Sure, I'm aware that this is fundamental research (likely not on actual protein folding, but haven't looked at the details), which will probably not have short-term outcomes, not on the timescales required for a 1st gen vaccine (perhaps not even 2nd gen).

That does however not take away from my main point, if anything it is in support of what I was trying to emphasize: such computational research is better suited for a compute cluster.