In recent years, Musical Acoustics has made dramatic advances in physical modeling techniques using Ultrascale Computing methods, such as Graphic Processing Units (GPU) or Field Programmable Gate Arrays (FPGA). These methods allow solving differential equations, including highly nonlinear frequency-dependent viscoelastic damping, pointing to internal damping as maybe even more important than eigenmodes regarding instruments’ sound characteristics. Very recent suggestions of adding Acoustic Metamaterials to musical instruments not only enlarge sonic possibilities but also create sounds impossible with traditional instruments. Along with machine learning methods using psychoacoustic features of timbre perception, solving the inverse problem is within reach, estimating how to build a musical instrument starting from sound idle. These advances are accompanied by new nonlinear dynamical models of musical instruments able to predict dynamical processes with much higher precision than linear models. All these recent developments are about to lift musical instrument building to another level of customer sound-designed instruments with precise spectral and transient properties and sounds, which are impossible to realize with traditional mechanical instruments.