A theory, sure. Another thing to remember about biological systems is that complex features like sleep rarely serve just one purpose. There will generally be one cause that was killing lots of members of a population which the feature was developed to mitigate, but other "reasons" may have grown over time as evolution "discovered" other ways the feature can help.

I think in general, good-sounding theories produced by the "speculative evolutionary biology" game are actually very often right -- but only describe one facet of a supremely messy system with boundaries that don't conform cleanly to the mental models we try to impose.