The detection of primordial gravitational waves is one of the biggest challenges of the present time. The existing (Wilkinson Microwave Anisotropy Probe) observations are helpful on the road to this goal, and the forthcoming experiments (Planck) are likely to complete this mission. We show that the 5-year Wilkinson Microwave Anisotropy Probe $TE$ data contains a hint of the presence of gravitational wave contribution. In terms of the parameter $R$, which gives the ratio of contributions from gravitational waves and density perturbations to the temperature quadrupole, the best-fit model produced $R=0.24$. Because of large residual noises, the uncertainty of this determination is still large, and it easily includes the R=0 hypothesis. However, the uncertainty will be strongly reduced in the forthcoming observations which are more sensitive. We numerically simulated the Planck data and concluded that the relic gravitational waves with $R=0.24$ will be present at a better than 3$sigma$ level in the $TE$ observational channel, and at a better than 2$sigma$ level in the `realistic $BB$ channel. The balloon-borne and ground-based observations may provide a healthy competition to Planck in some parts of the lower-$ell$ spectrum.