Why would it be hard to test? We have our own anti-missile technology, so it's ostensibly as simple as not putting a payload on the missile, then launching it at your own test range.
The physical environment these weapons were designed for is extreme and only possible to simulate piecemeal.
Each stage needs to function in the presence of nearby nuclear detonations, resulting from both adversary and friendly weapons.
These detonations are expected to cause severe shock, thermal, radiation, and electromagnetic transients.
In the case of the most important targets, it is guaranteed that numerous detonations near the target, from ABM systems and friendly impacts, will occur, and these systems have been engineered and are expected to perform reliably under such conditions.
This weapon is an ICBM. The payloads are delivered to orbit then launched at the target from there. You're already facing severe shock, thermal, radiation and EM transients just to get to orbit. Once there, you're ultimately dropping MIRVs, the design of which is considerably simpler.
The delivery vehicle and the reentry/payload vehicle have entirely different life cycles and deployment concerns.
A key operational requirement for any fixed US ICBM is that you can launch during and after an enemy nuclear attack on your silo fields (silos dispersed and hardened, redundant command links, etc) and penetrate defended areas protected by nuclear-armed ABM systems (e.g. A-135/ABM-4 Gorgon). That means resistance to high radiation flux from nearby detonations at both launch and reentry, as well as the need to survive flying through the expected debris clouds kicked up by previous detonations.
If the enemy is using ICBMs you are very likely to get yours launched well before their weapons make the first impact. If that weren't true, then the high flux conditions do not last for a substantial period, so you're describing a problem that would occur if an enemy warhead hit at the precise moment yours was leaving the silo. Your enemy cannot possibly have this precision in timing.
> If the enemy is using ICBMs you are very likely to get yours launched well before their weapons make the first impact
This is a terrible assumption to make if you’re trying to deter nuclear war.
Unless any random outage or terrorist/conventional strike against one’s early-warning radars, or errant satellite launch by a low-grade nuclear power, is automatic grounds for a universal MAD offensive.
>Unless any random outage or terrorist/conventional strike against one’s early-warning radars, or errant satellite launch by a low-grade nuclear power, is automatic grounds for a universal MAD offensive.
They are, that's what Stanislav Petrov is famous for helping prevent. A single early warning satellite had an anomalous reading and everyone wanted to start nuclear Armageddon, only prevented because he didn't believe the US would send only a few warheads, to the point he was willing to bet his country on it.
Famously the US plans early on didn't really have any distinction as to who fired, and the only retaliatory option available was to go full send with everything on Soviet cities.
Specifically about satellite launches, yeah, that's a genuine concern, which is why they are talked about publicly, even when it's a classified spy satellite going up, why you are very loud and public about testing ballistic missiles anywhere you do so, and why both the US and USSR worked very very hard on making computers to quickly estimate the landing point of a ballistic launch.
Meanwhile the entire point of nuclear missile submarines is that ground based missiles are an explicit target of known enemies, and thus not likely to survive a first strike. They are not intended or planned to survive such a first strike, which is why we had SAC flying B52s 24/7 for like 40 years. Any missiles that aren't out of their tubes before enemy detonation are mostly assumed lost.
In fact, nuclear submarines have gone a long, long way to improve the situation, because now you don't have to rely on those ground based or air based warheads as much, so you can be more conservative in your judgement. If you're wrong and the soviets really did initiate nuclear war, oh well, Tridents will show them the error of their ways.
The biggest "eh, we should wait before we launch" cause is simply the lack of tension between world powers. Despite Russia's blustering, in official capacities they have not signaled that they are looking to launch nukes. They have not significantly increased their readiness, towards a large scale anti-NATO war, to the point that they are literally removing defenses from the border with Finland in order to divert those resources to invasion.
There has only been one "Oh shit oh god" moment that I know of; when an S300 missile (which I believe later turned out to be Ukrainian) landed in Poland and killed a farmer. There was a crisis meeting of NATO members.
Moreover, putting nuclear weapons in orbit would be a violation of the 1967 Outer Space Treaty. As an aside, Atlas and Titan were capable of reaching orbit, and were used for the Mercury and Gemini missions respectively. Minuteman, on the other hand, was not powerful enough to put a payload in orbit.
