Untether AI provides energy-centric AI inference acceleration from the edge to the cloud, supporting any type of neural network model. With its At-Memory compute architecture, Untether AI has solved the data movement bottleneck that costs energy and performance in traditional CPUs and GPUs, resulting in high-performance, low-latency neural network inference acceleration without sacrificing accuracy.