If the average number of sexual partners in a community is 5, then a lot of links exist for the virus to flow through. But the links are of much "higher quality" if the partners are simultaneous, because in the serial case, a lot of paths through the graph are, when you look at them closer, made impossible by the chronology. If Marie sleeps with Jim, and then Jim sleeps with Agnes, and Agnes carries HIV, the virus can't flow from Agnes to Marie (along what appears to be a valid path in the sexual network) unless Jim goes back and sleeps with Marie again.
So it's not just about lifetime sexual partners, though it's certainly true that high average lifetime sexual partners contributes to spread.
is a man with many casual dates / FWBs / one night stands, not a man with two long term partners.
Note that one-night stands can never, by definition, be simultaneous with each other, though a succession of one-night stands can each be simultaneous with a long-term girlfriend.
The difference is that Westerners generally form small, closed sexual networks, while Africans participate in a gigantic sexual network including most of the local sexually active population.
The term of art here is the fairly awkward "multiple simultaneous sexual partners", but that concept isn't quite enough - a typical Westerner with multiple simultaneous sexual partners might be a man with a wife and a mistress, but when neither of them is cheating on him, that's still just a tiny, closed three-person network. But in Africa, the typical case is that most people have multiple simultaneous sexual partners, so even with the low transmission rate from heterosexual intercourse, the paths still exist for the virus to flow from one person in the society to (almost) anyone else.
(I don't think mettamage deserved to be downvoted. I specified gay sex for a reason, but the followup question "why not heterosexual sex?" is pretty obvious. It has a lower transmission rate, but that's a minor issue compared to the behavioral difference.)