About Mythic
Mythic is a semiconductor company developing analog compute-in-memory technology and Analog Processing Units (APUs) for artificial intelligence inference. Its architecture stores AI model weights in flash memory and performs computation directly where the data resides. The company describes this approach as a way to reduce the energy spent moving data between separate memory and processor components, addressing a major source of inefficiency in conventional chip architectures.
The company’s product portfolio includes the Mythic M1 APU, designed for AI workloads in edge devices and other power-constrained environments. Mythic says the M1 delivers 25 TOPS of inference performance, and its product page describes the chip as being used by customers across automotive, robotics, defense, and edge AI applications. The company also offers an M.2 card for embedded applications and presents its products as a platform for deploying AI closer to where data is generated.
Mythic positions its architecture for use cases extending from edge devices to enterprise computing. Its website describes applications including cameras, drones, vehicles, robotics, industrial systems, and data-center or enterprise AI inference. The company says its chiplet approach and software stack are intended to support deployments at different scales. These are company-reported product capabilities and intended applications rather than independent performance assessments.
Mythic also develops software to support the use of its hardware. The company describes a software stack that optimizes and compiles trained neural networks, with a development flow that includes familiar tools such as PyTorch. This software is intended to help teams deploy neural networks on Mythic hardware without requiring them to manually manage all hardware-specific conversion steps.
Mythic was founded in 2012. Its company materials trace the origins of its analog compute technology to co-founder Laura Fick’s doctoral research at the University of Michigan. LinkedIn identifies Austin, Texas, as the company’s headquarters, while Mythic’s public materials also describe a presence in other locations. The company acquired Videantis, a German processor IP company, in 2026, according to Mythic’s announcement.
Mythic’s stated focus is enabling AI to run reliably and efficiently in real-world environments. Its team page describes a multidisciplinary group of engineers, scientists, and specialists working on analog processing units and compute-in-memory architecture. Public company materials emphasize building a platform for practical AI performance and reducing the energy burden associated with conventional AI computing; they do not provide detailed policies or broader claims about workplace culture.
