Chips&Media is the leader in the video hardware IP. We have long been perceived as a Video Codec IP specialist with firm market leadership for the last decades. The recent introduction of new video IPs, including deep learning-based computer vision, computational photography, and image signal processing, are transforming Chips&Media as a total Video IP provider. Chips&Media’s deep learning inference computer vision IP is designed to provide fully hardwired acceleration for computing-intensive and application-specific neural network algorithms developed in houses. Our IP’s dedicated architecture reduces external memory bandwidth significantly, minimizing power consumption. These characteristics place Chips&Media’s computer vision IP optimally suitable for automotive and surveillance applications such as embedded vision with rapid processing of extensive visual data from multiple sources (i.e., cameras), where high performance, power, and area efficiencies are essential.