Worth noting that in the US, probably more half the software development done gets to be called "R&D". We have very beneficial R&D Tax Credit (26 U.S. Code §41), and it very broadly applies. I wonder to what degree this is common elsewhere: GPT-5.5 says basically only Canada comes anywhere close to allowing such a broad classification of work to be considered R&D. It feel like some of the difference seen here might perhaps be definitional differences? The US gets to call so much more R&D, because of our laws/enforcement, & companies are religious about pursuing R&D designation.

Software Development does have development in the name, so perhaps it's not entirely unreasonable a perspective? But it is asymmetric right now, that the US applies it so broadly to software tasks.

I suspect the US would still be ahead even without this particular oddity, whether software dev was counted so or not, but it definitely seems liable to skew our ability to measure & compare accurately. It's not really possible to measure fairly.