Analog Tech has introduced a compact edge AI computer designed to run machine learning workloads without relying on cloud connectivity. The device is designed for industrial and commercial environments that require on-site data processing due to network limitations. By running analysis directly on the hardware, the system reduces reliance on cloud connections for time-sensitive operations.
Compact industrial computer pairs Intel Core Ultra processor with dedicated accelerator card for on-site machine learning workloads
The new hardware carries the model designation AironiA AIR-AD-AI-001 and occupies a 5-liter chassis volume. It pairs an Intel Core Ultra 5 processor with a dedicated PCIe dNPU accelerator card built around the AxeleraAI Metis AIPU architecture. The dedicated accelerator handles heavy image processing, leaving the central processor free for general computing duties.
The AI accelerator delivers 214 TOPS of INT8 performance using a Samsung 5nm process node. Operating at 8 to 15 watts allows the unit to run cool enough for its compact chassis design. It supports up to 24 simultaneous video streams processed through YOLOv5 object detection models.
The platform includes 32GB of DDR5 SO-DIMM memory and a 512GB M.2 NVMe SSD for storage. Network connectivity relies on dual wired ports offering both 2.5GbE and 1GbE speeds. Multiple video ports enable the use of several screens, which is useful for monitoring stations on production floors.
The manufacturer markets the device specifically for dedicated AI processing tasks instead of everyday computing needs. Engineering choices focus on consistent, reliable operation for image analysis rather than maximizing overall speed.
Discussion
0 comments
Log in to join the thread with a thoughtful take, question, or correction.