Aeration control in stored grain was largely developed for temperate climates, where cool air is used to reduce grain temperature. In hot semi-arid stores, however, moisture can be the dominant hazard and the coolest air may also be the most humid. This study evaluated whether aeration should be triggered by the drying capacity of incoming air rather than temperature alone, and assessed the contribution of wall insulation. A psychrometric drying-capacity criterion was tested in two geometrically identical octagonal steel silos containing 3 Mg of barley each; one was bare steel and the other jacketed with 4 cm expanded polystyrene. The silos operated side by side through three regimes in 2025. On a common 768 h autumn block, continuous, fixed-night and temperature-cooling controls were evaluated as counterfactual simulations using the logged weather and grain record and compared with the physically deployed drying-capacity controller. Without ventilation, the insulated silo damped bulk-temperature fluctuations more strongly than the bare silo (53% versus 13%). Simulated continuous ventilation selected air at 72% relative humidity, temperature-cooling control at 75–86%, and fixed-night ventilation at 85%. In contrast, the drying criterion selected air at 39% relative humidity in the bare silo and 44% in the insulated silo. The deployed controller operated for 75 and 40 h, respectively, using air averaging 45% and 36% relative humidity. Fan operation was 61–79% lower than the simulated fixed schedule and 75–95% lower than the simulated temperature-based or continuous strategies. Sensor-inferred grain water activity fell below the indicative mould-growth boundary for more than 92% of hours, and incoming dew point remained below grain temperature except for one marginal hour. Grain sampling every three days detected no insect infestation, although localized moist pockets occurred in the non-insulated silo. From this single instrumented deployment, the results indicate that drying capacity was a more effective aeration trigger than temperature alone under the semi-arid conditions investigated. The low-cost psychrometric controller selected drier air and reduced fan operation while under generally favourable physical storage conditions during this deployment; microbiological safety was not assessed.