Taipei, Taiwan – March 12, 2026 - ADLINK Technology Inc., a global leader in edge computing, has launched the ADLINK MXM family of embedded graphics and AI accelerator modules based on NVIDIA Blackwell architecture. Offered in industry-standard MXM 3.1 Type A/B form factors, the new modules deliver scalable performance and energy efficiency for next-generation AI, robotics, and edge computing applications. They are designed to power intelligent systems that operate locally across a broad range of use cases, including robots and industrial machines, smart cameras and inspection systems, autonomous vehicles, retail infrastructure, and medical imaging and healthcare devices.
To address the wide range of deployment environments seen across various AI application demands, ADLINK’s MXM Modules powered by NVIDIA RTX PROTM Blackwell Generation Embedded GPUs deliver exceptional flexibility in power and performance configuration. With support for power budgets ranging from 45W to 150W, the platform enables system designers to precisely tailor performance profiles based on application priorities, whether maximizing compute capability or operating within strict thermal and energy constraints.
ADLINK’s EGX-MXM-BW2000, for example, sets a new benchmark for compact edge AI performance by delivering up to 100W of performance in a MXM 3.1 Type A form factor (82x70mm), maximizing compute density for space-constrained embedded systems. At the high end, the EGX-MXM-BW5000 features NVIDIA Multi-Instance GPU (MIG) technology that allows developers to partition GPU resources for parallel workloads, improved isolation, and more efficient resource management in advanced edge AI deployments. This robust combination enables efficient execution of large language models (LLMs), deep learning pipelines, and other data-intensive workloads directly at the edge.
“Having worked closely with customers for many years across diverse embedded applications—from medical imaging to industrial automation—we have witnessed a major shift in how AI is deployed at the edge,” says Edgar Chen, Managing Director of the Edge Visualization Business Unit at ADLINK. “As AI adoption accelerates, our customers demand higher GPU performance, larger memory capacity, and uncompromising reliability. With our MXM modules based on NVIDIA Blackwell architecture, we deliver a powerful combination of performance, efficiency, and flexibility, empowering system designers to confidently build next-generation edge AI solutions tailored to their environments.”
Built on NVIDIA Blackwell, the new MXM modules deliver breakthrough AI inference performance with FP4 precision, achieving up to 2x throughput while reducing memory bandwidth demands. The latest NVIDIA CUDA Cores, 4th-gen RT Cores, and 5th-gen Tensor Cores provide powerful GPU acceleration for AI inference, computer vision, and real-time data analytics, with peak FP32 performance ranging from 9.2 to 49.8 TFLOPS. All models include next-generation GDDR7 memory, providing high bandwidth and low latency for consistent, high-efficiency processing of complex, data-intensive workloads at the edge.
Depending on the configuration, the GDDR7 memory subsystem supports up to 24GB, memory interfaces up to 256-bit, and bandwidth reaching up to 896GB/s, significantly boosting throughput by while minimizing latency for demanding applications such as deep learning, high-resolution imaging, and advanced simulation.
All modules support up to four DisplayPort 1.4a outputs, enabling multi-display visualization for industrial, transportation, and autonomous systems. With 5-year availability and wide operating temperature ranges (Standard: 0˚C to 55˚C, Extended: -40˚C to 85˚C), the platform ensures long-term reliability, lifecycle stability, and dependable operation across both commercial and harsh industrial environments.
Product availability includes:
- EGX-MXM-BW500 - NVIDIA RTX PRO 500 Blackwell Generation Embedded GPU, MXM 3.1 Type A, 82 x 70mm
- EGX-MXM-BW2000 - NVIDIA RTX PRO 2000 Blackwell Generation Embedded GPU, MXM 3.1 Type A, 82 x 70mm
- EGX-MXM-BW4000 - NVIDIA RTX PRO 4000 Blackwell Generation Embedded GPU, MXM 3.1 Type B, 82 x 105mm
- EGX-MXM-BW5000 - NVIDIA RTX PRO 5000 Blackwell Generation Embedded GPU, MXM 3.1 Type B, 82 x 105mm