**How simulations drive the discovery of novel materials, and novel physics**

First-principles simulations are one of the greatest current accelerators in the world of science and technology. To provide some context, one could mention that 30,000 papers on density-functional theory are published every year (this corresponds to an investment of roughly 3 billion US$ PPP); that 12 of these are in the top-100 most-cited papers in the entire history of science, engineering, and medicine; and that initiatives based on open science for codes, data, and simulation services are multiplying worldwide.

I’ll highlight some of our own scientific and technological perspectives on this, starting with the goals and the infrastructure needed to deliver on the promise of materials discovery, and applying it to the case study of ~1800 novel two-dimensional materials (including e.g. the first Kane-Mele quantum spin Hall insulator) and their possible applications in electronics or energy.

I’ll then argue how the need to compute some of the most relevant materials properties – in this case transport – forces us to critically re-evaluate some of the stalwarts of condensed-matter physics: learning that phonons are just a high-temperature approximation for the heat carriers, or discovering that the Boltzmann transport equation can be generalized to describe simultaneously the propagation and interference of phonon wavepackets, thus unifying the description of thermal transport in crystals and glasses.