Should multiple displays be a release-blocking criterion? I can imagine a desire to ship a new release on-schedule because of significant, content that works fine, but with multiple display support deferred for an update because upstream problems are likely to take a while to fix. I am uncomfortable with the notion that a release must wait for all desktops to work with multiple displays on erery architecture. It is desirable, yes, but this sounds like a QA attempt to control development.
Your criterion should be limited to hardware that works in single display configurations.
I think two displays is a reasonable limitation. Even with that, test coverage will be extremely sparse - there are just too many devices to make tests of different combinations practical. (If we cannot test it, we should not claim a release meets a criterion.)
If I can install three graphics adapters in a machine, and each supports four displays, would you require that I can run 12 displays on each desktop? Lovely if it works, but too rare a use case to be a blocker, or even to test on a regular basis.
Slots capable of driving a graphics adapter are very limited, but what about USB devices? With a few hubs, I could connect scores of displays, and your criterion asserts they will work.