Reducing the usage of fossil fuels is a central issue in ongoing policy debates. This in particular holds for Dutch horticulture, given its energy-intensive production. We analyse differences in energy usage and price responsiveness of horticultural firms by estimating energy demand functions using a Bayesian random coefficient model. Beyond, the effects of a proposed energy tax are assessed. Allowing for firm-specific energy price coefficients gives a better model fit compared to conventional models with fixed slope parameters. This confirms that firms respond differently to energy prices, which is taken into account in simulating the effects of more restrictive energy policies. The results show that larger-sized firms use less gas per square meter yet also point at a considerable spread in additional energy expenses between firms.