The Kiruna mine produced about 19 Mt processed iron ore in 1999 using sublevel caving. General failure in the hangingwall has been widely studied and its mechanism is rather well understood. Failure in the footwall is more recent and not perfectly explained. Plane shear failure and Hoek and Bray's circular failure charts were successfully used to match the first surveyed failure surfaces. The footwall is not free yet but covered by a caved waste rock abutment whose exact behaviour is poorly known. Based on new field data and numerical models (FLAC), the following paper presents recent improvements in understanding the Kiruna mine's footwall failure process.