this doesn’t account for the carbon footprint of making a entirely new car vs keeping an old one running well.

Part of the problem is deliberate Planned Obselecence as an industrial manufacturing strategy. Cars - particularly American cars - begin to fail after ten to fifteen years. Finding parts becomes more difficult over time, finding skilled mechanics even more so, and risks of accident (particularly on highways with speeds exceeding 55mph) lead to cars getting totaled before they've been fully exhausted.

I'll spot you that simply yanking new ICE cars off the road and replacing them with electrics is wasteful. But when you're talking about a ten year old vehicle, the math for those next ten years gets fuzzier as the risks inherent in ownership rise.

Incidentally, this is why mass transit improvements are an overall better play. Swapping old cars for new is never going to be as efficient as swapping cars for buses and trains, which are maintained as a fleet rather than as an oddball assortment of flavor-of-the-month private vehicles.

Those figures also assume all virgin materials for batteries. The reality is that as more batteries are built, they will reach a critical point where battery recycling is a major source of elements for new batteries. We're only just now coming to that point where there are 10+ year old EVs out there that have batteries that need to be recycled.

Also those studies all look at the super inefficient 3rd world exploitation of minerals and labor to get lithium. There are new techniques being developed out in the Salton Sea (desert in southern california) that extract lithium from ground water pumped in a closed loop. The expectation is that production technique alone will be enough for the entirety of the next few decades of American need. And that's a far, far more efficient technique.