Essentially the take away is that each mile driven is fetching about .05c per mile. IF Tesla where to sell this data and I dont believe they will directly sell this data, they could fetch .05 x 15,000mi per year per car or roughly $7500. She actually says per camera but I im not going to be that optimistic. What I do agree with is that the data is valuable to Tesla and $7500 per car sold is probably a minimum as to how much its worth to them as they would need to acquire this data if they where not generating it. I guess it is possible that they could sell the data or share in a partnership where Tesla is getting something back, such is map data or machine learning algos or something beyond my rudimentary ability to guess. The value is there, but what form it takes is debatable. Does ford have this value from its cars? They will need to buy this data from someone or the resulting solution. The more uses the can find the data the more valuable it is. The first is probably high def maps and the next is probably FSD. But what else could you do with this data? Advertising is one place where you know roughly how much people make and where they are driving and when. When they stop to charge, you could popup a groupon on the screen that offers some savings while they charge for example. Very rudimentary, but the number of things you can do would be nearly infinite which makes the data more and more valuable. How many ways can they leverage the data, can they find enough solutions to pay for the entire cars marginal cost? Maybe.