I'd question a 100x decrease, but anywhere from 2-10x due to less obstacle avoidance and external factors seems reasonable. (Although newer designs by places li

I'd question a 100x decrease, but anywhere from 2-10x due to less obstacle avoidance and external factors seems reasonable. (Although newer designs by places like Hitachi promise to cut cost using smaller vehicles at higher frequency, probably in an effort to make inroads in NA). Something else that's worth considering is whether this scales. I'd expect a loop like the Las Vegas Loop to be somewhat larger to deal with concerns with ventilation and evacuation, something this loop doesn't have to deal with yet. I think we also want GoA 2/GoA 3 capable cars, which should be easier than FSD and be achievable by Tesla in the short term. Driving manually through small tunnels is not for everyone, and even with guidewheels, I think automation is pretty important for this kind of system. The technology to realize these loops in the way Musk wants them isn't quite there yet, but I'd expect them to be here soonTM.