Additive manufacturing (AM) of fiber-reinforced polymer composites enables various complex geometries in an out-of-autoclave process, eliminates the need for molds and tooling, and eliminates material waste. Accompanying these improvements is the ability to tailor the placement of fibers (intricate fiber steering) within a part to improve its performance for a given state of loading. AM is inherently well-suited for integration with optimization schemes, as the designer can send a geometrically complex model directly to the 3D printer for fabrication. This study investigates the printability and performance of three benchmark geometries and loading scenarios optimized for stiffness. These designs are achieved via topology (part geometry) and fiber placement optimization of AM continuous carbon fiber reinforced polymer composites. Some of the manufacturing constraints were implemented in the optimizer. Post-processing was, however, necessary to apply other AM constraints before parts could be successfully printed. The optimized designs were manufactured, on a printer built in-house, and tested in special load fixtures while parts’ displacements were recorded using digital image correlation (DIC). The stiffnesses found through experiments agreed with the finite element analysis (FEA). The optimized AM parts displayed up to 100% improvement in specific (per weight) stiffness, translating to a maximum of 50% weight reduction over conventional composite parts with unidirectional laminas. This study demonstrates that bridging topology and fiber path optimization with AM of continuous fiber-reinforced composites enables significant weight reduction that would otherwise be impossible.