AimsTesticular germ cell tumours are the most common solid malignancies in young men of age 14–44 years. It is generally accepted that both seminomas and non‐seminomas arise from a common precursor, the germ cell neoplasia in‐situ, which itself is the result of a defective (primordial) germ cell development. The stem cell‐like population of the non‐seminomas, the embryonal carcinoma, is capable of the differentiation of all three germ layers (teratomas) and extra‐embryonic tissues (yolk‐sac tumours, choriocarcioma) into cells. In contrast, seminomas are thought to have a limited differentiation potential. Nevertheless, several studies have highlighted their ability to undergo reprogramming to an embryonal carcinoma or differentiation into other non‐seminomatous entities. Here, we demonstrate that in approximately 5% of seminomas, the yolk‐sac tumour driver gene FOXA2 is detectable at the protein level, indicative of an occult yolk‐sac tumour subpopulation that putatively arose from seminoma cells, as the presence of other GCT entities could be excluded. The presence of these subpopulations might render the tumour more aggressive and argue for an adjustment of the therapeutic concept. We used our data to update the model of germ cell tumour pathogenesis, especially regarding the developmental potential of seminomas. Additionally, we suggest to include detection of FOXA2 into standard routine diagnosis of seminomas.