I think that first thing we need to address is that cold-welding isn't really as easy and practical as means of permanently welding pieces together as it might seem immediately apparent from its name (it's also know by others, e.g. stiction). What happens is that materials that can join into a cold-weld (similar metals) only do that on immediately adjoining surfaces, where atoms of one piece quite literally fill in the gaps between atoms of another piece so they're at that point essentially a single piece. So if you want to achieve a strong weld, the two surfaces you're joining would have to match perfectly. This obviously limits the use of printing with predominantly random surface geometry particles, since the structures wouldn't be as strong as it might be anticipated. Second thing is, that in order to reduce the unwelded space between two surfaces, either a lot of pressure is required, or a lot of heat, at which point you're already welding in more conventional sense and it most certainly can't be considered a "cold" weld. Assuming the former requirement (we sort of disqualified the latter) is met, that means that you'd be either firing small (probably nano-scale) particles at great speed to the surface you'd want them to cold-weld (impact cold-welding), or apply a great deal of pressure once the particles are more or less in the place intended. With additive printing, the latter is again not practical (we'd be back to now ancient days of needle-printing, not state-of-the-art additive manufacturing one).
So it would have serious limitations. But there are some fields that I can think of where the phenomenon of cold-welding could be of some use, for example with self-assembling nanostructures or self-healing metallic surfaces. Nanostructures are interesting for multiple applications, from medicine to create artificial manifolds on which crystalline and molecular structures could grow (say artificially grown tissues for transplantation, more potent medicine,...), to computing (data storage, central processing units, you name it), even batteries with greatly increased charge surfaces. And with self-healing surfaces, these could again be used in all kinds of fields, from astronomy to again computing, wherever nano-scale precision is required and parts of equipment used can be damaged on scale that the surfaces used could be repaired by self-healing to resume their function in mint condition.
So, in short, yes I imagine it could be useful, on nano-scale, but has serious limitations for large industrial-scale applications and it won't replace traditional welding techniques, such as laser and electron beam welding that could also be simplified if they were used in the vacuum of space (they are done in vacuum anyway).
