Putting the motor in the wheels must produce a hefty increase in rotating mass and in un-sprung mass, both undesirable. It's necessary in the case of a retrofit, but in a from-scratch design I doubt an automotive engineer would try this.
There's certainly some downside, but the compromise can be worth it for packaging reasons, and there are a few mitigating factors.
For one motor torque scales with the square of radius, so a relatively light pancake motor can have quite high torque. You can also downsize the friction brakes thanks to regenerative braking. Finally you can distribute your motors across all the wheels.
In exchange you get a drivetrain with one moving part per wheel (yes, leaving aside bearing needles, you pedant!).
I don't foresee this making its way into sports cars, but I could see the minivan of the future having a drivetrain like this.
That video is about a show car, and it says the technology will show up in he 2010 model year. But it didn't happen.
Electric cars face big challenges in terms of cost and weight; I expect that a design like this, which increases both, is far, far away.
The Tesla cars, which cost a huge amount of money, don't do things this way. The Toyota hybrid system, probably the most developed in the world at the moment, doesn't do things this way. The people who designed those things were smart people.
