Edge AIChips & modules

Analog compute-in-memory AI processors for power-constrained vision and industrial systems

Mythic builds analog matrix processors that store neural network weights in flash memory and perform multiply-accumulate operations inside the memory array. The M1076 chip and its PCIe and M.2 cards run convolutional networks for video analytics, drones, industrial inspection and defense systems with a fraction of the power a GPU needs. The company supplies a compiler and runtime that map trained models onto the analog tiles.

Profile added on October 17, 2026 from the company's official site; products, pricing and limits have not been verified yet. Spotted an error? Tell us or see the editorial policy.

Focus: analog compute in memory, ai accelerators.

Tags: analog ai · compute in memory · video analytics · drones · defense

More edge ai companies

Ambarella

Computer vision SoCs with on-chip neural network engines for cameras, robots and vehicles

Santa Clara, United StatesFounded 2004

Aspinity

Analog machine learning chips that classify sensor data before it is digitized

Pittsburgh, United StatesFounded 2015