Automatic milking systems (AMSs) have become increasingly common in the US in the past few years. Recent surveys from Idaho, one of the largest dairy-producing states, as well as from other states and countries, suggest that: 1. among farms adopting robotic milking systems, few are reporting less labor usage after adopting this labor-saving technology; 2. small farms rather than large farms are adopting (or more interested in adopting) robotic milking systems. In this article, we propose a series of new modeling strategies, which introduces the role of general-equilibrium effects to explain these new stylized facts. We show that: first, farms adopting labor-saving technology may, in fact, use more labor to compensate for the loss in the value of labor; second, when smaller farms experience more labor efficiency gains or value their leisure time (or off-farm income) more, they are more likely than larger farms to adopt the new technology. We contribute to the technology-adoption literature in two important ways. First, to our knowledge, this is the first article that introduces general-equilibrium effects to the technology-adoption literature. Second, this is also the first article that provides a theoretical perspective to explain the stylized facts in the adoption of robotic milking systems.