Although machine learning has gained great interest in the discovery of functional materials, the advancement of reliable models is impeded by the scarcity of available materials property data. Here we propose and demonstrate a distinctive approach for materials discovery using unsupervised learning, which does not require labeled data and thus alleviates the data scarcity challenge. Using solid-state Li-ion conductors as a model problem, unsupervised learning utilizes a limited quantity of conductivity data to cluster a broad range of Li-containing materials that narrows a high-throughput screening of a large candidate list to a prioritized list for further accurate screening. Our unsupervised learning scheme discovers sixteen new fast Li conductors with conductivities of 10^-4 to 10^-1 S cm^-1 predicted in ab initio molecular dynamics simulations. These compounds have structures and chemistries distinct to known systems, demonstrating the capability of unsupervised learning for discovering materials over a wide materials space with limited property data.