← All hardware
Plug-and-play Edge TPU inference over USB for $59.99.
Pros
- Dead-simple USB integration with any host
- Very low power and silent operation
- Cheap and well-documented for vision tasks
Cons
- Locked to TensorFlow Lite / Edge TPU compiled models
- Only 4 TOPS, eclipsed by newer accelerators
- Aging platform with limited recent updates
✓ Where it shines / best for
- Adding efficient vision inference to a Raspberry Pi or PC
- Low-power smart camera and IoT prototypes
- Education and quick TensorFlow Lite deployments
✕ Not the best fit for
- Generative AI or LLM workloads
- Non-TensorFlow/PyTorch-native pipelines
- High-TOPS robotics applications
Features
- ✓ Google Edge TPU ASIC coprocessor
- ✓ 4 INT8 TOPS at ~2W (2 TOPS/W)
- ✓ USB 3.0 (USB-C) plug-and-play interface
- ✓ Runs TensorFlow Lite models (MobileNet v2 at 100+ fps)
- ✓ Works with Linux, macOS and Windows hosts
- ✓ Compatible with Raspberry Pi and other SBCs
- ✓ Compact, fanless, portable form factor
- ✓ Coral compiler and pretrained model library
Pricing
| Plan | Price | Billing | Notes |
|---|---|---|---|
| Coral USB Accelerator | $59.99 | one-time | Current MSRP (reduced from $74.99); sold via Seeed, Mouser, Newegg, RS. Dev Board and M.2/PCIe Edge TPU variants sold separately |
Pricing verified from the official source. Prices change often — confirm on the vendor's site before buying.
Specifications
| power | ~2W (2 TOPS/W) |
| memory | Host-dependent (no large on-board DRAM) |
| interface | USB 3.0 (USB-C) |
| architecture | Google Edge TPU ASIC |
| ai_performance | 4 INT8 TOPS |
Sponsored
A full review is being generated for this product and will appear here shortly.