I work on servers almost every day.
These heatsinks are meant to be in a case, which usually has 2-8 fans per RU on the front, and 0-8 on the rear along with power supplies loaded with fans . The heatainks themselves sometimes have fans as well. These are extremely thick high backpressure and and very short fans at 40mm or so and most run at extreme RPMs.. Their reliability is below that of an 80-120mm brand name fan by a lot. The servers generally run at 80-100% load at all times, because an underutilized server is like throwing 20k away day one and a hundred bucks daily after that for space, bandwidth, cooling, and electricity.
Most datacenters push a wall of chilled air into these things (cold side) and isolate and evacuate heat on the hot isle.
Servers run five 9s (99.999%) uptime only because they have redundancy, meaning you can lose about 30-50% of a device and it keeps working. Everything can get hotswapped with the exception of CPUs normally.
They may work if you constructed a large ridgid channel for a 6" or 8" inline to pull air through them at full blast - but if you apply a 50W load to them and they just have 1 fan it's likely you will have a very poor and load cooler. If you apply one of them to a desktop processor it will overheat or thermally throttle.
We have thousands of these for free, and I opted not to use them. They would probably passively support a 20-35W load just because they are a hunk of metal.