Knowledge of relationships among plants has improved dramatically but for many groups, it still rests solely on information from the plastid genome. Consequently, only parts of the organisms’ evolutionary history are revealed. For the ancient gymnosperms of Ephedra (Gnetales), previous conclusions were based on both plastid and nuclear ribosomal DNA data, but results were typically poorly resolved and supported, presumably because of information poverty in the utilized gene regions. With the aim of resolving phylogenetic questions using more data, we sequenced the plastid genome and the nuclear ribosomal cistron for 50 specimens of Ephedra, largely covering the diversity of the genus. Phylogeny and node ages were estimated using maximum likelihood and Bayesian methods. However, instead of clarifying a few remaining uncertainties, we were left with new questions and incongruent results. Age estimates of the crown group of Ephedra vary considerably depending on utilized software and specified tree prior. Furthermore, previous estimates of the phylogeny of Ephedra have largely reflected information from nrITS, despite utilization of plastid gene regions as well. With analyses based on the entire plastid genome, completely new results emerge. Earlier conclusions of deep divergences in the genus are not supported, and there are ample examples of phylogenetic incongruence. Our study overturns conclusions in previous work and highlights that we still know fairly little about evolution in the ancient Ephedra lineage. How many species of Ephedra are there, and how are they related? How old is the crown group? Many species appear affected by a history of hybridization/introgression and/or polyploidy, but other processes may result in similar patterns and reasons for the detected incongruences must be further analyzed, preferably using population‐level sampling and low‐copy nuclear data.