Global health policy experts tend to organize hunger through scales of “the individual,” “the community,” and “the global.” This organization configures hunger as a discrete, measureable object to be scaled up or down with mathematical certainty. This article offers a counter to this approach, using ethnographic cases to illustrate that the calculated referent of “hunger” does not hold stable. In the highlands of Guatemala, where obesity has become a matter of concern, many people treated hunger as a sensation connected to family and history. For doctors working in the region, hunger was determined through body mass indices and global risk statistics. For global health experts it was different still, operating as an indicator derived from agricultural and population data. I draw these different, yet connected, versions of hunger together to explore dilemmas of scaling an object that isn’t solid but is made and unmade variously. This allows me to illustrate that global hunger is not a summation of hunger in the world, but its own version of hunger. I further suggest that “multi-object ethnography,” which allows for the persistence of uncertainty, can help to develop policy responses to hunger(s) that will, in some cases, be more appropriate and effective than scale-based evaluation.