To rationalize and restrict health care spending, policy makers in many jurisdictions have withdrawn insurance or funding for selected health care technologies. Numerous analytic frameworks and applied exercises have emerged to guide decisions about “which” services to cut. But in their focus on choice-making processes, these efforts have paid little attention to the problem of defining and dividing the set of technologies to choose among. If technology assessment refers to methods for weighing services for their relative value, the term technology assembly might be used to refer to methods for framing the technological trade-offs to enroll in such contests. This article examines technology assemblies found in several types of theoretical and applied rationing exercises (including Oregon’s Medicaid rationing process, economic evaluation literature, citizen “values” surveys, and Canadian provincial deinsurance policies). Based on this review, some key conceptual conventions and problems in technology assembly can be identified. The boundaries between health technologies are fuzzy, interlocked, layered, and continuously moving. Consequently, the defining features of technological trade-offs are inevitably socially constructed and negotiated. Trade-offs can be arranged along numerous dimensions, and the divisions typically correspond to broader political, administrative, and ethical dilemmas in health policy. Examples include equity among demographic classes, concepts of need, legitimacy of therapeutic goals, and so forth. Insights into the process of constructing technological trade-offs may help policy makers better question what technologies they are looking at and why, before moving on to the task of determining which ones to cover.