Menu

Introducing ML
Custom SoC Platform for Remote Sensor Applications

Overview

chipIgnite ML is a groundbreaking platform designed to bring the power of custom machine learning (ML) chips to remote sensor IoT applications. Delivering an optimized system-on-chip (SoC), it addresses the limitations of current general-purpose solutions, enabling product developers to create highly efficient, battery-powered ML devices at the edge.

The Challenge

Remote sensor products today rely on general-purpose chips that consume too much power and are too bulky for many applications. The market is dominated by top-tier companies using off-the-shelf chips, limiting the possibilities for smaller players to innovate. Long-tail applications—where specialized, low-volume products are needed—struggle due to the high costs and technical barriers involved in creating custom chips.

The Solution: chipIgnite ML

chipIgnite ML offers a customizable platform that delivers the following benefits:

  • Longer Battery Life:
    Achieves up to 10x longer battery life, enabling devices to run on a single coin cell battery for years.
  • Smaller Form Factor:
    Designed for compact applications, with the size of a key fob, versus current products that require multiple AA or AAA batteries.
  • Tailored to Specific Needs:
    Fully customizable SoC that integrates analog interfaces and accelerates ML-based event detection for low-data-rate signals.

Key Features

RISC-V Based CPU

Utilizes a 32-bit core for efficient processing.

Advanced ML Acceleration

Custom logic for event detection and ML acceleration tailored to specific applications.

Ultra-Low Power Design:

Features clock-gated logic and low-power states to maximize battery life.

Customizable Analog Interfaces

Integrates programmable op-amps, ADCs, DACs, and more for direct sensor interfacing.

Flexible Development Process

An automated design flow allows non-experts to easily turn ML models into hardware solutions, using either open-source or commercial tools.

Efabless_Chip_noborder.png

Who is it for?

The chipIgnite ML platform is perfect for companies looking to build innovative remote sensor products across a variety of industries

Smart Homes & Buildings

Ideal for production line monitoring, inventory management, and predictive maintenance

Industrial Applications

Ideal for production line monitoring, inventory management, and predictive maintenance

Agriculture

Helps farmers optimize irrigation and monitor livestock health

Environmental Monitoring

Useful for tracking air and water quality, as well as structural health in infrastructure.

Retail

Enhances customer behavior analysis and inventory management

Personal Wellness

Powers fitness trackers, fall detection devices, and more

Competitive Edge

chipIgnite ML outperforms existing alternatives by providing application-specific custom ASICs that minimize active processing time, reduce power consumption, and streamline ML event detection. With power consumption as low as 10 microamps, chipIgnite ML is a game changer in edge ML technology.

Pricing and Availability

  • Proto-Only Option: Available as an open-source solution for prototyping and proof-of-concept projects. Pricing starts at $14,750.
  • Proto + Production Option: Includes full features for production-ready applications, starting at $30,000.

Both options include fabrication on an MPW shuttle, development boards, and access to EDA tools for seamless integration.

Timeline

SoC Platform Tapeout September 2024
Early Customer Development Release November 2024
First Customer Prototype Shuttle April 2025
First Production Release Late 2025

Get Started with chipIgnite ML

Unlock the potential of custom ML hardware for your next product. Contact us to learn more about our innovative platform and how it can help you accelerate development, reduce costs, and bring your unique product vision to life.