To add to this, weather conditions on November 5 are likely to be more temperate than February 29. November 5 can definitely still have bad weather, but you would be making in-person voting overall objectively statistically more difficult by pushing it into the end of February thanks to snow, where November doesn't even crack the top 5 months for the most snowfall (and we're talking early November here).

You would also be giving some president/Congress ~~three-ish extra months~~ nine-ish fewer months of a term for literally no reason by shifting it all the way to February 29, unless you wanted to roll out some decades-long scheme to incrementally push it there.

Edit: that also leaves the fact that you would have (presumably the first day of) one of the most watched events on Earth taking place on the same day as the general election, meaning arguably one of the few times when minute-by-minute 24-hour news coverage is necessary for the election, you'd get it interspersed with a ton of Olympics stuff, and you'd likely also have a decent chunk of people staying home to watch the Olympics instead of voting, further depressing turnout in addition to possible weather issues.

And February 29 doesn't happen on years ending in 00 unless the year ends in 000.

That's not correct. Centuries are not leap years, unless they're divisible by 400 (not 1000). So, 2400 will be a leap year, but 3000 will not.

February isn’t in summer.

Yeah, it is.