Large molecular assemblies are ubiquitous in living cells and play key roles in many biological processes. Indeed, multiple copies of protein subunits can organize into large macromolecular nanostructures, in shapes of fibrils, filaments, pores, capsids, etc. Understanding the mechanisms responsible for the molecular assembly of such systems is of primary interest in biology in order to gain insights into their crucial functions. Inspired by these remarkable architectures, supramolecular chemists and material scientists aim at designing synthetic self-assemblies with a broad spectrum of applications, ranging from regenerative medecine to drug delivery. The design of new nanostructures and the tuning of their functionalities require a detailed description and understanding of the weak interactions driving the molecular assembly process. For self-assembled nanostructures involved in cellular processes or engineered by supramolecular chemistry, the 3D structure determination is hampered by several technical challenges: (i) the assemblies usually lack crystalline order required to perform X-ray crystallography and (ii) the high molecular weight prohibits fast molecular tumbling, which restricts the use of solution Nuclear Magnetic Resonance (NMR). So far, hybrid approaches combining high-resolution structures of the isolated subunits and the molecular envelope obtained from electron microscopy have led to few low- to medium-resolution models of such assemblies. However, this approach lacks the experimental determination of the crucial subunit-subunit interfaces, which can lead to inaccuracies since the subunits can adopt different conformations in isolation compared to their relevant assembled states. We have recently proposed a new approach based on modern solid-state NMR techniques to solve atomic structures of complex self-assembled nanostructures, demonstrated by an atomic model of a bacterial filament (Loquet et al., Nature 2012). This breakthrough forms the genuine basis of the NanoSSNMR project. We now envision disentangling more complex supramolecular self-organizations, either involved in synthetic or cellular processes. The NanoSSNMR proposal will exploit state-of-the-art solid-state NMR methods and strategic isotope labeling and integrate hybrid approaches to elucidate the assembly mechanisms, revealing the atomic structures of two complex self-assembled nanostructures.