Derived from the Latin word for "faster," the word “ocius (OH-see-us)” in our company’s name, Ocius Technologies LLC, indicates that our research concerns making technology work much, much faster.
Our analog computer chips and design are supported by a Phase-2 DARPA project. Ocius Technologies is focused on analog computing for
High Speed and Low Power: Analog computing for computational electromagnetics at radio frequency in real-time. Performance of our analog computer chips show 400x speed improvements over the latest NVIDIA GPUs while consuming only 200 mW. (The GPUs consume about 50W)
Complex Simulations: Speeding up standard CMOS based analog computing for simulations like plasma containment (important in fusion energy) and high-performance AI/ML that is 100x-1000x improvement over the GPU farm approach used today.
Other Applications of Analog Computing: Any applications that require advanced mathematics, such as artificial intelligence and machine learning, cryptography, sensor processing at the edge, and physics simulations involving systems of linear and non-linear partial differential equations (PDEs) with applications in domains such as acceleration of nuclear fusion, 5G/6G communication system design, electronic warfare and advanced unhackable sensing/computational platforms for intelligence and defense applications.
Also, see a more complete listing of our research projects here.
Note: A publication from our team titled “A Radio-Frequency Analog Computer for Computational Electromagnetics” is scheduled to appear in the IEEE Journal of Solid-State Circuits (JSSC).