There's the perfect solution in an ideal world, and there's what actually works given people's actual behavior. For the latter, you have to consider that if you make EVs less compelling, more people might just stick with their ICE vehicles, which are worse overall when you take all pollutants into account.
After the EV transition is complete, though, that'd be a good time to start limiting torque. Unless most people are just riding self-driving taxis by then, which probably won't be doing fast acceleration anyway.
Alternatively: the time to limit torque is now, while consumers are still getting to know EVs and the other substantial benefits they present. Those other benefits ought to be enough to drive adoption, and the longer we indulge the most harmful excesses, the harder it will be to reverse them in the future when the trap has sprung.
I don't really think either of these arguments is obviously correct, so I lean toward the option that immediately limits harm. But I'm also under no illusion that anything other than short to medium term profits are driving automakers' decisions, so in practice we will surely get the worst of both worlds.
After the EV transition is complete, though, that'd be a good time to start limiting torque. Unless most people are just riding self-driving taxis by then, which probably won't be doing fast acceleration anyway.