The development of new technological materials has historically been a difficult and time-consuming task. The traditional role of computation in materials design has been to better understand existing materials. However, an emerging paradigm for accelerated materials discovery is to design new compounds in silico using firstprinciples calculations, and then perform experiments on the computationally designed candidates. In this paper, we provide a review of ab initio computational materials design, focusing on instances in which a computational approach has been successfully applied to propose new materials of technological interest in the laboratory. Our examples include applications in renewable energy, electronic, magnetic and multiferroic materials, and catalysis, demonstrating that computationally guided materials design is a broadly applicable technique. We then discuss some of the common features and limitations of successful theoretical predictions across fields, examining the different ways in which first-principles calculations can guide the final experimental result. Finally, we present a future outlook in which we expect that new models of computational search, such as high-throughput studies, will play a greater role in guiding materials advancements.