it's not about what "could work", a 15fps game "could work".
it's "what is successful". And for a mobile audience, UI/UX engineers have spend decades determining that after a certain amount of lag, that experience is lost.be it for technical or political reasons, Safari and by extension, much of HTML 5 cannot meet this UX criteria. So an application has an massive advantage.
>but if someone could write a performant JS/web assembly transcoder that could run on the web
Sounds like a Progressive Web App: https://web.dev/what-are-pwas/
They've come far but are far from perfect. It's also moot in this case, since certain optimizations would need to rely on understanding Safari's infrastructure. Which is closed source. They can optimize it for webkit, but any web dev can tell you that interplatform quirks are the 2nd 90% of the battle.
You can’t do advanced workloads in the background with native apps on iOS. Apple strictly limits what an app can do in the background to conserve battery life. Safari doesn’t stop running in the background.
I don’t know about performance, but if someone could write a performant JS/web assembly transcoder that could run on the web, since iOS devices routinely trounce all Android devices with respect to browser performance, I’m sure it could be done.