Otherwise, you might as well cut most R&D spending and focus on reverse engineering. Getting really good at following fast is probably a lot cheaper than getting really good at inventing new things that turn into products.
Also, then you could spend a smaller fraction of money to "acquire" (through various means) the schematics and trade secrets and produce your own version of the next big thing. Sure, that is probably unethical, but without any semblance of protection via patents, trademarks, other IP law, it seems like that is what you would get.
That being said, as it gets cheaper to distribute information and turn that information into physical goods, the amount of value in the production of physical goods is going to approach the cost to manufacture those goods. There will always be some money in service, repair, etc. but a lot of things value will be destroyed along the way and it is not yet clear if that is a net positive for society at large.