Providing a complete scheme to describe the methanol-to-hydrocarbon (MTH) reaction network is impeded by an insufficient understanding of the incipient chemistry. Advanced synchrotron-based photoelectron photoion coincidence spectroscopy and photoionization mass spectrometry (MS) identified intermediate species in a temperature-dependent reaction strategy. Combined with theoretical calculations over the H-ZSM-5 zeolite, we provide a comprehensive understanding of the initial hydrocarbon formation, including the initial C–C bond, olefins, and aromatics. Our results indicate that surface methoxy species (SMS) can react with formaldehyde to yield surface acetyl and gas phase ethenone (ketene). Methylketene and dimethylketene are experimentally captured in a real MTH reaction, and they can be transformed into olefins. Observing the generation of initial olefins accompanied by ketene and carbon monoxide completes the evidence that ketene methylation followed by decarbonylation leads to the initial olefins in MTH. Furthermore, a kinetically favorable route to form initial aromatics is identified experimentally and theoretically, which involves tandem Prins, Diels–Alder, and hydrogen-transfer reactions with butadiene and (methyl)cyclohexenes as key intermediates.