We consider the uncertainty of theoretical calculations for a specific difference of the hyperfine intervals in the 1s and 2s states in a light hydrogen-like atom. For a number of crucial radiative corrections the result for hydrogen atom and helium ion appears as an extrapolation of the numerical data from medium to low Z. An approach to a plausible estimation of the uncertainty is suggested using the example of the difference D 21 = 8E hfs (2s) − E hfs (1s).Key words: Tests of quantum electrodynamics, Hyperfine structure PACS: 12.20. FV, 21.45.+v, 31.30.Jv, 32.10.Fn Studies of the hyperfine structure in light hydrogen-like atoms are of interest because of a possibility for a precision test of the bound state Quantum Electrodynamics (QED). However, in spite of a record accuracy achieved in experiments on the hyperfine interval in the ground state of hydrogen, deuterium, tritium and helium-3 ion (see, e.g., [1]), we should acknowledge that such a test cannot be really successful until effects of the nuclear structure are known accurately enough.In fact, the bound state QED effects are lower than those of the nuclear structure. Obviously, the latter cannot be known well enough. However, there is a hopeful opportunity to perform such a test if we consider data of two measurements: for hyperfine intervals in the ground and metastable states (i.e. for the 1s and the 2s states).