Primary plastids descend from the cyanobacterial endosymbiont of an ancient eukaryotic host, but the initial selective drivers that stabilized the association between these two cells are still unclear. One hypothesis that has achieved recent prominence suggests that the first role of the cyanobiont was in energy provision for a host cell whose reserves were being depleted by an intracellular chlamydial pathogen. A pivotal claim is that it was chlamydial proteins themselves that converted otherwise unusable cyanobacterial metabolites into host energy stores. We test this hypothesis by investigating the origins of the key enzymes using sophisticated phylogenetics. Here we show a mosaic origin for the relevant pathway combining genes with host, cyanobacterial or bacterial ancestry, but we detect no strong case for Chlamydiae to host transfer under the best-fitting models. Our conclusion is that there is no compelling evidence from gene trees that Chlamydiae played any role in establishing the primary plastid endosymbiosis.