That bit about Aquinas denying the intelligibility of prime matter even in the light of God, is quite interesting. I'll use it as an indirect point of departure for my response.
The three "models" of creation that I'm familiar with are (A) creation "from nothing," (B) creation by emanation, and (C) creation by "organizing" pre-existent substance (which substance is not itself emanated). There's an LDS author, Brandon Sanderson, who penned a character in one of his novels as entertaining the thought that (A) and (C) can be had to coincide, but generally those two options are held to be quite distinct.
At its most extreme, exnihilation is the (internally contingent) creation even of "pure possibilities." Let us imagine a grid of variables, at the center of which is a single ultimate constant, 𝕲 (for God). Style creation as generically assigning a value to a variable (c.f. Quine's "to be is to be the value of a bound variable"). There is no other grid that we are working with in the background; this is the grid of all other grids. There is no source for assignments to the variables, besides what is available to God at the heart of this grid.
Per (B), then, we can also imagine 𝕲 itself as coding over a subgrid of Its own, let's say {X, Y, Z}. A popular pre/non-Trinitarian attempt to exalt Jesus (and the Holy Spirit, but less often...) is to hold that He was emanated from the Father, hence created modulo (B), but that 𝕲's creation of other things was either exnihilation (or maybe (C)) or "incomplete" in a way that the emanation of the Son was not. At any rate, let us hold Y = 𝕲, and say that God can quasi-create Itself, by emanation, in the sense of assigning values to X and Z from 𝕲. And then too, or instead, 𝕲 can create things through/in the rest of the universal grid, in some like way.
Alternatively, if the grid is pre-existent matter "surrounding" God (we're close to the relevance to defining "prime matter"), then maybe God can determine some variables according to other ones + 𝕲's own essence and character. By contrast, (A) requires that God decide even what the possible values of all the variables are. How could that be?
Worse, some extremists about divine sovereignty will go on to imply that not only are all the items imaged in the grid created from nothing, but that God created the grid itself from no grid at all, and decided that it would be a fact that things exist in terms of this grid. But leave those impious fanatics aside for a moment. If at least the grid in a way exists generally, no matter how 𝕲 has decided to determine the Xs, Ys, and Zs thereof, then this grid is functionally equivalent to the concept of prime matter: per the Aquinas bit, beyond 𝕲's understanding not by being a truth beyond 𝕲 so much as not sustaining (without contradiction) the predicate of intrinsic intelligibility at all. 𝕲 does not "understand" prime matter because there is nothing to understand: to know something, and to appreciate the ramifications/significance of this knowledge, there must be something to be known, but prime matter is as much as an array of question-functions so vaporous that God Itself does not need to care what the answers to said functions are, for there are no answers as such (the closest thing being, "Does X, or Y, or Z, exist?" which then for God is decided by the power of creation).
Addendum: a set-theoretic look at the matter
Suppose that "is closed under the predicate 'is open under all other predicates'" is meaningful. For sets, this would mean a couniversal set, I suppose. Now, I often wondered how prime matter could never be defined by a specific property (I think this was something like Locke's contention, or even Aquinas', among others), but is that to which properties variously "attach," since "is that to which properties attach" seems propertyish (at least). But so take the following list:
- Is closed under the predicate "is closed under all other predicates"
- Is closed under the predicate "is open under all other predicates"
- Is open under the predicate "is closed under all other predicates"
- Is open under the predicate "is open under all other predicates"
The first and fourth options defeat the point of the "all other" moment of quantification, I would think. However, if there is a set that is open on all terms aside from its own general status, and this is not pointless, could there not be elements that reciprocate this relation, and these would be prime matter? For they would not have any defining property other than having no other defining properties; they might be predicated of sets in a neo-Humean way, maybe (I tend to interpret Humean bundles as sets of tropes, but maybe that's just me!), or then properties would calcify, in sets, over these generic elements, perhaps. In other words, these elements would have only the fairly insubstantial property of being themselves as such, but when they were included in sets with more vivid tropes, they'd "glue" an object's qualities together, forming the object.