Do development costs really need to be recovered through sales? I thought most of this research was done on federal grants.

The early research is done on federal grants. What costs real money is late-stage development and especially clinical research, startups and universities are not at all equipped for that.

The financing model is problematic. Return on investment is much higher and faster in IT, besides there is a mismatch between what the market favours and where future clinical needs are (the antibiotic crisis, neglected tropical diseases).

Isn’t the solution then to build a national infrastructure for late stage trials as opposed to putting life-saving drugs behind a ludicrous paywall, whose price is determined by the opaque malfeasance of the healthcare system in the US?

Late-stage trials are enormously costly and negative results are common. To publicly fund them, we'd have to be ok with governments spending billions and getting nothing for it. And after the fact there'd be a long line of people saying how $treatment was obviously never going to work and only an idiot would ever have let it go forward for stage 3 trials.

I'm not opposed to the idea, but I think you'd struggle to make it politically achievable.

The government mostly funds basic research, but then there is a lot of work to turn promising leads into products. Googling a bit, in 2017 the pharma companies spent $70 billion on research while the NIH spent $38 billion.

> I thought most of this research was done on federal grants

And if not, why not? There may be folks who think public research grants shouldn't exist, but since they do, what else would be worthy of such support if not public health research?