I used to run the FreeBSD box for sendmail.org. When I left that job in 2001 it had already been running for 2+ years.
Considering that the datacenter it was in is now the Dropbox office, I'm guessing it had to be shut down and moved at some point, but 2+ years seemed like a really long time even then!
I don't get the "even then" part of your statement. If anything, big uptimes are a thing of an older, more monolithic era when patching was less common. It was extraordinarily common to have multi-year uptimes on important servers, whereas these days they seem far less common
In my experience, the past decade has been a time of trying to be more rigorous than ever about regular patching, routine reboots, and respinning VMs to ensure that your provisioning systems work as intended, so you never again end up with these monolithic irreplaceable systems.
These days, the only time I find servers with big uptimes is when they've been neglected -- they're some old bastard child of some former employee or the ancient rickety crap some department is too afraid to touch... And even then, it doesn't raise an eyebrow until it's 1000+ days.
Because even 15 years ago it was rare for a server to have such a high uptime, since you usually had to unplug and move it every once in a while, and hardware still didn't last years and needed replacement.
Considering that the datacenter it was in is now the Dropbox office, I'm guessing it had to be shut down and moved at some point, but 2+ years seemed like a really long time even then!
FreeBSD is just really good at lasting forever.