Mythic logoMY

Mythic

About

Mythic, founded in 2012, is a semiconductor company developing a fundamentally different approach to artificial intelligence inference. The firm's core innovation is analog compute-in-memory architecture, which collapses traditional compute and memory into a single plane. This approach aims to eliminate the memory bottlenecks that constrain conventional digital processors during AI workloads.

The company's flagship product, the M1076 Analog Processing Unit (APU), stores AI model parameters directly in the processor. Mythic states the chip delivers up to 25 TOPS (tera operations per second) and achieves energy efficiency up to 100 times greater than industry-standard graphics processing units. The technology targets applications across data centers, automotive systems, robotics, and defense.

Mythic has raised over $125 million in funding to date. The company maintains offices in Austin, Texas; Palo Alto; and Bangalore, positioning it within key US and Indian technology hubs. Its technical focus spans analog computing, AI inference, compute-in-memory architectures, and energy-efficient AI processing.

Similar companies

Gauss Labs logoGL

Gauss Labs

Gauss Labs develops industrial AI solutions and software to make factories more intelligent by analyzing machine-generated data, with a focus on manufacturing and semiconductor sectors.

1 job
d-Matrix logoD-

d-Matrix

d-Matrix builds computing platforms using digital in-memory compute technology, designed specifically for efficient and scalable AI inference workloads.

TensorWave logoTE

TensorWave

TensorWave builds a cloud platform for AI workloads powered exclusively by AMD Instinct accelerators, operating one of the largest liquid-cooled AMD GPU clusters in North America.

NVIDIA USA logoNU

NVIDIA USA

NVIDIA designs GPUs and accelerated computing technology used in gaming, artificial intelligence, high-performance computing, and various industrial applications.

Liquid AI logoLA

Liquid AI

An MIT CSAIL spin-off building efficient foundation models designed to run from data centres to edge devices, moving beyond traditional transformer architectures.

Featherless AI logoFA

Featherless AI

Featherless AI provides a serverless inference platform for instant access to over 20,000 open-source AI models, featuring rapid hot-swapping and significant cost reduction.