Heme proteins are almost ubiquitous both in Archaea and in Bacteria. The last universal common ancestor (LUCA) of the two prokaryotic domains was, therefore, assumed until recently to already have made use of heme cofactors, a notion bolstered by molecular phylogenies of several heme-bearing enzymes. The discovery of a second pathway for heme biosynthesis, predominantly present in Archaea, was subsequently interpreted to indicate independent origins of heme biosynthesis in each of the two prokaryotic domains (Lane and Martin (Cell 151:1406–1416, 2012)), implying that the LUCA might have been entirely devoid of hemes and heme proteins. In this contribution, we outline the presently available evidence in favour of either scenario and propose a new model reconciling the seemingly contradictory messages sent by molecular phylogeny of heme-bearing enzymes and the biosynthesis pathway dichotomy. A possible relation to other biosynthesis pathway dichotomies is suggested, and the far-reaching repercussions of the ultimate resolution of the controversy on our understanding of free energy conversion in the LUCA and at life’s origin are emphasized.