Friction is due to the boundary condition. Even if the pipe doesn't become wetted by the fluid (there is an air space between pipe and fluid), the fluid's surface is still held in place by the microscopic contact points which give rise to the lotus effect.
Now, the surface is fluidic too, so it's not perfectly still. The boundary condition is more complicated. Parts of the surface (between contact points) can still move. That surface movement, in turn, shears against the air space, which is much less viscous so we can probably ignore it.
Over a long time scale, I wonder if the effect is persistent. Consider: all gases have nonzero solubility in liquid. Slowly, the air space shrinks as its gas is absorbed by the fluid (being deposited elsewhere in the system). (Maybe this won't happen everywhere in a closed system, but at least somewhere.) For it to be persistent, it has to have so much energy in surface tension that an actual vacuum can be formed, filled only with the fluid's vapor. Which I suspect just can't happen with most fluids.
It's not a useless question -- it's certainly possible with the right combination. Take mercury for example: it has so much surface tension, and so little surface energy, that it doesn't really stick to anything nonmetallic in the first place. Even smooth surfaces in vacuum. (On the other hand, it strongly wets most ordinary metals!)
Also, we can test a continuum of surface energies down to zero, by heating any fluid towards its critical point, where surface tension disappears completely (gas and liquid densities become equal and phase separation ceases). So the question is, what combinations of what fluids, at what temperatures and pressures, and with what surfactants if any, can make use of this?
Applications where it can benefit, assuming the boundary layer remains intact -- include viscous, viscoelastic or inhomogeneous fluids: if the bulk fluid has higher effective viscosity than the liquid surface, then there can be more shear right near the surface than in the bulk. That is, it has laminar flow but with a non-parabolic velocity profile due to a nonlinear shear property.
Otherwise, if the fluid wets the pipe surface, it's just any other rough-pipe condition, and your surface treatment is at best wasted, or makes things worse.
Surface roughness induces turbulence at somewhat lower velocities, so has a bit more drag than a smooth pipe. There is a length scale that goes with that: very small features diffuse energy rather than shed vortices. A nanoscale surface treatment might be effectively smooth for water, but not, say, for air. A rusty pipe might be effectively smooth for honey perhaps, but not water.
Tim