What if the non-coincident loads actually are "both of the following" ? Say, heating and cooling that are unlikely to be used simultaneously, and two EVSE branch circuits controlled by an EMS. All on the same feeder.
Well, if there's an EMS, for the EVSEs we don't need to consort with 2026 NEC First Draft 120.6. A different sections covers that. That leaves the heating and cooling, which would be on different branch circuits, so (2) would not apply. And if the heating and cooling are unlikely to be in use simultaneously, then (1) applies, and they are non-coincident. So we use 120.6 for the heating and cooling.
Here's a different example, maybe what you are driving at: there's one branch circuit that supplies two EVSEs that are mechanically interlocked so that only one can be powered at once (maybe they have different incompatible plugs for different vehicles, like one is for the original Tesla Roadster). Then the EVSEs are non-coincident under (2). And the heating and cooling are non-coincident under (1). All is well, no problem with the text of (1) and (2).
Of course that raises the question of how you deal with two different sets of non-coincident loads, but the 120.6 language is broad enough that you can just do the obvious thing: consider 4 different cases, heating and EVSE1, heating and EVSE2, cooling and EVSE1, and cooling and EVSE2.
For the case of two EVSEs, you don't have to resort to that complexity, all you need to do is compare them and use the larger rating, then do the usual thing for heating vs cooling. But we can certainly come up with other examples with other types of loads where you really do need to consider all 4 cases.
Cheers, Wayne