
AI Edge Hardware
AI-Box Edge AI Hardware
Model AI-Box
AI-Box pushes AI to the shop-floor edge for quality inspection and predictive maintenance.
- NPU2~4 TOPS compute
- InterfacesEthernet/Serial/USB/CAM
- InferenceOn-device models
- Latency<50ms
SPEC
Parameter Table
NPU2~4 TOPS compute
InterfacesEthernet/Serial/USB/CAM
InferenceOn-device models
Latency<50ms
DeploymentOn-premise
FAQ
Product FAQ
Can it run models locally?
Supports on-device model inference — data stays on-site, low latency.
How does it work with gateways?
Gateways collect; AI-Box does edge intelligence analysis, forming a smart production line.
Which frameworks are supported?
Supports models exported from mainstream inference frameworks.
Does it need internet?
Fully offline on-premise deployment is possible.