This conclusion makes sense, today, when the average car is used 5% off the time. However, what about a future where 50% of the cars on the road might be Uber-like self-driving cars, where the business model would dictate that higher utilization is better. In this case, these cars could be driving people around non-stop with virtually no need ever to pull over and charge (may allow up to 100% utilization of the vehicle). Just a thought.
Do you really believe that the utilization rate is even throughout a day, even in that future scenario? Even with future ride sharing services, the idea that people will not own cars is one thing, but the reservation levels would have to be quite high in order to make it likely that you can make it to work. The difference is that the additional car, especially for someone that isn't working could be handled by a ride sharing service. Utilization rates would increase, but that doesn't help the wireless case, since there would still be a wide variance between peak and non-peak utilization and therefore, many vehicles would have a chance to charge. As it stands right now, to drive 200 or so miles, a Model S can charge in 5 hours. That's roughly 5 times the current utilization rate with about 1/5th of a day in charging.
The energy transfer losses is quite high with wireless. A robotic plug solves this problem quite neatly without the massive energy loss. Maybe one day, when renewables are abundant, then paying that much extra money to waste that level of power makes sense.
Wireless charging of busses on predicable bus stop routes can make sense. The energy density of batteries might make it quite difficult, but the numbers would have to be crunched... intuitively, it might make sense, but the case is close enough that one really does still need to run the numbers.
Power in the roads almost never makes sense in the foreseeable future. Pack swap would make more sense before power in the roads.