Alif MCUs Combine Arm Cortex and Ethos

Targeting IoT, wearables, consumer electronics, and robotics, Alif Semiconductor is building microcontrollers that consume less than 1 W and provide up to 250 GOPS using Arm Cortex CPUs and Ethos AI units.

Anand Joshi

Armed with $235 million in funding from prominent VC firms such as Kleiner Perkins, Alif Semiconductor is on a mission to add AI accelerators to microcontrollers (MCUs). The Silicon Valley headquartered company is building a portfolio of chips that combine Arm’s Cortex CPUs and Ethos AI core. It’s targeting IoT, wearables, consumer electronics, and robotics.

Alif’s MCUs consume less than 1 W and provide up to 250 billion operations per second (GOPS) to run AI applications at the edge. They’re configurable to include a combination of Cortex-M, Cortex-A, and Ethos cores. The chips are sampling now and have seen early adoption for computer vision. The company plans to begin production in 2H23.

Offerings from startups as well as prominent semiconductor vendors that are already shipping crowd the market for low-power AI chips. By providing multiple processing cores, the Alif designs will allow customers to run a complete application pipeline consisting of preprocessing, neural networks, and postprocessing, along with an OS. The multicore capability gives the company a differentiator over its competition. Arm’s Ethos intellectual property (IP) is still unproven, however, and Alif must demonstrate a performance advantage along with a seamless integration into Arm’s application-development flow to become a major player in this market.