Conventionally, buoyant foundations are designed based on Archimedes’ principle, by which the buoyant force is computed by simply measuring the weight of groundwater that is displaced. However, soil at shallow depths is generally unsaturated and should be considered as a multiphase porous medium with heterogeneous properties. The efficacy of the Archimedes’ calculation of buoyant force is questionable because recent field evidence indicates that the measured buoyant force is actually much smaller, and the foundation design is always conservative. This article presents a series of model tests on shallow foundations embedded in different surrounding types of materials, such as pure water, clay, and sand to clay-sand composites. Both the transient and sustained hydrostatic uplift forces were measured. It has been demonstrated that the measured buoyant force increased with time and was always less than the theoretical value. A reduction coefficient was consequently suggested to scale down the Archimedes’ buoyant force for use in design. For shallow foundations in clay, a value of 0.3–0.4 should be used to evaluate the transient uplift force, and a higher value of 0.7–0.8 is recommended for the estimation of the steady-state buoyant force. For sand, the reduction coefficient varies between 0.85 and 0.95 for transient and sustained uplift conditions, respectively.