JP Data LLC
101 Jefferson Dr, 1st Floor
Menlo Park CA 94025

Phone
(408) 623 9165

Email
info at jpdata dot co
sales at jpdata dot co

Competition between low power and ultra-low power AI chip vendors

According to a new report that we published recently, ultra low power AI chip market is expected to reach $3.2 billion by 2027. The ultra-low power segment is defined as chips that consumer less than 100 mW. These chip are targeted towards always on applications such as wakeword recognition that are always running in the background.

The ASP of such chips is expected to be ~$1. In essence, the ultra low power chip market is currently going after what is historically MCU (microcontroller) market and adding intelligence to that. This is enabling wide range of new applications – such as a microware that understands spoken command or a smart-watch that detects a person falling.

Ultra low power segment, is likely to face its stiffest competition from low power companies. Most of the devices ultra low power chips go into are already having some sort of application processor. Qualcomm for example also offers an ultra low power mode in which majority of the chip is shut down and only part of the chip runs algorithms. So why would someone consider a separate chip to enable always on functionality? Is it worth adding extra chip and a BOM when low power can accomodate ultra low power mode?

There are two possible outcomes:

  1. Ultra low power chip companies find applications where the application processor is the ultra-low power chip itself. This could be a small camera with long battery life at a low cost. In other words, these chips create a new class of devices where low power chips can’t work. Synaptic’s Katana platform is trying to enable new set of applications with this.
  2. Ultra low power chip companies justify their existence for extra BOM by enabling a range of new use cases that one could not achieve via low power chip. For example, being able to run complete speech recognition. This of course requires some out of box thinking – perhaps inventing and productizing new architectures and only time will tell if this works.

Irrespective, a new class of chips is borne and is here to stay. We’ll monitor where and how far it goes.