I think this mixes up low odds of success with risk, the latter rather being an expectation value than a probability (i.e. probability of success or failure times the value of the things at stake). Risk can be low even when the odds of success are not good just by nothing being really at stake.

Think about their rocket landing program: the probability of landing a rocket on a boat in the first attempt might have been low but the risk was almost negligible because the launches that SpaceX practiced landing on had been paid for by a customer already. Now, if SpaceX had bet the company on reusability by paying for all these rockets out of their own pocket instead and hoping they manage to get it right before cash runs out, then this could have been called risky.