This paper explores the environmental impacts of retaining or replacing buildings by extending the consequential replacement framework (CRF) for life-cycle assessment (LCA) from individual buildings towards the urban scale. It addresses a gap in previous studies by examining the broader consequences of building-level retention or replacement decisions in the built environment, namely greenhouse gas (GHG) emissions from associated infrastructure construction. The case study, which is located in Finland, investigates alternative ways to densify a growing city, some of which can be achieved within the urban structure, while others require greenfield construction elsewhere in the city. The results reveal that refurbishing and extending already existing buildings is worthwhile in terms of GHG emissions, even if some of the additional floor area targeted in densification were to be built on a greenfield site. After 50 years, the accumulated GHG emissions of scenarios that develop existing buildings are 0.2–12% less than those from replacement scenarios, depending on whether the targeted densification is minor or major and the amount of infrastructure construction involved. This is primarily due to the 11–35% smaller upfront embodied GHG emissions of refurbishment compared with new construction—which 50 years do not offset. Policy relevance This paper reinforces the notion from previous research that developing the existing building stock holds major emission-savings potential, even more so than replacement with energy-efficient new build. It highlights the potential contribution of existing building stocks in low-emission city development, even in growing city contexts. The case study results challenge the common belief in urban planning that if replacement involves densification, it must be a more low-emission alternative than retention, and conversely, if retention leads to greenfield construction, it must be higher in emissions than replacement. Low-emission urban planning should base its conclusions on urban LCA. This paper introduces an infrastructure-extended CRF that accounts for building retention and replacement scenarios, ensuring methodological robustness. The approach is essential for policymakers to base their decisions on evidence and prioritise low-emission development alternatives. Which policymakers hold power over such decisions and which instruments are suitable to pursue these goals depend on the context.