As today has been a day for analogies:
dog : proximal environment :: human : Reality
A dog's ontology is presumably quite limited. A human's ontology is apparently maximal. While our average intelligence is quite a bit more than it is for dogs, it's not "godly" by comparison (maybe "demi-godly" but not godly, since for instance we can't create dogs from raw materials right now). But our ontology is maximal. If the analogy should carry over, as it seems it should, as in
(dog, proximal environment) : (human, Reality) :: (human, Reality) : (superintelligence, ?)
something should presumably have to replace the question-mark against all logical possibility. And then even if something did replace it against all logical possibility, humans are at most right now merely demigods compared to dogs.
Might this put a substantial hitch on some theoretical maximum of intelligence?
Update 1 2014-10-17: While no one has pointed this out yet, during my discussions so far I realized that the frame of mind at around the time of generating this curiosity could readily concede to "The analogy breaks down if superintelligence (including up to hyperintelligence) necessarily must originate from at least as low as human-level intelligence." ... As a corollary, this doesn't bode well for traditional conceptions of God.
Update 2 2014-10-17: Responder @ChrisSunami points out that a hunter-gatherer tribe 10 thousand years ago would not have a maximal ontology [in the relevant sense]. Since the main analogy broke down from Update 1, and in light of this hunter-gatherer point, a new analogy becomes forthcoming.
(early human, has no maximal ontology) : (modern human, has a maximal ontology) :: (modern human, standard intelligence) : (derivable descendent "human", superintelligence)
This analogy happens to have the bonus side-effect that, as most arguments so far have been based on the magnitudes of ontological contents rather than the magnitudes of ontological expanses, it now accommodates this concern by having built into it the altogether irrelevance of the magnitudes of ontological contents.
So, to answer the original question, there's nothing to worry about. In comparing modern humans to dogs or to early humans with tight extremophilic analogies, ultimately no great filter to superintelligence or beyond is implied.