Hydrodesulfurization (HDS) is the process by which sulfur-containing impurities are removed from petroleum streams, typically using a heterogeneous, sulfided transition metal catalyst under high H_2 pressures and temperatures. Although generally effective, a major obstacle that remains is the desulfurization of highly refractory sulfur-containing heterocycles, such as 4,6-dimethyldibenzothiophene (4,6-Me_2DBT), which are naturally occurring in fossil fuels. Homogeneous HDS strategies using well-defined molecular catalysts have been designed to target these recalcitrant S-heterocycles; however, the formation of stable transition metal sulfide complexes following C–S bond activation has largely prevented catalytic turnover. Here we show that a robust potassium (K) alkoxide (O)/hydrosilane (Si)-based (‘KOSi’) system efficiently desulfurizes refractory sulfur heterocycles. Subjecting sulfur-rich diesel (that is, [S] ∼ 10,000 ppm) to KOSi conditions results in a fuel with [S] ∼ 2 ppm, surpassing ambitious future governmental regulatory goals set for fuel sulfur content in all countries. Fossil fuels contain naturally occurring organosulfur impurities, with quantities varying depending on the type of feedstock. These sulfur-containing organic small molecules poison catalytic converters and generate polluting sulfur dioxides when combusted. Hydrodesulfurization (HDS) is the industrial process by which sulfur impurities are removed from petroleum fractions prior to their use as fuels. Currently, HDS is performed by treating petroleum with H_2 at high pressures and temperatures (that is, 150–2,250 psi and 400 °C) over heterogeneous catalysts such as cobalt-doped molybdenum sulfide supported on alumina (that is, CoMoS_x∕γ-Al_2O_3; Fig. 1a). However, certain organosulfur species, in particular dibenzothiophenes alkylated at positions 4 and 6, are not efficiently removed. Homogeneous strategies employing sophisticated, well-defined transition metal complexes—including those based on platinum, nickel, tungsten, molybdenum, palladium, ruthenium, rhodium, iron, cobalt, and others—have been extensively investigated. While these studies have provided valuable mechanistic insights, several fundamental issues, such as the formation of stable organometallic S–M species upon C–S bond activation by the metal centre (Fig. 1b), generally restrict industrial implementation of such methods. Rare examples of desulfurization of dibenzothiophenes alkylated at the 4 and 6 positions by homogeneous transition metal catalysis utilized either Ni compounds in combination with superstoichiometric alkyl Grignard reagents or Ni or Co phosphoranimide complexes in the presence of superstoichiometric KH. These issues pose a formidable challenge for the development of new HDS methods. Moreover, increasingly strict governmental regulations require limiting the sulfur content in diesel fuel and gasoline (in the US: typically