Cryogenic electron microscopy has become an essential tool for 3D structure determination of biological macromolecules. Despite many technical advances, the difficulty to reliably prepare samples with uniform ice thickness and the excessive sample loss during grid preparation still present a barrier for routine high-resolution imaging and limit the current throughput of the technique. Nanofabrication techniques employed for Micro-/Nanoelectromechanical Systems (M/NEMS) provide opportunities to miniaturise and automate cryo-EM sample preparation. We have recently shown that MEMS-based nanofluidic sample supports with well-defined geometry can be used to prepare cryo-EM specimens with uniform ice thickness from picoliter sample volumes, and allow for high-resolution structure determination. Despite these promising prospects, several key challenges remain to be addressed in order to transform this approach into a viable alternative to widely used holey support films. I will show where we currently stand with our developments, report on recent efforts in addressing some of the outstanding challenges and elaborate on others that give us a hard time to solve. I will also showcase recent results demonstrating the potential of new chip generations to further automate the cryo-EM workflow, and to explore new frontiers for cryo-EM applications such as time-resolved imaging and high-throughput screening.​