Depends how elastic the supply is. If those 1 million units, what if only 50k turn over each year? So 50k vs 55k units available could have a doubt digit influence on price.
You should't assume that airbnb conversions would turn over 20x as frequently as the existing apartments. So it wouldn't 50k vs. 55k, but 50000 vs. 50250.
Sure, there might be a big spike in supply on airbnb-banishment-day, but then those apartments get rented and we're back to approximately the place we started, as far as long-term rents go, only without airbnb the tourists and property owners are worse off.
