When we first research red belts they're great. They solve throughput issues without having to worry about the actual issues, and belt weaving is much less complicated with only one belt. But eventually, we get past that, and at that point, at a high level of play, I wonder what use cases red belts have. Obviously, you would eventually want to use them in a big enough base, but let's assume reasonably small bases (pre-rocket).
Yellow belts are more than twice as cheap, and space is not a huge issue, which would make it seem like red belts are just the clear loser since, well, there's not much else going for a belt. You could argue that cost isn't a big issue either, but it turns out it is, as much as people would like to ignore it. For example, in my deathworld marathon playthrough (currently at utility science), yellow belts account for like 5% of all iron ever produced. Red belts are 5x more expensive for the same amount of throughput... I think the implications are clear.
However, there are cases where red belts are actually cheaper as pointed out by Putin's Chest here on the discord. For example, using one red belt for steel is cheaper than two yellow belts since the steel that has to pass over the belt is simply so expensive. Having more belts mean a bigger buffer, which ultimately results in more frontend expense. So it turns out there are use cases for red belts. And in resource-rich worlds maybe using them in some places and not others is a pretty feasible strategy as well. So what are the use cases of red belts?