Having built a telescope mount or two, I would not worry about stiction of the tapered roller bearings. I cannot say that I have noticed any tendency to stick before movement begins. In fact, the movement is so free that it is necessary to make careful provisions to balance the tube assembly perfectly, or the scope will easily drift. In practise, you will have to add a brake to each axis that lacks a worm gear drive and clutch assembly. You will have more difficulty trying to control the breakaway friction of your drive clutches than you will with the bearings.
I used standard Timken tapered roller bearings. If you get using something like a 4" bore bearing, you are looking at a housing of some serious size and weight to hold it all. You might want to consider moderate weight of the mount unless this is a permanent installation.
IMO, 35kg is no significant load. I know that ATM books recommend huge axles and mounts, but I believe, now with a few years of machining experience behind me, that flexure is an overblown issue with small telescopes. The real flexure issues come about in sneakier ways than the "bending of your 2" shafts"
For example, every bearing fit must be tight: that means no "slide together" assemblies, because that is looseness. This means that assembly is permanent, because you'll have to heat most OD fits to get them together, or Loctite them together with a space filling compound, which makes disassembly a very difficult procedure.
Where the T for the declination axis joins the RA (polar) axis must be tight. A slide fit, one that is not too difficult to assemble when holding a rather awkward contraption up at mid chest height, means that you've got .001 clearance between the housing and the polar axle. That is where you will get the sneaky flexure of the mount. It will not announce its presence with a clonk or a click, but it will introduce random errors in positioning of the tube, depending again, on where you are pointing the scope and where the weights are.
Your tube itself needs to be incredibly stiff to not introduce some small measure of flexure. The spindly aluminum saddles that one often sees atop these mounts is a joke for stiffness. A lot of this stuff, you cannot really test until you've built the first mount, and then test its stiffness with a dial indicator in various attitudes. That's when you discover what is what.
First you get good, then you get fast. Then grouchiness sets in.
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)