Materials by design has long been a holy grail of computational materials design. The idea is that one could predict both the chemical composition, structure, and desired functional physical properties of materials based on "ab initio" electronic structure calculations.
There is a nice Physics Viewpoint, "Materials prediction scores a hit", by Filip Ronning and John Sarrao. The two pages are worth reading and digesting. The authors puts in context the recent successful prediction of superconductivity in a high pressure phase of iron tetraboride.
Why is predicting superconductors so hard? Particularly, in strongly correlated electron materials? It is a problem with multiple energy scales. Basically, superconductivity is an emergent low-energy phenomena that is an instability in a metallic state, that itself involves emergent low-energy scales.
Given the above one can debate the merits of the White House Materials Genome Initiative, but be excited about the recently announced $90 million dollar initiative "Emergent Phenomena in Quantum Systems" of the Gordon and Betty Moore foundation. The focus is on Quantum materials with a significant emphasis on solid state synthesis.