Radiation hardening is a requirement to protect the arsenal from EMP attacks in a large scale conflict with a nuclear component. This will hopefully be a rare or nonexistent occurrence, and hardening should not be considered a requirement for bombs used on Al Qaeda and IS (and nearly all other expected enemies). There is a place for radiation hardened munitions in the arsenal for deterrance, and there is also a place for some that are not, and these should be the ones we drop in "low intensity" conflicts, with the occasional release of a hardened munition to act as a quality control test.

Reliability is obviously a concern, but the procurement process should force contractors to open their designs to qualified competitors after a certain period of monopoly profits to recoup their research expenses, sort of like the same way that brand name drugs are eventually forced to compete with generics. Just because these designs are generally national security secrets doesn't mean that the original developer should own the equivalent of an infinite patent on the technology.

Even if you do that and build one. You'll need to spend many many times over that budget lobbying, making friends, getting noticed and networking, formalizing through red tape and so.

Imagine today you have that device in your table. Ok. What is your next step. Just show up at the Pentagon with it in your bag?

A lot of that costs money (unfairly probably) and that is one of the reasons soldiers sometimes end with crappy hardware that costs some astronomical amounts.

Sigh. I should have pointed out that I used to be an Air Force engineer. Although my post was a massive oversimplification, it was not for any of the reasons you pointed out. No, dev time would not be more than 1 engineer, at least not for the bomb guidance part. Most bombs that are dropped do not need massive accuracy - anywhere in a 10m radius will do the job nicely. There ate a few targets where you want 1m accuracy, but that's the exception. A phone 'a sensors should be able to do the job sufficiently well.

No, where you're really going to get burned in dev is hooking the bomb up to the aircraft's targeting system. That's a world of pain right there. The best solution might probably involve making a Bluetooth widget that hooks up to the aircraft and then talks to the bomb via radio. They could be pre-paired in the factory. Not ideal though. You're also going to need to spend a ridiculous sum of money on sales, just to get your foot in the door.