Atomic-layer-deposited (ALD) ZrO₂-based dielectric layers have been employed for a long time as the dielectric materials of dynamic random access memory (DRAM) capacitors. However, since the inborn permittivity of ZrO₂ is not sufficiently high (< 40), further scaling of the memory devices is facing the limit. Interposing leakage-blocking layers such as Al₂O₃ within the dielectric films also disturbs achieving a high enough capacitance density. Recently, SrTiO₃ (STO) has been considered as a promising candidate for the dielectric materials in next-generation DRAM capacitors due to the exceptionally high dielectric constant (>150 in the thin-film state) and acceptable energy bandgap (3.0 - 3.2 eV). However, adopting STO is still not easy to succeed in 3D structure due to the difficulty and complication of the ternary oxide ALD. In addition to the examples of the nonideal growth behavior in complex ALD and the difficulty of multicomponent ALD, the overall history of STO ALD would be presented.