True, the microballoons do reduce the epoxy shear strength, but it's still much higher than the shear strength of the foam, so it works well. What happens is that the microballoons key into the interstitial spaces in the sanded foam, so keying the resin/foam joint and adding strength. They also act to reduce the sharp transition in modulus between the relatively stiff resin and the relatively flexible foam, I think, as the thin layer of foam/microballoon/resin mix is less stiff than the resin, but stiffer than the foam.
This method is used for high load bearing structures is virtually all the Rutan designs. The key to producing good wing spars is to ensure that there is a decent shear web incorporated into the foam, so that the spar caps are properly supported and not wholly reliant on the strength of the foam core. The easy way to do this is split the foam, lay-up a glass web on the split faces, overlapping on to the spar cap surfaces, bond the two foam halves back together with the glass sandwiched and then lay up the spar caps. The result is a glass I beam set inside a foam core, which is both stiff and strong.
If I were making a composite electric bike, then I'd forget about trying to simulate tubes and go for a stressed skin monocoque, with space for all the components inside the "frame". This is easy to do with the glass on foam method, because you just carve the foam to shape, layup glass on the outside (using epoxy, not polyester or vinylester resin) and when it's cured pour some petrol (gasoline) in to melt out the foam core, leaving a hollow shell. The only thing to watch is to make sure that the access holes are cut into non-load bearing areas. Should make a nice light "frame" though, with no hassle about where to fit parts around awkward tubes. If you want extra stiffness in some areas, then a variation on the wing spar method above would allow internal webs to be added, or some foam can be left in, if it's sealed up from the petrol.
Jeremy