Edge AI processes data on or near the device that collects it — reducing cloud dependence, latency, and bandwidth.
Edge AI is AI that runs at the edge of the network — on the device or a nearby gateway — instead of in a remote cloud. AWS defines edge ML as running models where data is generated to cut latency and bandwidth costs.
Think factory sensors scoring defects on the line, a phone transcribing speech without uploading audio, or a security camera detecting motion locally. The model runs close to where data is created.
Compared to cloud AI, edge AI trades unlimited compute for speed, privacy, and reliability when connectivity is poor. NVIDIA's edge computing documentation emphasizes local inference for latency-sensitive workloads. It overlaps heavily with localized AI and on-device AI; edge often implies IoT and enterprise deployments, while localized AI often means personal devices.