I somewhat agree with this vision, but in my eyes, correctly performed experimental work would give more certainty when it comes to conclusions and give a chance for unforeseen discoveries.
That said, with computers becoming faster and faster, the idea of using only computation to find materials seems to become even more useful and the speed of discoveries may outweigh most benefits that experimental discovery offers.
I would expect this intensive computational work towards a ‘google maps’ for materials design to become viable within the next few decades, but of course with final experimental verification for promising materials that will be used in the real world.