This paper examines the tech cluster phenomenon by considering three partially answered questions. We first ask how to define a tech cluster—that is, what properties are required to be a tech cluster? This delineation is harder than it appears at first glance and raises some key questions and issues. We start with the scale and density of local activity and then extend into the frontier nature of the work being undertaken and its ability to impact multiple sectors of the economy. We illustrate our definition through some common metrics like patents, venture capital funding, and employment in sectors that are intensive in research and development or in digital-connected occupations. We also note some interesting clues from emerging metrics (for example, high-growth entrepreneurship, artificial intelligence researchers) and recent efforts to measure tech clusters globally. We then ask how tech clusters function, with a focus on traits that extend beyond those associated with traditional industrial clusters. Not surprisingly, knowledge spillovers are a powerful force in tech clusters, and recent work explores how knowledge transmits across firms situated in a tech cluster and how density impacts the types of innovations created. Tech clusters facilitate powerful scaling for the best designs when they combine modular product structures with high-velocity labor markets. Universities, high-skilled immigration, and global production linkages also feature prominently in the functioning of leading US centers. Finally, we turn to the roots of tech clusters and inquire into the mix of initial ingredients required for their formation. Leading tech clusters are far from permanent and have frequently emerged in new places following the advent of new general-purpose technologies. Today, the rapid growth of Toronto as an artificial intelligence cluster suggests that there may be limits to Silicon Valley’s grip on this frontier. Yet despite the government having played an important role in this history of many tech clusters, top-down attempts to re-create Silicon Valley have mostly failed (Lerner 2009). Our historical examples suggest that local officials instead may wish to facilitate the scaling of nascent industries that have taken root, even if due to random chance, rather than attempt to engineer a cluster from scratch. We conclude with some thoughts on future research opportunities, including the question of whether tech clusters are at their high-water mark or are likely to strengthen further. The implications of the ongoing COVID-19 crisis for tech clusters could be profound. Our discussion focuses primarily on the US economy, but much of what we describe applies to other countries as well. We ground our discussion firmly within the economics and management disciplines, occasionally reaching out in incomplete ways to other social sciences as we go.