Non-Von

Non-Von is a Dartmouth-affiliated startup bridging the gap between algorithms and computing hardware. Our goal is to drastically improve AI and multivariate sensor processing applications, on the edge, and eventually in datacenters. Developing hardware from studies of brain circuitry - brains implement intrinsically parallel, power-sensitive algorithms, we combine novel software that uses such algorithms with novel parallel neuromorphic hardware architectures: Non-von Neumann machines. This combination of our custom algorithm designs with matching hardware, reduces price and improves watt performance by an order of magnitude. Our approach allows very large numbers of independent processing streams, none bottlenecked by a typical monolithic memory unit. Non-Von is building on almost a decade (and several million dollars) of R&D toward this goal; there are three patents with three more pending. Our team consists of expert computer architects, experts in A.I., experts in our target market segments, and strategic partners with half a century of experience in related low-power designs.