With respect: that's exactly the kind of glib, drive-by analysis I'm talking about. What you're saying simply isn't true. People have been testing composites in pressure chambers and every load imaginable for decades.
Yes, clearly they don't act like steel. But saying that it's impossible is just too easy.
Just watch. It's going to be a routine design flaw, like they added a joint or a wiring hole in a bad place and didn't re-run the model.
(Edit because this was misunderstood: I'm not saying what the failure is. I'm saying that we DON'T KNOW what the failure is, and that given the choice between makes-me-feel-smart pontification about materials and a routine design flaw in a prototype device, I'm going to bet on the routine failure every single time.)