Aspinity raises $5.3m Series A

1 min read

Pittsburgh based Aspinity, a developer of ultra-low-power analogue machine learning processors, has announced that it has raised $5.3 million in Series A funding.

The funding was led by Anzu Partners, a venture capital and private equity firm that focuses on breakthrough industrial technologies. Other participating investors in this round include Amazon’s Alexa Fund, Birchmere Ventures, Mountain State Capital and Riverfront Ventures.

Founded five years ago, Aspinity has developed a type of machine learning that is enabled by proprietary analogue circuit technologies, to dramatically improve battery life in smart electronic devices that are always on and always sensing the environment for acoustic, vibrational, or other ambient triggers.

These devices are continuously consuming power while waiting for triggers such as voice to wake-up a voice-controlled TV remote or wireless earbuds, the sound of glass breaking to activate a smart home security system, or a change in vibrational frequency to provide early warning of equipment malfunction on the factory floor.

While such smart devices are growing in popularity, with forecasts suggesting a base of 1 billion voice assistant devices in use by 2023, they struggle with short battery life because of the power required to continuously digitize and process all incoming sensor data, relevant or not.

“Aspinity’s mission is to help system designers solve a major pain point that users face on a daily basis — the necessity of recharging or replacing batteries way too frequently in their always-listening devices,” said Tom Doyle, founder and CEO, Aspinity.

“Our new investment will allow us to speed deployment of our first neuromorphic analogue-processing chip, which completely changes the way that incoming sensor data are handled. Instead of a system that immediately digitizes all data for further analysis, our chip uses near-zero power to analyse the sensor data while it is still in its native analogue format, only waking the digital system when important data are detected.

"This makes our architecture a game-changer for manufacturers who want to make smart always-on devices with batteries that last up to 10 times longer. Just imagine wireless earbuds that last for months instead of a day on a single charge or a voice-activated TV remote that runs for years without battery replacement.”