AV1 already wins handily as a codec, and the only thing keeping it from being adopted more broadly than is currently the case is lack of hardware decoders on older hardware. This problem naturally solves itself as old hardware gets replaced.
Even then, dav1d is a remarkable piece of software, and software decoding is pretty viable for AV1 thanks to it. Many places have already adopted AV1, and you should expect to keep seeing it get adopted as time goes on.
Better hardware encoder support would help, too. It’s insanely inefficient to encode without that dedicated hardware, compared to h264/h265, where dedicated hardware support is there.
I was hoping Apple would add it when they shipped the M4, and now M5, but nope.
Hardware encoder support I think is generally less critical. Decoding is the process that needs to happen real-time, while most encoding can be done far in advance, unless you’re live broadcasting or operating at YouTube-scale.
While I agree, my point was that encoding needs to be more efficient, both in time, and resource consumption. That isn’t quite there yet, for AV1. It is improving, albeit slowly.
It’s too bad the GPU prices are utter insanity due to the LLM pyramid scheme poaching global RAM. I read an article yesterday that said Apple is likely eating that RAM overhead as a loss to ensure their long term strategy.
AV1 already wins handily as a codec, and the only thing keeping it from being adopted more broadly than is currently the case is lack of hardware decoders on older hardware. This problem naturally solves itself as old hardware gets replaced.
Even then, dav1d is a remarkable piece of software, and software decoding is pretty viable for AV1 thanks to it. Many places have already adopted AV1, and you should expect to keep seeing it get adopted as time goes on.
AV1 has recently gotten involved in a lawsuit by Dolby saying that they’re breaking like
54 of their patent, so there’s some issues there as well: https://arstechnica.com/gadgets/2026/03/av1s-open-royalty-free-promise-in-question-as-dolby-sues-snapchat-over-codec/These things always happen because of how dumb software patents are. There’s no guarantee the lawsuit will stick, nor do I necessarily expect it to
It’s even worse than that: the USPO abrogated their responsibility to evaluate patents for prior art and conflict with other patents to the courts.
So they just issue patents willy nilly and expect courts to decide which ones ‘win’.
Better hardware encoder support would help, too. It’s insanely inefficient to encode without that dedicated hardware, compared to h264/h265, where dedicated hardware support is there.
I was hoping Apple would add it when they shipped the M4, and now M5, but nope.
Hardware encoder support I think is generally less critical. Decoding is the process that needs to happen real-time, while most encoding can be done far in advance, unless you’re live broadcasting or operating at YouTube-scale.
On a media server encoding is typically done in real time
It depends on what the receiving unit can decode. Sometimes there will be transcoding, but it’s usually something you want to avoid.
While I agree, my point was that encoding needs to be more efficient, both in time, and resource consumption. That isn’t quite there yet, for AV1. It is improving, albeit slowly.
non apple chips can pretty much all hardware encode AV1 nowadays. it’s really just apple doing its own thing again
It’s too bad the GPU prices are utter insanity due to the LLM pyramid scheme poaching global RAM. I read an article yesterday that said Apple is likely eating that RAM overhead as a loss to ensure their long term strategy.