That's simply not true either. The entire hearing aid industry, for example, is pretty much entirely additive manufacture.
Or for a smaller example, GE uses AM to produce jet nozzles not only with an architecture impossible to achieve in SM, but also at lower cost. Where previously they needed a complex of machines to produce a single nozzle, now they only need one. That's not just R&D, that's a business decision. The next 777 is slated to have an entirely 3d printed GE engine, and I have my doubts that Boeing would have made that contract if they weren't seeing cost savings.
Speed and resolution are catching up to traditional SM processes, and as noted in some cases actually superior. Cost will inevitably fall as greater adoption across industry occurs, and there is an undeniable carrot to drive adoption- the ability to do away with re-tooling factories. Considering all the other benefits possible with AM like programmable properties, minimal resource waste, portability...
It's only a matter of time.