At the ultimate limit of magnetic recording, suitable storage media will consist of nanometer-sized entities, each of which will carry one bit of information. Materials with a high magnetocrystalline anisotropy energy are required to guarantee thermal stability of the ferromagnetic state at realistic operating temperatures. The face-centered tetragonal (fct) L10 FePt alloy belongs to the promising class of materials that offer the perspective of storing one magnetic bit per nanoparticle. Widespread activities have therefore arisen worldwide, targeting novel strategies for both the synthesis of suitable magnetic nanostructures and their organization into superlattices by means of parallel processes. Here, we present a new approach for the synthesis of size-selected L10 FePt nanoparticles based on the self-organization of spherical micelles formed by diblock copolymers, thereby significantly extending a previous technique to produce large-scale arrays of elemental nanoparticles. Our approach overcomes the typical drawbacks of the current colloidal routes towards densely packed arrays of ferromagnetic FePt nanoparticles while still guaranteeing areal densities exceeding 1 Tbits inch (1 inch≈ 2.54 cm). Since the first presentation of magnetic data-storage devices five decades ago, the areal density of digital information has increased by eight orders of magnitude to reach values of about 200 Gbits inch, as found in present hard disk drives. A few years ago, an efficient method was developed to synthesize FePt nanoparticles on the basis of wet-chemical synthesis (hereafter referred to “colloidal”), which involves particle stabilization by an organic-ligand shell. The significant advantage of this approach, allowing a simple preparation of densely packed 2D nanoparticle arrays from corresponding particle solutions, is, however, compensated by some serious drawbacks related to the thin ligand shell (1–3 nm) which serves as a spacer between the nanoparticles. As a consequence of the resulting small interparticle distance, the nanoparticles exhibit a strong tendency to aggregate during heat treatments. Thermal annealing at 500–600 °C is, however, generally required in order to transform the assynthesized, chemically disordered (Fe and Pt atoms randomly distributed over the lattice sites) face-centered cubic (fcc) structure, which results in superparamagnetic behavior, into the magnetically attractive L10 phase. Furthermore, undesirable collective magnetic dynamics arise at such small interparticle distances through dipolar coupling; collective modes, however, are clearly at odds with the idea of storing magnetic data in individual nanoparticles. Finally, the heat-treated colloidal FePt nanoparticles are found to be highly oxidized and contaminated by carbon because of the thermally induced decomposition of the organic shell. Recent alternative routes for the synthesis of L10 FePt nanoparticles include their formation in cluster beams and from wet-chemical procedures at elevated temperatures. However, even in these cases, the particles still need to be encapsulated with an organic-ligand shell, either after deposition from a cluster beam or during synthesis, to guarantee their ordered organization. Typical interparticle spacings are thus the same as described before, resulting in dipolar coupling and, additionally, the particle–ligand interactions can deteriorate the magnetic properties. Our new approach for generating ultrahigh-density arrays of L10 FePt nanoparticles overcomes all these drawbacks. It is based on the flexibility of macromolecular chemistry to design diblock copolymers with adjustable block lengths, which form reverse micelles when dissolved in an apolar solvent like toluene. After loading their cores with a suitable metal salt, such C O M M U N IC A TI O N