It doesn't have to be bullshit. You have a sample size of 1 and realistically we can't expect 100% accuracy. We also don't know how often your IP is changing - maybe it was a proxy endpoint recently?
It's a static address that hasn't changed for 3+ years. The only thing I can think of the makes it proxy-ish, is that I have a 6 person household here.
Edit- The IP my employer uses as its outbound nat IP that we vpn from while WFH (so it actually is a proxy) gets a clean bill of health while our corporate email server (not a proxy at all) is tagged as one.
It's not just IPs that have reputation scores but IP ranges and AS entities. Residential IP ranges are often dinged by default because of the risk of infected home PCs sending spam.
There absolutely is a reputation associated with individual addresses and ranges of addresses. And that reputation can be derived from numerous legitimate means (honeypot activity, spam, reverse dns info), but whatever these folks are doing regarding "proxy/vpn detection" is indistinguishable from snake oil.
