As I understand it, Tesla's autopilot is currently not capable of self driving. It requires constant monitoring and intervention. In its current state, it has trouble in many situations.
The 6,000,000,000 miles clock will not start until the car requires no human intervention (or at least constant monitoring i.e. level two) . Once the software reaches the state, they will need 6,000,000,000 miles over which time they can show the vehicle was safer than human drivers.
In other words, they need 6,000,000,000 miles at whichever level (three, four or five) of capability whatever they are aiming for. They are still very much only at level two.
My $.02 on the 6B miles - that's an Elon throwaway comment intended to illustrate directionality and order of magnitude (my interpretation), using his understanding, as of the writing of the 2nd Master Plan. The 6B miles IS NOT an established regulatory standard by any regulatory body that I'm aware of, nor is it even an aspirational statement by a regulatory body.
I'm on board with the idea of making guesses at how quickly regulatory approval will come, both the first approver, and how quickly other bodies will fall over themselves to get on the train. But be clear that they're all guesses, and there are all sorts of external factors that influence when the guesses come true.
Here's an article about Texas' legislature:
The Texas Legislature just ended its session. Here's what to expect next.
They meet every other year, for about 1/2 the year, in odd numbered years. They're done for this year, so either Texas calls a special session of the legislature (something I get the impression they don't like to do), or the SOONEST Texas would even consider Texas wide approval of autonomous driving is the front half of 2019. And if it doesn't make their list of things to do, then the NEXT opportunity is 2021.
I'm not picking on Texas - just using them as an example of legislative calendars and the pace at which things can get done.
Guess away, absolutely. Be clear that you're guessing, and try to be clear about where your guard rails are at, so we can be clear about when things take more (or less) time than expected, we can also see where the assumptions stop working.
For my part, an autonomous Tesla network is purely theoretical for my valuation / view of the company. The legislative impediments are too big and too unknown, the amount of work between what we see today in the wild and what we'll need to see in the future to consider real autonomy is also too big and too unknown for me. I realize that the market grants market capitalization value to companies that are doing this well, and from that point of view, Tesla should get credit for their progress.
In terms of an actual business model and business plan, my valuation of Tesla continues to be based on $0 from this sort of automation. Not because I think the option value is actually 0, but because I don't yet have any real information on what impact on Tesla it can have. If I needed this business to have value to Tesla today in order for Tesla to be worth owning, I personally would call this the top and exit the company.
I see several years of huge growth for Tesla whether anything comes of this potential future business or not. The option value is pure gravy over and above the company, and the level of guesswork layered on guesswork for me to back into a more specific value is high enough, and sufficiently unnecessary, that I value the company without it.