To give a less mathematical response, it is the latter. For a QM entity to be in an exact place with an exact momentum is just not something that is even defined, in QM. The uncertainty is intrinsic to nature.
From reading Carlo Rovelli's "Helgoland", it seems the "relational interpretation" of QM in effect denies that a QM entity even has any properties on a continuous basis.
According to this view, all properties become manifest only in the course of interactions involving the QM system in question. In between interactions, it is not necessary to assume that it has any properties at all. The wave function (when the QM operator for a property is applied to it) tells you what range of values of the property the system may manifest when it interacts.
Tying properties to interactions gets rid of the tiresome issue of "the observer", which has led all sorts of people astray over the years, even sometimes to the extent of speculating that QM gives a special place to conscious observers. This last is something that has spawned an entire industry of quantum woo (Deepak Chopra et.al.). But, on the contrary, any "observation" necessarily requires an interaction. It is the interaction that counts, not an act of observation.
The further implication of this is that what we call "reality" is made up of the interactions going on all the time between QM entities. Which is not as crazy as it sounds. After all, something only provides evidence that it exists when it interacts with something else.
Classically, we interpolate between interactions by assuming that objects possess properties with defined values all the time, in a continuous manner. But in QM it seems the only thing that unambiguously persists is the wave function, which represents the potential properties exhibited when an interaction "collapses" it.
This, at least is my understanding. I find it elegant, as it seems to resolve a number of the paradoxes that QM throws up.