We present a new kind of video transcoding called “video vectorization”, which transcodes video to a vector-graphics format and leverages open and existing standards such as SVG and OpenGL for hardware accelerated rendering on end users’ devices. Doing so can enable bit rate reductions of over an order of magnitude for animated and screencast video content without the need for a new codec or decoder on end users’ devices. We explain at a high level how the technology works, and provide some real world demos of vectorized cartoon and animated videos, and provide bitrate / quality / speed /performance comparisons with AV1, h.265, h.264 and VP9. We go over some of the advantages of the approach, including scalability and ultra-low bitrates, as well as some of the disadvantages – namely it’s restricted applicability to “vector-friendly” content. We show how vector graphics can be used in parallel with existing codecs within a single video stream for “hybrid content” (e.g. news tickers and sports scoreboards), as well as how it can practically integrated into modern streaming architectures, from packaging vector streams within an MP4 container, to streaming via HLS/DASH, to integration into popular javascript players. Presented at Demuxed 2020.