Cloud kitchens or dark kitchens are the latest innovations in the food delivery landscape. Cloud kitchens are delivery-only food businesses. They have no storefront or dining area, resulting in lower capital and operational costs. Cloud kitchens also bring together a mix of eateries and cuisines. They combine orders and encourage bundling. Order bundling, or the delivery of multiple orders on a single trip, has the potential to reduce the negative externalities of food delivery. This study used an agent-based simulation to create a scenario for meal delivery from a cloud kitchen. The agent-based simulation was developed from scratch, using Python. This simulation is used for measuring the effects of introducing a cloud kitchen on emissions and parking loads. The results show that the introduction of a cloud kitchen causes longer distances to be travelled post-delivery, and therefore an increase in emissions when compared to a baseline scenario. However, cloud kitchens have lower fleet size requirements and attract a large idle fleet from neighbouring eateries. Cloud kitchens, by virtue of being located away from thoroughfares, reduce congestion resulting out of on-street parking of delivery vehicles. Thus, the introduction of cloud kitchens has both positive and negative effects on the generated externalities.