We will develop the conceptually new paradigm for ultra-dense and ultrafast magnetic storage that will exceed the current technology by two orders of magnitude in storage density (going from terabit/inch2 to tens of terabytes/inch2) and by about four orders of magnitude in operation speed (going from low GHz to THz for read/write). This will be achieved in an all-optical platform that allows deterministic, non-thermal, low-energy, ultrafast magnetization switching at few nanometers and potentially down to a molecular length-scale. The main building block of the envisioned memory unit in this new paradigm is the spinoptical nanoplasmonic antenna that concentrates pulsed polarized light at the nanometer length-scale and enables non-thermal spin-orbit mediated transfer of the light’s angular momentum (orbital and/or spin) to the nanoscale magnetic architectures. In this way fs-pulsed light, assisted by the plasmonic optical spin-selective antenna and the local electromagnetic field enhancement, allows for the precise control of the magnetic state of nanometer sized / molecular magnetic structures. The project aims to elucidate the fundamentals of the emergence and manipulation of light’s orbital and spin angular momenta to achieve non-thermal momentum-transfer-driven ultrafast switching process, to demonstrate its practical realization, and will map its suitability for future upscaling towards industrial implementation in devices.