If you picked A, you assumed the old order would stay the order. Instead the money went where performance gains were clearest. That means a company with large fabs can be sidelined by investors if it is not central to the AI compute stack.