Haplodiploidy tends to arise in lineages with gregarious broods and maternally-inherited endosymbiotic bacteria. When broods are gregarious, maternally-transmitted symbionts often evolve amale-killing phenotype. Here we explore the possibility that male-haploidy originally arose as a male-killing bacterial phenotype, by modeling the conditions under which a male-haploidizing symbiont can invade a diploid host population. We propose both a population genetics model and a game theory model. We find that haploidizing endosymbionts can invade a population hosting neutral symbionts under a fairly wide range of realistic parameter values. We derive a simple invasion condition requiring that the product of the transmission of a haploidizing endosymbiont by its viability is greater than the viability of a neutral endosymbiont. Furthermore, we show that both symbiont transmission and host survivorship to haploidization are under selective pressure to increase, causing the system toevolve towards a male-haploid population as observed in haplodiploid species with paternal genome elimination. Finally, we found out that for certain combinations of parameter values male haploidization can benefit both symbiont and host although there is a constant conflict of interests over male survivorship.