If the risk to such a thing failing is so large, this seems a point against your argument. We MUST be able to reduce the risk of such systems failing, and do so

If the risk to such a thing failing is so large, this seems a point against your argument. We MUST be able to reduce the risk of such systems failing, and do so provably. If we cannot reduce the cost, we must reduce the risk, such that the calculation makes the system affordable again.