As the name suggests, it's R&D. R&D in lay man terms means hit and trial with the technology. Even a hit rate of single digit % would be considered as good as R&D outcomes are known to give benefits for longer durations. But on the other hand you would constantly be investing in R&D. 90% of the investment may go down the drain but 10% may bear fruit. So to take care of both scenarios, a SLM seems appropriate unless we have some mechanism to distinguish R&D investments as fruitful and not so fruitful.