Traditional construction cost estimating methods that rely on historical cost data are not well suited for underground transit feasibility studies, because not only do construction costs vary widely because of subsurface, geographic, and other project-specific parameters, but also because such construction costs are not generally available in cost databases. Furthermore, the inherently expensive and unknown nature of underground construction often leads to inaccurate cost estimates, which in turn can lead to a significant budget shortfall as the project moves from planning and design to construction. What does this mean for the planner producing an EIS/EIR or the engineer involved in conceptual design? It means that tunnels and other associated underground work cannot be estimated on a “per foot basis” as is customarily done during the early stages of an above-ground transportation project. Labor factors and geologic conditions will often dictate tunnel construction costs; tunnels in Boston have different labor factors than tunnels in Los Angeles, and tunnels in downtown San Francisco through soft-ground will have very different costs than tunnels through weak rock on the west side of San Francisco. This paper summarizes the unique aspects of underground construction relating to cost and concludes that while production-type estimates are crucial for developing realistic project budgets, such estimates do not require a high degree of project definition. Tunnel construction, and in particular transit tunnel construction, is by its very nature complex, risky, and often fraught with geologic unknowns. In urban areas, tunnel transit has a distinct advantage of minimizing surface disruptions compared to surface or cut-and-cover transit configurations. However, tunnel construction is an expensive endeavor. While social, political and environmental forces have favored tunnel alignments for many transit systems, the costs of construction have often exceeded budgets, preventing the development of new transit lines or the extension of existing systems, and eroding the public’s confidence in the ability of tunneling projects to be accurately forecasted. It is therefore in transit agencies’ best interests to get the most realistic and accurate cost estimate possible, particularly during the planning stages of a project. An estimate of construction cost for a tunnel alignment on a per foot or per mile basis is nothing more than a “guesstimate” that will likely prove to be grossly inaccurate in hindsight because it ignores the geography, geology, allocation of risk, and market conditions associated with a particular project.