Google marketed it that way but I could never reproduce a meaningful size savings without noticeable quality loss. You need to serve a LOT of video before even the top-end 10% savings was worth it, especially if your traffic was spread across many items so doubling your storage cost cancelled out a fair chunk of the total. I have no doubt that YouTube saw a savings but I don’t know how many other sites did, and I would be curious what the savings was relative to the extra power used by the millions of client devices which could’ve streamed H.264 at 10% CPU versus having the fan on high.
If users don't have hardware accelerated video decoding, it's so bad that it actually hurts the experience. I can't imagine that being worth the space savings. There doesn't have to be a good reason YouTube does it, it might just be someone wanting to insert their tech, which I'm pretty sure is the reason Meet uses it.