Video Codec Wars: AV2 Is Faster but Patent Risk Remains
The dav2d decoder for AV2, the successor to AV1, landed on HN today and the comment thread immediately zeroed in on two problems. First, AV2 decoding is roughly five times more complex than AV1, meaning real-time software decoding on current hardware is borderline. Second, and more pointed: AV1 was designed as royalty-free, but Sisvel and then Dolby and Snap successfully asserted patent claims against it anyway. The question being asked is how AV2 avoids the same fate.
The pattern is that the 'open, royalty-free codec' story keeps colliding with patent trolls and patent pools who treat 'royalty-free' as a challenge rather than a conclusion. Companies that built infrastructure around AV1 on the assumption of zero royalty exposure got a nasty surprise. AV2 promises better compression (roughly 25% better than AV1) but at the cost of hardware that does not yet exist and a patent landscape that is, charitably, unclear.
For anyone building video infrastructure today, this is a real planning problem. The codec you choose today determines your hardware requirements and your legal exposure for the next five to ten years.
So what?
Do not commit to AV2 for production video infrastructure without a clear legal opinion on patent exposure, and do not assume 'royalty-free' means 'legally safe.' If you are already on AV1, the Sisvel and Dolby precedents mean you should have that conversation with a patent attorney regardless. The compression gains from AV2 are real, but the hardware and legal readiness is not there yet.