Fans are cheap. Realistically. Mass produced. Each magnet used is not identical in ability, different resistance in the coils (extremely slight) but all of that adds up to being a ±% for each fan. A fan rated at 1800rpm is really rated at 1800±% rpm. So you could see anything from @1750-1850rpm on average. With voltage or pwm signal being consistent between 2 fans, that means at any given setting you could easily see upto @ 100rpm differences, even if they were identical fans on a splitter. You could have 1800rpm fans on a splitter read 1750rpm. Swap the fan leads (splitters only read 1 fan) and suddenly that header with no other change is now reading 1850rpm.
Its not something to worry about unless there's a serious difference, like 1800rpm fan at 100% duty cycle reading 200rpm.
It's one of the reasons I prefer fans grouped in splitters, so all 3 fans on intake are 1 header, both top exhausts are 1 header etc. That way I assume all 3 intakes are the same speed ±, all top exhausts are the same etc. You'll know if individual fans have an issue, big difference in temps or visible rotation differences etc.
It also means just 1 control temp, motherboard temp sensors are at variable points around the board, it's easier to turn up/down all the fans in a group with just 1 curve, and not have different sensors controlling adjacent fans. Also frees up a lot of headers.
With an aio, you only need 3 headers. Cpu_fan for the pump or fans, sys_fan1 for pump or fans, sys_fan2 for fans. Depending on the AIO.