The spread of agriculture is an important topic of archaeological research, but relatively few studies address the drivers behind the spread of specific species empirically. Here we use published isotopic data to consider whether the millets spread from their putative domestication centre in the East to western Eurasia for use as a staple food. We show that the consumption of significant quantities of millet was both far more sporadic than the earliest appearance of millet might suggest and delayed. This is not to say that millet was not consumed, rather that any consumption was below the level of isotopic detectability, and thus millet cannot generally be considered a staple. Nevertheless, individuals who regularly consumed millet occur both as typical members of their population and as unusual individuals. The reasons for this pattern open up new questions about, and avenues of research into, the spread of agriculture.