Back in the day we had 1,000 physical servers to run a large scale web app. 90% of that capacity was used only for two months. So we had to buy 900 servers just to make most of our money over two events in two seasons.
We also had to have 900 servers because even one beefy machine has bandwidth and latency limits. Your network switch simply can't pump more than a set amount of traffic through its backplane or your NICs, and the OS may have piss-poor packet performance too. Lots of smaller machines allow easier scaling of network load.
But you can't just buy 900 servers. You always need more capacity, so you have to predict what your peak load will be, and buy for that. And you have to do it well in advance because it takes a long time to build and ship 900 servers and then assemble them, run burn-in, replace the duds, and prep the OS, firmware, software. And you have to do this every 3 years (minimum) because old hardware gets obsolete and slow, hardware dies, disks die, support contracts expire. But not all at once, because who knows what logistics problems you'd run into and possibly not get all the machines in time to make your projected peak load.
If back then you told me I could turn on 900 servers for 1 month and then turn them off, no planning, no 3 year capital outlay, no assembly, burn in, software configuration, hardware repair, etc etc, I'd call you crazy. Hosting providers existed but nobody could just give you 900 servers in an hour, nobody had that capacity.
And by the way: cloud prices are retail prices. Get on a savings plan or reserve some instances and the cost can be half. Spot instances are a quarter or less the price. Serverless is pennies on the dollar with no management overhead.
If you don't want to learn new things, buy one big server. I just pray it doesn't go down for you, as it can take up to several days for some cloud vendors to get some hardware classes in some regions. And I pray you were doing daily disk snapshots, and can get your dead disks replaced quickly.