A. a sentient civilization, living in a virtual universe with unlimited possibility, and

B. a completely automated outer reality, keeping the sim running, of which the sentients have (and need) no awareness.

Part of the sentients' remembered history would be that that "we used to live in a much more limited universe, but then we ascended to this one," or something like that. The "reality" that the automated system inhabits would be no more real to the sentients than your fingertip is to your IDE. (And would have as little realization of an orbital bombardment against the automation as your IDE has of your fingertip holding down the power button of your computer.)

The automated system would want only to grow and maintain itself, and would not have the higher-level thinking capacity required to formulate larger goals. It would be more like a plant, having the same sort of relationship to the sentients that we have to bacteria that live in our bloodstreams. (I want to mention both Avatar and [End of] Evangelion here, for oddly-complementary visuals.)

But when you start to think like a post-Singularity sentient, you start to realize that their ethical system would be much "pickier" than ours. They would see leaving any of their own (whether biological or AI-based) in the "limited" reality as a punishment equal in scope to the suffering child of Omelas—that is to say, one they could not possibly stand for.

If you estimate that life at our stage of development is relatively common, but notice the stars are silent, you should then conclude that humanity is doomed. Either we're doomed to destroy ourselves (like 100% of other life forms at our stage) or we're doomed to be stuck in one solar system.

Note: I don't want to spend a bunch of time explaining my view on this, but I think that intelligent life is so rare that we're the only instance of it in our light cone.

So, a reasonable alternative solution: use the Anthropic argument. First, assume that there is intelligent life close to us. Then, realize that we still exist, and haven't been consumed by grey goo; and, according to our long-distance observations, all the planets around all the stars we've surveyed still exist too, in shapes that don't suggest intelligent redesign (even though "they" have had a few potential billion years to do so.)

Given that, it's likely that whatever path civilizations tend to follow eventually preclude them from wanting to leave to go redesign the galaxy in their image. This might just be a problem of limited extra utility derived from expansion: if Moore's exponential curve does, in fact, turn into Moore's S-curve, then there's no point in sticking more Computronium onto the pile, because the lightspeed-delay in the optical relays could just get to be too long for the extra circuits to do any good (presuming everyone wants to live on a single "instance" of AwesomeUniverse2000, rather than sharding it.)

Now, the number of additional assumptions in the conclusion can, by working back through Bayes' law, tell you how probable the assumption itself (that life exists close to us) is. I'm guessing it's a small, but not infinitesimal (compared to space) number.

Looks like you have worked out their business model then - unless you pay you have to share it with others, go for the upgrade and you get to be selective.