Why does a single missile cost $80k? Is it just profiteering or is there an actual technical/economic reason for this?
This isn't a fragile laptop you can sell a hundred million copies of.
I wouldn't be surprised if the 'low volume' part of that statement can be revised a year from now.
They're being used up at a pretty fair clip right now; Ukraine has been provided 7,000 so far, which is about a third of the US inventory. Current production volume is only sufficient to replace those in a bit over a year, which is giving a few folks the heebie-jeebies, and who knows how many more will be sent, and whether there will be a push to enlarge the US stockpile.
So, current production volume is almost certain to be ramped up to replenish the inventory faster, and I would bet that at the same time production capacity will be modernized and ramped up even more (since it looks like when you need to use them, you may need to use them a lot), leading to the per-missile cost coming down, which of course may mean that they'll be handed out and deployed even more widely and used up even faster (per the Jevons paradox), which at some point even starts shortening the design iterations.
There's a reason the Commodore 64 remains the most popular personal computer ever sold: There was really only one model, and it sold millions. Even though the market has expanded hugely since the 80s, it has also fragmented, and no personal computer, not even the original iMacs, has ever reached that level of popularity since.
So, given that the production volume for Javelins may be doubled (at least, and potentially much more if they start being exported more freely), it isn't unreasonable to expect some cost savings.
They aren't cheap, but they aren't expensive for what they are.
If your goal is to just inflict large cost upon the occupiers until they get bored after a decade then ATGMs are great. If your plan involves routing them with a counter-attack ASAP you'll probably find them lacking.
Unit costs tend to be estimated by dividing the total amount of cash spent on a weapons program and divide it per deployment unit.
Let's put it this way: if you invest a million bucks on a R&D program for a donut factory and in the end you close shop after getting a single donut out of the program, that would mean that donut cost 1 million bucks to make.
In any case, the processes and diligence required are extensive which creates a huge barrier of entry. It creates an environment where it's much easier to go with a company that knows all of your requirements and compliance needs.
Like everyone, these procurement people are probably overworked and would rather pay a bit more to reduce their risk and handholding required. They're more likely to be personally blamed for failure of a new vendor, than one that's been used extensively. The upside is limited for them.
Is this bad? Well... allowing silicon valley to self-drive cars is about my upper tolerance. Space-X is getting there, so they're a good example of what's needed to succeed in that space.