Here's the rear hub of that riders wheel in the video above.
Absolutely no way you could do or would want to retro-fit any form of usable motor into the small void available within the hub shell. Maybe a couple of watts, but think how small the rotor and stator would have to be to be self-contained in what is essentially no more than 5mm gap around the axle which would still have to be there to withstand the forces going through the bearings anyway.
Now lets look at Katusha riders frame and Zipp rims of his Canyon Aeroroad CF SLX and see if you could get a usable motor in there to spin a wheel round.
The Rotor (alternating polarity magnets) part of the motor has to be in the rim of the Zipp Firecrest 404s he's riding. The Stator (Coils) of the motor has to be within the seat stay and chain stay VERY close to the Rotor magnets when the wheel is turning. Anything more than 1-2mm gap between rotor and stator is incredibly inefficient because the force of permanent magnets exponentially falls very rapidly when not close to one another. For every 1mm of gap about 20% of the usable force of the motor is lost.
So Lets look at the gap between stays and rim on Katusha's Aeroroad CLX.
There's about a 25mm distance between where the stator could be in the seat stay and where the rotor magnet will pass it inside the Firecrest rim, so essentially ZERO usable magnetic force from the seat stays possible using the chart above. Obviously you could use stronger, much heavier magnets, but even so incredible heavy then for the wheel and inefficient still anyway.
Looking at the chain stays of the CLX
You could argue they might be around 10mm from the rim, so a very slight usable magnetic force there, but tiny in reality still. There's only two chain stays and maybe 3cm of space to fit the stator into where it would align with the magnets in the rim. As Varjas explained, this type of wheel is not considered a motor like a crank-based one, he classes them as boosters because the wheel is really only being propelled from one stator in the stays not around the whole circumference you really require for any usable power.
Here is what a usable rim and frame-based motor looks like in the real world. Notice how the magnets all have to be glued visible to outside of rim to maintain the 1mm gap (even inside the rim you've already increased the gap by 1mm and lost 20% efficiency by doing so). Notice how the seat tube and stays are wrapped around nearly 90 degrees of the wheel to so the stators are working more efficiently with the magnets, not just the 5 degrees available in a typical off-the-shelf frame like the Canyon.
You could argue you might squeeze a couple of usable watts into a typical Zipp rear hub within the couple of cm3 available and a couple of watts in Canyon Aeroroad chainstay given how inefficient it would have to be due, but even that is extremely difficult to do invisibly in the space involved to a simple tablet the UCI uses to detect motors. If you're protecting the magnets from magnetic field detector, you're by default also protecting them from being able to interact with the stator which needs that magnetic field to work off or going down using coils instead of magnets in the rotor, but then you've got the issue how to power the coils wirelessly and that presents a whole new challenge in itself.