I'm reminded of the first calls (90s->2002, not the later programs) for bids to replace the KC-135. There were only two real contenders, Airbus and Boeing. Airbus's first reaction was essentially: "why even bother."
A primary reason why payloads are so expensive is because launch services - used to be - so expensive. With lowering the launch price payloads can become cheaper.
You're right that tomorrow, a different one could be built with the new launch cost in mind. The entire "new space" movement revolves around lowering launch costs and building much cheaper satellites to take advantage of lower launch costs.
A computer that is substantially cheaper would be hobbyist-grade computer.
Of course I'm only talking about traditional military rather than forces like A* Qda which might make do with whatever they can get their hands on, which is probably what you're referring to.
I suppose a very cheap satellite could be terrorist-grade. ;)
Just assume that military is the standard, and in order to qualify as consumer, you need to add extra features.
So far your argument is symmetric.
Yes, you may always want a unique, expensive, cutting edge satellite. But you won't always need that - military doesn't directly set the goal to spend as much as possible. X-37 demonstrates a lot of capabilities while being much cheaper than previous reusable space vehicle. So with launch costs going down other, cheaper opportunities open - to everybody, including military. That's beside the idea that cheaper rocket doesn't mean worse rocket, reminding Japanese cars of few decades back.