So either chess.com is wrong, despite having a pretty good ability to detect cheating, or Hans is lying. I’m inclined to believe the latter.
Also, if chess.com had all this evidence why did they wait until after Magnus’s cryptic tweet to act on it? Do they only run their algorithms on games after Magnus makes cryptic comments on players playing the games?
The far more obvious reason is that they took these actions because they don’t want to potentially lose Magnus and his company’s involvement in the tournament.
That being said I do think it’s unlikely he cheated in Sinquefield. I should stress that OTB cheating and online cheating are indeed psychologically different. One can be a momentary lapse in judgement which can be a slippery slope, the other requires extensive pre-meditation and is more likely to be detected.