I would be happy to join the crowd seeing this as yet another artificial scarcity attack. Or at least attribute it to the lack of thinking.
But in this case the explanation is well known and documented, and it is neither of the above. It’s a reasonable decision, which didn’t age well. The person responsible,
Dave Plummer, gave the details.
Unlike with other fake limitations, Microsoft never prevented users from creating bigger FAT32s even with their official tools – only that single GUI tool got affected. Never profited from it. Never used it against anybody. By the time anybody asked, the limit became a de facto standard. By the time it became relevant to wider audience, memory cards were using exFAT.
Was the decision bad? No. Historian’s fallacy applies. And Plummer was still very cautious and generous setting the value to 32 GB. Remember this is 1995. Your internal HDD was 250 MB to 1 GB, which you would further divide into partitions. That was a time, when I sent a screenshot of a program reporting CPU to be 1000 MHz as a joke to a local computer magazine. And I would die from laugher, if anybody told me about 32 GB flash storage.
Was the decision invalid? Yes, I think so. The limit should have been 2
28 · 512, which is 128 GB, not 32 GB. Unfortunately Plummer no longer recalled his thought process when recording the video.