Much as I like SpaceX's innovation and Musk's vision, I think I can relate to how the air force guys must be feeling about this. This is basically like being told by the beancounters that you can't get more of these shiny 100%-reliable Unix/Mac workstations that you are used to, and have to make do with virtualized Windows clients, because while slightly inferior, their three times lower price more than makes up for it.
Welcome to the commodisation of Space.
You might have plug-and-play appliances in your office providing local storage, VPN services or edge filing, you might even get some shiny Apple Macs for specific workloads. But your datacentre is going to be full of disposable whitebox servers because going hyperscale with HPE or Dell isn't necessary or affordable for your run-of-the-mill heavy lifting. It's cheaper to have spare boxes to spin up whilst you repair/replace failed servers than to pay for a branded box with additional redundancy.
Shiny branded boxes with expensive software licenses and lovely management tools become relegated to special workloads which can't be spun out to a cloudy cluster.
And so it is with space. For small satellites, anything that you're serially producing and can afford to stuff "+1" on the end of the production run because you lost a launch, you go with the ludicrously cheap option.
Consider - if you're at NASA building a one-off probe or bespoke bit of kit, you need to know that it's going up because you only get one shot. Likewise if it's a uniquely huge bit of kit that needs a heavy launcher. You pay the premium.
If you're building a dozen identical commsats, a string of common-framework surveillance sats or a global positioning constellation, you've got a production line going, and if you lose a launch, you just get the insurance and add one or two more satellites to your order, which is more than covered by the fact you're paying 1/3 on the launch.
Sure, it's nice to have the fancy workstation, but the world moves on...