One of the main advantages of cold-formed steel profiles is the great flexibility of cross-sectional shapes, attributable to the manufacturing process allowing achievement of almost any desired cross-section. The cross-sectional shape is the key element in enhancing the strength of cold-formed steel profiles as it controls the three fundamental buckling modes: local, distortional (for open profiles) and global. However, research on optimisation of cold-formed steel profiles has been mainly restricted to the conventional C, Z or Σ cross-sectional shapes and seldom new cross-sectional shapes have been considered. This report presents the optimisation of cold-formed open columns using the recently developed self-shape optimisation method that aims to discover new profile shapes. The strength of the cold-formed steel sections is calculated using the Direct Strength Method (DSM) and the rules developed in this research to automatically determine the local and distortional elastic buckling stresses from the Finite Strip and constrained Finite Strip Methods are discussed. The rules are verified against conventional and optimum sections yielded in this research, and found to accurately predict the elastic buckling stresses. The optimisation method is applied to singly-symmetric (mono-symmetric) cold-formed steel columns and the operators behind the method for the special case of singly-symmetric open profiles are introduced in this paper. “Optimum” cross-sections for simply supported columns, 1.2 mm thick, free to warp and subjected to a compressive axial load of 75 kN are presented for column lengths ranging from 1,000 mm to 2,500 mm. Results show that the “optimum” cross-sections are found in a relatively low number of generations and typically form non-conventional “bean”, “oval” or rounded “Σ” sections. The algorithm optimises for distortional and global buckling, therefore likely subjecting the cross-sections to buckling interaction. A manual attempt to redraw the “optimum” cross-sections to include limitations of current manufacturing processes is made. Future developments of the method for practical applications are also discussed.