We report an original alkane elimination approach, entailing the protonolysis of triisobutylaluminum by the acidic hydrides from Cp*IrH4. This strategy allows access to a series of well-defined tri- and tetranuclear iridium aluminum polyhydride clusters, depending on the stoichiometry: [Cp*IrH3Al(iBu) (2)](2) (1), [Cp*IrH2Al(iBu)](2) (2), [(Cp* IrH3)(2)Al(iBu)] (3), and [(Cp*IrH3)(3)Al] (4). Contrary to most transition-metal aluminohydride complexes, which can be considered as [AlHx+3](x-) aluminates and LnM(+) moieties, the situation here is reversed: These complexes have original structures that are best described as [Cp*IrHx](n-) iridate units surrounding cationic Al(III) fragments. This is corroborated by reactivity studies, which show that the hydrides are always retained at the iridium sites and that the [Cp*IrH3](-) moieties are labile and can be transmetalated to yield potassium ([KIrCp*H-3], 8) or silver (([AgIrCp*H-3](n), 10) derivatives of potential synthetic interest. DFT calculations show that the bonding situation can vary in these systems, from 3-center 2-electron hydride-bridged Lewis adducts of the form Ir-H -> Al to direct polarized metal-metal interaction from donation of d-electrons of Ir to the Al metal, and both types of interactions take place to some extent in each of these clusters.