Hypersonic flight environments are very, very difficult to replicate in wind tunnels. We can replicate portions of the environment, but due to the highly coupled nature of the responses seen by flight vehicles at these speeds, there's always a tradeoff. To really understand what's going on, you need high enthalpy, high pressure, high Mach number flow for a reasonably long period of time. Experimental scaling is only so helpful depending on what you're trying to investigate. For a fully coupled response, you basically end up with a scaling ratio of 1 since the different physics clash when trying to scale the many parameters you need to achieve similarity. Realistically, flight test is the only way to truly get the conditions and response you need, but that's very expensive and risky and you can't build a huge factor of safety into the vehicle or it won't be able to meet the needed performance. Instead, we use a combination of simulation tools and targeted experiments to try and reduce the risk/cost for flight tests.

Vietnam? The belief that low, slow and stealthy would win the day vs high and fast? Look at how many people claim that SAMs won out against the B-70 Valkyrie (basically, a heavy bomber analogue to the SR-71).

More likely high and fast is the answer, it's just ICBM's are faster and go higher than hypersonic rocket planes...