Carbon dioxide is the ultimate source of the fossil fuels that are both central to modern life and problematic: their use increases atmospheric levels of greenhouse gases, and their availability is geopolitically constrained(1). Using carbon dioxide as a feedstock to produce synthetic fuels might, in principle, alleviate these concerns. Although many homogeneous and heterogeneous catalysts convert carbon dioxide to carbon monoxide(2), further deoxygenative coupling of carbon monoxide to generate useful multicarbon products is challenging(3). Molybdenum and vanadium nitrogenases are capable of converting carbon monoxide into hydrocarbons under mild conditions, using discrete electron and proton sources(4). Electrocatalytic reduction of carbon monoxide on copper catalysts(5) also uses a combination of electrons and protons, while the industrial Fischer-Tropsch process uses dihydrogen as a combined source of electrons and electrophiles for carbon monoxide coupling at high temperatures and pressures(6). However, these enzymatic and heterogeneous systems are difficult to probe mechanistically. Molecular catalysts have been studied extensively(6-23) to investigate the elementary steps by which carbon monoxide is deoxygenated and coupled, but a single metal site that can efficiently induce the required scission of carbon-oxygen bonds and generate carbon-carbon bonds has not yet been documented. Here we describe a molybdenum compound, supported by a terphenyl-diphosphine ligand, that activates and cleaves the strong carbon-oxygen bond of carbon monoxide, enacts carbon-carbon coupling, and spontaneously dissociates the resulting fragment. This complex four-electron transformation is enabled by the terphenyl-diphosphine ligand(24,25), which acts as an electron reservoir and exhibits the coordinative flexibility needed to stabilize the different intermediates involved in the overall reaction sequence. We anticipate that these design elements might help in the development of efficient catalysts for converting carbon monoxide to chemical fuels, and should prove useful in the broader context of performing complex multi-electron transformations at a single metal site.