First, let the record show that I loathe Trump, think he's a lying snake, and "the wall" is a particularly stupid idea.
However, Dave's analysis does show something interesting: that, GIVEN that you are building the wall, adding solar could be in the ballpark of being able to defray the costs slightly.
I spotted two issues with Dave's analysis. First, and less important, he used TMY data from a coastal location, so I suspect that will show lower insolation overall and probably lower peak temps than a more inland reference. Since those impacts confound each other, it'd be hard to say which dominates without digging into am or detailed analysis, but I'd wager a real system would be a bit better.
Much more serious, The $1.50/W is for utility scale systems which are almost ALWAYS ground-mounted and generally layed out "compactly" to reduce the amount of wiring and the distance to inverters. Of course, a wall-based system would be spread out over 2000 km, which means inverters every 1/2 km or so, and appropriate DC wiring AND AC wiring to what would likely have to be multiple transmission interconnections. Oh, and those would be interconnections to transmission that does not exist yet. The NREL numbers for a commercial (not utility) system are probably more realistic for this type of installation, plus the cost of transmission, which is generally not part of a commercial system's price. I would not be surprised if such a system was 3-4 times as expensive as a state-of-the-art ground-based utility system.
That, I think, blows the PV economics out of the water. It is not viable in its own right, even if the wall were not counted.
So:
Wall: stupid
Wall + PV: even more stupid
Sad.