While that is often true in software it is very seldom true in hardware. Think hard drives. The first HDD with HAMR (using lasers to heat up the platter to make it easier to isolate and flip a bit) required many millions of dollars to figure out how to do it. And when they started, no one knew if it was possible. Once they knew it was possible, creating a factory to build them was simple.
Building a fab costs ~$8bn now, does Intel spend that in research? They report (like everyone) R&D, but I bet the majority is D. Delivering a product is very different from making a component (like a platter) - the safety testing, reliability engineering and all the other elements of development are expensive and labour intensive.
You are thinking "expensive" is equal to "more money". Expensive, in this context, means far more than money. For example, risk. Sometimes company-ending risk.
I said "R" is expensive because it is often packaged with uncertainty. You have no idea if something is going to work or if people will even be receptive to the product.
Take SpaceX's landing rockets as an example. Hard and very expensive research. Lots of details nobody will ever know about. Lots of failures before success. And then, even with success, you have to make it repeatable and reliable.
One the "R" is done, "D" or building and scaling is, in relative terms, cheap.
Sure, it can cost billions to tilt-up a factory to build hard drives. Yet, you have a product and you have a market. You have an opportunity. You didn't have any of that before.
Wright Brothers: Hard and expensive R&D requiring YEARS of dedication. Once everyone knew you could build such planes (and how to do it) the "D" was cheap because it was done.
Sikorsky and helicopters is another AMAZING example of that. Helicopters got launched because of the Sikorsky's hard work and the unbelievable involvement of none other than Sergei Rachmaninoff (as in the pianist and composer) taking a leap and investing in the "R" that allowed Sikorsky to complete his work [0].
"R" is very expensive, in more than financial terms, that's why the Chinese copy rather than innovate.
That we are allowing them to get away with it is, well, disturbing.
"I think R cost is low, creating a poc is about 100th of the cost of a product (a proper product) but the risk of R is huge and therefore it's very hard to sustain and justify."
Which I think you have provided some good supporting statements for. I must really be missing something here.. Not sure what though!
