In the late 90s the US shifted its missile defense strategy to stopping small numbers of missiles, either fired from rogue nations like North Korea or accidenta

In the late 90s the US shifted its missile defense strategy to stopping small numbers of missiles, either fired from rogue nations like North Korea or accidental launches from more established powers like Russia. Defense against all out ICMB strikes from Russia or China was explicitly stated to not be a goal, as it could potentially jeopardize the status quo of mutually assured destruction and lead to increased resistance in nuclear disarmament talks. For the mission for which GBI was designed to do, it performs adequately. While critics often cite an average lifetime success rate of about 50%, they fail to mention that there were lots of failures early on in the project and that over time the success rate has steadily improved. Further, for the intended mission, launching multiple interceptors at a single target to ensure a kill is perfectly reasonable. Preventing our destruction by large numbers of missiles is the job of our massive nuclear deterrent. The Arrow 3 system was made in partnership with the very same people who made the GBI. While their specifications are slightly different as they are expected to deal with different threat vectors, they're using the same technology. The Arrow 3 isn't going to perform any better than the GBI will. Further, the Arrow 3 couldn't work as a drop in replacement - it would need new launch facilities built in new locations, new radar systems and control infrastructure, and of course a whole separate supply of spare parts and maintenance equipment. If we want the GMD program to defend against larger numbers of missiles, it makes far more sense to just build more GBI missiles, but again, there is no compelling reason to believe that we should want GMD to defend against more missiles than it currently does. The only legitimate concern is that GBI was not designed to deal with hypersonic missiles. Arrow 3 is also not capable of this. For the moment, only Russia and China might reasonably be expected to have such capabilities, and they are deterred by the threat of nuclear retaliation. The MDA is already working on anti-hypersonics, but the threats we face now and in the near future do not require it, and there is no pressing urgency to shift more resources to its development than are already allocated.