Also called an "intelligent processing unit" (IPU), "neural accelerator" and "neural engine," a neural processing unit (NPU) is a set of circuits or a separate chip designed for AI processing.
Such chips are called neural processing units because they were not originally graphics processing units (GPUs) that were adapted and enhanced for AI. In addition, neural processing units are often designed to operate in desktop and mobile devices, providing AI computing that is normally in AI data center servers. See
GPU.
For example, in 2016, Qualcomm introduced its Zeroth Machine Intelligence Platform and Snapdragon Neural Processing Engine that enabled mobile devices to execute AI applications on their own rather than going to a data center in the cloud. For another non-GPU type of AI chip, see
neuromorphic computing.