Here we investigate the extent to which infrared heating used to warm plant canopies in climate manipulation experiments increases transpiration. Concerns regarding the impact of the infrared heater technique on the water balance have been raised before, but a quantification is lacking. We calculate transpiration rates under infrared heaters and compare these with air warming at constant relative humidity. As infrared heating primarily warms the leaves and not the air, this method increases both the gradient and the conductance for water vapour. Stomatal conductance is determined both independently of vapour pressure differences and as a function thereof, while boundary layer conductance is calculated using several approaches. We argue that none of these approaches is fully accurate, and opt to present results as an interval in which the actual water loss is likely to be found. For typical conditions in a temperate climate, our results suggest a 12–15% increase in transpiration under infrared heaters for a 1 °C warming. This effect decreases when stomatal conductance is allowed to vary with the vapour pressure difference. Importantly, the artefact is less of a concern when simulating heat waves. The higher atmospheric water demand underneath the heaters reflects naturally occurring increases of potential evapotranspiration during heat waves resulting from atmospheric feedback. While air warming encompasses no increases in transpiration, this fully depends on the ability to keep humidity constant, which in the case of greenhouses requires the presence of an air humidification system. As various artefacts have been associated with chamber experiments, we argue that manipulating climate in the field should be prioritized, while striving to limit confounding factors. The excess water loss underneath infrared heaters reported upon here could be compensated by increasing irrigation or applying newly developed techniques for increasing air humidity in the field.