Coralboard™ for Edge AI — Coming Soon!

Developer-ready edge AI board powered by Synaptics Astra™ compute and Google Coral NPU technology

Image of Synaptics Coralboard for Edge AISynaptics Coralboard is an edge AI development board manufactured by Grinn for Synaptics and developed in partnership with Google Research as the next generation of the Coral platform. Built around the Grinn AstraSOM-261x system-on-module (SoM) featuring the Synaptics Astra SL2619 edge AI processor and Torq™ AI engine with the industry’s first production implementation of the Coral NPU, Coralboard gives developers a practical starting point for rapid prototyping, product evaluation, and production-oriented edge AI development.

Coralboard is designed to accelerate development from an AI concept to a deployable intelligent device. With full NPU offload, developers can run AI workloads efficiently while freeing high-frequency CPU cores for application orchestration, agentic workloads, and device software stacks. An open-source edge AI compiler based on the industry-standard multi-level intermediate representation (MLIR) framework gives teams a more accessible path to model deployment, while the platform’s low-power design supports always-available, on-device ambient AI experiences.

Coralboard is intended for embedded developers, AI/ML engineers, system architects, product teams, and device manufacturers building the next generation of intelligent edge products. Teams can begin with Coralboard for evaluation and rapid prototyping, move to the Grinn AstraSOM-261x SoM for product prototyping and integration, and scale to full-production designs using the Synaptics Astra SL2619 edge AI processor. This flexible path helps customers explore real-world AI workloads, validate product concepts, and move from experimentation to production using a common edge AI technology foundation.

Features
  • Edge AI development board for evaluation, prototyping, and product development
  • Grinn AstraSOM-261x SoM built around the Synaptics Astra SL2619 edge AI processor
  • Torq AI engine with the industry’s first production implementation of the Coral NPU
  • Full NPU offload helps free high-frequency CPU cores for agentic orchestration and application software stacks
  • Open-source edge AI compiler based on the industry-standard MLIR framework
  • Designed to support low-power, always-available on-device ambient AI experiences
  • Integrated hardware and software platform for accelerating development from AI concept to working prototype
  • Scalable product path: Coralboard for experimentation, Grinn AstraSOM-261x SoM for prototyping, and Synaptics Astra SL2619 processor for production deployment
  • Hardware expansion support for camera, display, audio, sensor, USB, wireless, and embedded I/O development
  • Linux-based software development support with Synaptics Astra SDK resources
  • Suitable for on-device inference and local data processing at the edge
Applications
  • AI-enabled vision systems
  • Audio and voice AI devices
  • Sensor-driven edge AI products
  • Ambient AI and always-available intelligent devices
  • Smart home, appliances, and consumer devices
  • Industrial automation, process control, and machine vision
  • Smart retail, kiosks, and connected commercial devices
  • Healthcare device prototyping and diagnostic support systems
  • AI-enhanced gaming and interactive experiences
  • Robotics and commercial UAVs for industrial automation, inspection, and logistics
  • Human-machine interface and user-experience prototyping
  • Connected embedded products with local AI inference
  • Multimodal AI application development
  • Edge AI model and workload evaluation
Published: 2026-07-06