
Reduced Energy Microsystems makes it possible for low-power devices to run computer vision and AI by supplying ultra-energy-efficient inference silicon. The company is a fabless semiconductor vendor that combines proprietary asynchronous resilient circuitry with custom neural network architecture and software tooling to execute state-of-the-art vision and inference workloads in a tiny power envelope. REM's chips target embedded applications such as augmented reality, body-worn cameras, and autonomous robots and are designed for integration by device manufacturers. The technology focuses on on-device inference and traditional vision pipelines to avoid cloud dependence and reduce energy use. REM operates in the embedded vision semiconductor market, addressing customers that need high-performance, low-power AI at scale.

Reduced Energy Microsystems makes it possible for low-power devices to run computer vision and AI by supplying ultra-energy-efficient inference silicon. The company is a fabless semiconductor vendor that combines proprietary asynchronous resilient circuitry with custom neural network architecture and software tooling to execute state-of-the-art vision and inference workloads in a tiny power envelope. REM's chips target embedded applications such as augmented reality, body-worn cameras, and autonomous robots and are designed for integration by device manufacturers. The technology focuses on on-device inference and traditional vision pipelines to avoid cloud dependence and reduce energy use. REM operates in the embedded vision semiconductor market, addressing customers that need high-performance, low-power AI at scale.
Business: Fabless semiconductor company building ultra–low‑power silicon for on‑device computer vision and AI
Founded: 2014 (San Francisco)
Founders: Eleazar Vega Gonzalez; Dylan Hand; William Koven
Funding signal: Grant funding including U.S. National Science Foundation; participated in Y Combinator
Status: Closed (recorded in public profiles)
Embedded computer vision and on‑device AI for power‑constrained devices.
2014
DeepTech
Multiple grant entries recorded; public profiles list NSF grants.
2344576
Total funding amount and last funding date recorded in company snapshot.
“Participated in Y Combinator; received NSF grant funding”