I find it difficult to project financial results from level 5 automated driving. I've mostly stayed out of any of this discussion because I think we are too far out to wrap our heads around it for stock valuation. Its fun to talk about future stuff like this, but not really something to be looked at from a financial modeling perspective.
I think we will have level 4 for a very long time. Tesla's FSD is basically a level 3/4, which means a licensed driver must be available in the car. The take over is not immediate, but it won't be at the level where you can expect to throw your kids into the robot car and give it an arbitrary address and it will be fine going there completely on its own. If we are at level 3 for 2-3 years, then level 4 for another 5-7, we're talking 7-10 years before we can really talk about Tesla Network. That's too far out to make any projections.
The thing is, we don't need level 5 to have an extremely robust sales level and stock price. As it stands, there are cadres of people that buy Tesla's in order to deal with mind numbing traffic each day. As the vehicle gets better and better at handling 95+% of the every day commute, to the point where you can sleep or read a book safely while the car is driving itself in traffic, that's all that is necessary to drive sales to be production constrained. The next part is the ability to drop off all passengers and find its own parking spot and charge at a local shopping center. That's on private property and can be done with special toll tags/parking areas if need be, so that Tesla owners never need to hunt for a parking space. That's a game changer. No need for full Uber/Lyft replacement in some distant future in order for Tesla to be production constrained. Of course, the issue is then monetization by the mile or by the trip rather than by the sales/service of the vehicle. Sure, we've all seen the various calculations. I just don't think we'll have a handle on that for quite some time. And again, it depends on how quickly it arrives.
As for level 5 fully automated driving, I'm looking forward to red-eye drives. Especially at slow speeds. If you want to get to a destination that is about 300 miles away, you can get into the car at night and wake up at your destination. The car can go as slow as possible so that you can arrive at the right time in the morning. If that means driving 40 mph, then great... lower energy consumption, smoother ride. We can't do that today because humans don't want the drudgery of driving for an even longer period of time, not to mention the safety and strain of driving overnight. And multiple speeds of vehicles on the highways is dangerous today. But maybe that won't come to pass. It might be that every trip is a service fee, as I get a vehicle on demand that suits for the number of passengers, the distance, and the speed required. Maybe we go back to the pod concepts with exchangeable power modules. So much is unknown as we move forward into this new automated driving world.