BRINGING BLACKWELL-POWERED EDGE AI TO SMART MANUFACTURING AND VISION AI
June 23rd, 2026 , TAIPEI, TAIWAN – BIOSTAR, a leading manufacturer of edge AI solutions, industrial motherboards, motherboards, graphics cards, and storage devices, today announces the EdgeComp MS-NAT5000, a compact edge AI computing system that integrates the NVIDIA Jetson Thor T5000 module to bring generative AI, visual intelligence, and real-time sensor processing directly to industrial and autonomous applications.
Built for system integrators, AI software developers, smart manufacturing solution providers, and advanced autonomous machine developers, the EdgeComp MS-NAT5000 is engineered to support the next stage of edge AI deployment, where AI systems must process data locally, make decisions with low latency, and turn intelligence into real-world action.
By bringing NVIDIA Blackwell architecture to demanding edge AI workloads, the EdgeComp MS-NAT5000 delivers up to 2,070 TFLOPS of FP4 AI performance, paired with 128GB of LPDDR5X memory and a 14-core Arm CPU, the platform provides the computing foundation required for LLM, VLM, VSS, robotics, multimodal AI, and agentic AI-driven applications at the edge.
This high-performance architecture enables advanced AI workloads to run locally, helping reduce reliance on cloud processing while improving responsiveness, data control, and deployment flexibility. For industrial environments, the EdgeComp MS-NAT5000 supports applications such as real-time visual inspection, production-line image analysis, predictive maintenance, AI-assisted operation, human-machine collaboration, and on-premises generative AI services.
For physical AI and autonomous machine deployments, the EdgeComp MS-NAT5000 provides the performance and connectivity required for perception, navigation, sensor fusion, and intelligent control. From autonomous mobile robots and warehouse logistics systems to unmanned delivery vehicles, autonomous buses, agricultural machinery, and construction automation, the system is designed to support AI applications operating in complex real-world environments.
To support high-bandwidth sensor and vision workloads, the EdgeComp MS-NAT5000 features a 1 x QSFP28 100G (breakout to 4 x 25GbE links) interface and 2 x 5GbE LAN ports. Paired with NVIDIA Holoscan Sensor Bridge technology, the platform enables high-speed sensor-over-Ethernet streaming for cameras and connected data sources, supporting low-latency perception, inspection, and autonomous control workflows.
The EdgeComp MS-NAT5000 is also designed with flexible expansion and industrial connectivity for system-level integration. The platform includes 1 x M.2 Key-M 2242/2260/2280 slot for NVMe SSD storage, 1 x M.2 Key-B slot for 4G/5G modules, and 1 x M.2 Key-E slot for WiFi and Bluetooth connectivity. For industrial communication, it provides 2 x RS232/422/485 ports and 4 x CAN FD interfaces, enabling integration with controllers, sensors, machines, and automation systems.
In addition to HDMI and DP1.4a ports, its deployment flexibility is supported through an optional 2 x SATA III with SSD hot-swap tray and 12 to 36V DC input, making the EdgeComp MS-NAT5000 suitable for a wide range of industrial edge environments. The system is also designed to handle various AI workloads, such as navigation, vision processing, and language-based interaction, in parallel with improved workload isolation.
With the EdgeComp MS-NAT5000, BIOSTAR expands its EdgeComp portfolio with a compact, high-performance platform for next-generation industrial AI. By combining NVIDIA Jetson Thor T5000 computing power, on-device generative AI capability, real-time sensor connectivity, and industrial-grade expansion, the EdgeComp MS-NAT5000 provides system integrators and AI developers with a practical foundation for building intelligent edge solutions, where AI moves beyond analysis and into action.
Product Page: EdgeComp MS-NAT5000 Product Page
Video: BIOSTAR 40 Anniversary
