AFAIK the "manufacture a new molecule" part, or even the "find a molecule that looks likely to treat a disease" part, is in practice not even the main expense the patent secures; it's the "convince the FDA that the molecule is safe and effective" part, which is less engineering and more procedural/bureaucratic, and it takes the most time and money (after marketing expenses).
The FDA part is indeed expensive, and is a major part of the cost in what I do too (medical devices, hardware), but often these molecules are the result of decades of research. It's not really the case, the FDA part is hard, and takes years, but the molecule part often is even harder and takes even longer, costing even more.
I'm not trying to say the "molecule" part is easy (heck no!) just that the business of drug patents is another kind of ballgame entirely. Also, just about any innovation in any technical field has decades of research behind it, no? (I have worked at a medical devices company as well, but a few steps removed from anything that poked a human.)
The design cycle for hardware products is typically 6 months to 2 years maximum. Moat consumer products get refreshed pr redesigned every year, typically to coincide with CES and other things like the holiday rush. Same with the toy industry, there is a big annual show that you need to release your toy at or you wait to next year. In medical devices it's a bit longer, 2-3 years is common.
If you take much longer than that to bring the product to market you have to start redesigning the stuff you designed first. Components go obsolete, computers and displays change every year, software needs to be patched. It's why programs like the f35 and f22 are such messes, they were designed over a decade.
Most hardware products we interact with daily took less than 3 years to design and bring to market, they are not F22s. Some things take longer like an MRI machine but they typically are done in a iterative way so that proved versions of a product are brought out every year or two.
