Residential buildings account for about one-third of the final energy demand in Norway. Many cost-effective measures for reducing heat losses in buildings are known, and their implementation may make the building sector one of the largest contributors to climate change mitigation. To determine the sectoral emission reduction potential, we model a complete transformation of the dwelling stock by 2050 by applying both renovation and reconstruction with different energy standards. We propose a new dynamic stock model with an optimization routine to identify and prioritize buildings with the highest energy saving potential. We combine material flow analysis (MFA) and life cycle assessment (LCA) techniques to extend the sectoral boundary beyond direct household emissions. Despite an expected population growth of almost 50% between 2000 and 2050, sectoral carbon emissions in that period may drop between 30% and 40% for scenarios where the stock is completely transformed by either reconstruction or renovation to the passive house standard. Due to its lower upstream impact, renovation leads to a lower sectoral carbon footprint than reconstruction. Full transformation, however, is not sufficient to achieve an emissions reduction of 50% or more, as required on average to limit global warming to 2 degrees Celsius, because hot water generation, appliances, and lighting will dominate the sectoral footprint once the stock has been transformed. A first estimate of the additional impact of realistic energy efficiency and lifestyle changes in the nonheating part of the sector reveals a maximal total reduction potential of about 75%.