That is a very nice bearing. The size seems to be right as well. I currently want to keep costs as low as possible.  I aim to glue #54-090 from edmund optics to my prism. It is a perfect cylinder lens with a costs of 45 euro's. 

What will the lens go into that will make it an aerodynamic bearing? Also, will the effect persist in this reversed fashion? Why insist on an expensive lens? I have so many questions about this plan.

There will be no aerodynamic bearing at the start. If I call a CNC shop to make me a good sized rod, it costs me well over 50 euro's. In bulk the price is low but not an one off.  I am working on a PID controller and a kalman filter. My current focus is to improve the algorithm to stabilize the rotation. I don't want to break a part the current setup. I therefore postpone trying out a different mounting procedure.

Since you're making a PID controller, you might as well make a full motor controller. Considering the low amount of power needed and only going in a single direction, you may be able to get away with a small MOSFET, triggered by a single transistor, controlled by the FPGA, for each channel. This would make it easy to integrate the motor control and PID controller.

devil is in the details with these things..

Before you jump into designing your own aerodynamic bearing made of quartz, I recommend looking at COTS parts. I have no doubt that further down the line that this could be optimized into a single piece of quartz that is the prism, bearing, and holds the magnets but COTS may be the cheaper option without making thousands. However, I do wonder how such a shape would impact the laser light.

Here's an example COTS part: https://www.newwayairbearings.com/catalog/product/25mm-flat-round-air-bearings/