The difference is scale and cost. In 2 million miles of driving I've never had an internal combustion engine failure or a transmission failure. When transmissions fail they almost always do so because their fluid isn't changed often enough. Modern gas engines almost never fail when properly maintained. I've sold 5 cars that I've put over 300K miles on. Should any of them have actually had a major drivetrain component failure, it's unlikely the cost would have been $10K (DU cost) or $11K/$22K (battery cost).
I feel perfectly comfortable with my 320K mile Prius and my current 70K mile Lexus(last one was sold with 334K miles). It's unlikely they will ever have a major expensive issue. My current 2019 Prius should last similarly.
Not so with our batteries and DUs. My P85D has the original battery and DUs after 135K miles and that is unusual. Compared to the problems I'd likely have on my current Toyota and Lexus products, there's a very good chance that my P85D will require $40K+ in repairs in a very short period of time after the warranty expires. The likely out of warranty repair cost within a year or two of the warranty expiring is currently 90% of the current value and more than 100% if you figure the value in 2 years.
I'm not keeping it past the warranty. It will be gone at least 6 months before the warranty expires.
It's about probability of failure. You think nothing of having an ICE over 100k miles and 10 years etc. because you know from the experience of many many millions of cars before you that an engine just going bang isn't likely - it can happen but the probability is that low that you take the risk. Keeping a Tesla past 8 years (out of battery warranty) is way more unknown than that. There just isn't the history to tell us the probability of an out of warranty battery failure is 10%, or 1%, or 0.1% etc.
It's probably pretty low, but the cost to replace is so extreme that you really want to know that the probability is really low before being willing to take the risk. Third party repairs/replacements bring the risk down by bringing the cost down somewhat but we still have no idea of probability.
Add to that, users of a forum like this are definitely seeing close to ALL the cases of failure (pretty sure everyone in the situation would be seeking all the help/advice they can get before dropping $22k), and that's giving us an impression that it's very widespread. I am guessing there have been 30-50 cases reported on here, and that's out of c.25k deliveries in all of 2012 and 2013, so 0.12% to 0.2% failure. how that will change as years go by I don't know, but as of now I guess you could say that 1 in 500 batteries will bum out in the first year after warranty expires.
I would be extremely reluctant to spend $22k on a battery for an 8 year old car, because who knows when the suspension is going to fail, the sub-frame is going to crack, the motor is going to fail etc. but it certainly a lot less than spending $100k on a new model S. it's more than half of $40k on a 2019 model 3 SR+ though, which is still in full warranty for a while and lots of years of battery/drive warranty. When most of us bought our S's there was no 3, so perhaps the extra size and liftgate weren't actually needed, it's just what was available at the time.
I'm going out of warranty in June this year and so this is something I am thinking about a lot. I'm going to keep it but want to pin down some 3rd party options here in the UK for some peace of mind before I do!