Materials innovation has a scale-up problem, not discovery

Materials innovation faces a scale-up problem, not a discovery problem. Known materials for AI, quantum, and energy cannot be manufactured at production scale. Intel spent over a decade making hafnium-based high-k dielectric manufacturable for the 45-nanometer node in 2007. Atomscale builds systems that use existing data to guide growths in real time, extracting 43x more useful information than baseline unsupervised machine learning.