BIOSTAR has shared more details about its new compact edge AI platform EdgeComp MS-NAT5000, which is based around the NVIDIA Jetson Thor T5000 module, designed for system integrators, AI developers, smart manufacturing providers, and autonomous machine builders looking to deploy advanced AI workloads closer to where data is generated.
The unit is capable of delivering up to 2,070 TFLOPS of FP4 AI performance thanks to the NVIDIA Blackwell hardware paired with 128GB of LPDDR5X memory and a 14-core Arm CPU that forms a strong foundation capable of handling large language models (LLMs), vision-language models (VLMs), multimodal AI applications, robotics workloads, and agentic AI systems directly at the edge.
Meanwhile, the QSFP28 100GbE interface is both fast and flexible, since it can be broken out into 4x 25GbE connections, alongside dual 5GbE LAN ports. Integration with NVIDIA Holoscan Sensor Bridge technology enables high-speed sensor-over-Ethernet connectivity for cameras and other data sources, supporting low-latency inspection and autonomous operation workflows.
Other than that, one can find an M.2 Key-M slot for NVMe storage, an M.2 Key-B slot for 4G and 5G modules, and an M.2 Key-E slot for Wi-Fi and Bluetooth connectivity. Industrial communication requirements are addressed through dual RS232/422/485 ports and four CAN FD interfaces, allowing direct integration with sensors, controllers, and automation equipment.
Additional features include HDMI and DisplayPort 1.4a outputs, optional dual SATA III hot-swap SSD support, and a wide 12V to 36V DC power input range suitable for industrial deployments.
Instead of fully relying on cloud services, local alternatives always prioritize data control, while latency and responsiveness is dictated by user requirements. BIOSTAR positions the system for industrial use cases including visual inspection, production-line image analysis, predictive maintenance, AI-assisted operations, human-machine collaboration, and on-premises generative AI deployments.
The platform also targets physical AI and autonomous systems, supporting workloads such as perception, navigation, sensor fusion, and intelligent control. Potential deployment scenarios include autonomous mobile robots, warehouse automation systems, unmanned delivery vehicles, autonomous public transportation, agricultural equipment, and construction machinery.
