XML was abandoned because we realized bandwidth costs money and while it was too late to do anything about how verbose HTML is, we didn't have to repeat the mistake with our data transfer protocols.
Even with zipped payloads, it's just way unnecessarily chatty without being more readable.
Your memory is correct. Once compression was applied, the size on the wire was mostly a wash. Parsing costs were often greater but that's at the endpoints.
But one of those endpoints is a client on a mobile phone, which when we started with Internet on mobile devices wasn't a particularly powerful CPU architecture.
This is true, but if other formats work for those purposes and also network transmission, they'll start to edge out the alternative of supporting two different protocols in your stack.
if bandwidth was a concern, JSON was a poor solution. XML compresses nicely and efficiently. Yes it can be verbose to the human eyes, but I don't know if bandwidth is the reason it's not used more often.
JSON absolutely isn't perfect, but it's a spec that you can explain in ~5 minutes, mirrors common PL syntax for Dict/Array, and is pretty much superior to XML in every way.
Even with zipped payloads, it's just way unnecessarily chatty without being more readable.