How will future historians study the 2020s? The 59 zettabytes of data estimated to have been generated by the digital revolution until 2020 dwarfs what historians traditionally encountered. The challenge for the historian is not only due to information overload, but also difficulties in how to access big data as an archive, such as bit rot, sharding, replication, (in)compatibility, encryption, and the physical presence of digital information in the form of data centers and global communications infrastructure. This new reality prompts the need to rethink the established approaches to digital history, which, while innovative, is designed for converting documents to digital media and applying quantitative methods on sub-1.0 gigabyte data sets. As today's born-digital artifacts are vast, dynamic, and heterogeneous, research and training in the nature of big data from a historian's perspective are a necessity, not an option. In this talk, I will present the fruits of two years of research at the Big Data Studies Lab, where we investigate the preservation, authentication, energy demands, and societal implications of big data. Our approach is inspired by how book historians examine parchment, paper, ink, printing, and circulation but in the context of solid-state drives, 5D optical discs, and content-delivery networks.