Google unveils the Coral Dev Board Mini, a smaller, budget alternative to the Coral Dev Board with the Edge TPU, and the Accelerator Module, available H1 2020
Last March, Google took the wraps off of Coral, a collection of hardware kits and accessories intended to bolster AI development at the edge.
Context & Ripple Effects
Google's Coral line began with the Edge TPU chip announced alongside Cloud IoT Edge in 2018, then became a retail hardware program in the March 2019 launch of a $149.99 dev board and $74.99 USB accelerator. The Mini and the Accelerator Module complete that arc: the same Edge TPU now ships in a smaller, cheaper board for prototypers and as a module OEMs can design directly into products.
The move matters because it turns Coral from a developer kit into a supply-chain component — Google is no longer just selling boards to hobbyists but offering embedding-grade silicon to device makers building at the edge.
First-order effects
- Developers priced out of the original Coral Dev Board get a lower-cost on-ramp to Edge TPU prototyping, while hardware makers gain the Accelerator Module as a drop-in part instead of designing around a bare chip.
Second-order effects
- Rivals in edge AI boards face pressure at the low end of the market — a segment Nvidia later attacked directly with the $249 Jetson Orin Nano Super Developer Kit, showing how contested compact AI dev hardware has become.
Third-order effects
- If the module-plus-board pattern holds, edge AI silicon consolidates into tiered product families where the same accelerator spans prototype and production — the structure behind Google's continued custom-silicon push, including reported talks with Marvell on a memory processing unit working alongside TPUs (The Information's report on Google-Marvell TPU talks).
The trend: Edge AI hardware is shifting from standalone developer boards toward tiered lineups of cheap boards plus embeddable accelerator modules, pulling machine learning inference off the cloud and into devices.