Specifications
Vendor
Google Coral
Model
Coral USB Accelerator
Part Number
G950-06809-01
Coprocessor
Google Edge TPU
Performance
4 TOPS (INT8); approx. 2 TOPS / W
Interface
USB 3.0 Type-C (USB 3.1 Gen1, data + power)
Power
approx. 2 W typical
Host Support
Linux (Debian / Ubuntu / Raspberry Pi), macOS, Windows
Software
TensorFlow Lite, Coral runtime / examples
Applications
Edge inference, prototyping, retrofit AI to existing hosts
Overview
The Coral USB Accelerator adds Google's Edge TPU coprocessor to a host computer through a single USB 3.0 Type-C connection. The Edge TPU delivers up to 4 TOPS of 8-bit integer inference at roughly 2 TOPS per watt, letting small single-board computers, laptops and embedded hosts run machine-learning models locally.
It supports TensorFlow Lite models and the Coral runtime on Linux (including Raspberry Pi and Debian/Ubuntu), macOS and Windows, making it a quick way to add accelerated, low-power edge inference to an existing system without a hardware redesign. The compact stick form factor makes it easy to deploy and prototype.
Key Benefits
Google Edge TPU · up to 4 TOPS INT8 · approx. 2 TOPS/W · USB 3.0 Type-C · Linux / macOS / Windows · TensorFlow Lite.
Applications
Edge and on-device inference, rapid AI prototyping, retrofitting existing hosts (SBCs, laptops, industrial PCs) with accelerated inference, and computer-vision projects such as classification and detection.
Request a Quote — Google Coral USB Accelerator (Edge TPU)
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →