM series is a 3U VPX GPGPU (General-Purpose computing on Graphics Processing Units) blade designed to provide high-performance graphics and processing capabilities. Its primary application is to enable robust AI inference compliant with SOSA standards, making it a reliable solution specialized for high-load payload applications that require extensive real-time data processing in physically demanding industrial environments.