This study presents an experimental and numerical investigation on the buckling behavior of cold-formed thin walled steel storage rack upright frames. A total of twelve component tests were performed to examine the ultimate loads, failure modes, and load-displacement curves of upright frames subject to compressive loading conditions. In the test, the failure modes observed were global buckling (GB) and distortional-global buckling interaction (DB+GB). The accuracy of numerical models for upright frames was confirmed by the experimental results. Parametric analyses were then conducted, considering variations in upright cross-sections, lengths, thicknesses, steel grades, brace sections, and span lengths. In particular, the differences between flexural buckling (FB) and flexural-torsional buckling (FTB) in upright frames with identical geometric configurations but different boundary conditions were compared. The findings indicate that, when all other conditions remain constant, releasing the out-of-plane rotational restraints on the loading side results in a 20-40 % reduction on the ultimate capacity. The relationship between load-bearing capacity of upright frames and single uprights was also analyzed, and the influence of various upright and brace configurations on the ultimate loads was investigated. Additionally, the applicability of the direct strength method (DSM) for predicting the ultimate capacity of upright frames was assessed. The results show that while the current global buckling DSM curve accurately predicts the ultimate load, the existing distortional buckling DSM curve is inadequate for upright frames that fail for DB+GB. Therefore, a modified distortional buckling DSM curve that accounts for buckling interactions is proposed.