High crude oil prices and the eventual decline of conventional oil production raise the issue of alternative fuels such as non-conventional oil. The paper describes a simple probabilistic model of the costs of synthetic crude oil (SCO) produced from Canadian oil sands. Synthetic crude oil is obtained by upgrading bitumen that is first produced through mining or in-situ recovery techniques. This forward-looking analysis quantifies the effects of learning and production constraints on the costs of supplying synthetic crude oil from Canadian bitumen deposits. The results show the uncertainties associated with the future costs of synthetic crude oil. Carbon costs have a large impact of the total costs of synthetic crude oil, in particular in the case of synthetic crude oil from in-situ bitumen, due to the carbon-intensity of the recovery techniques. The influence of each parameter on the supply costs is examined. In the case of mined SCO, the maximum production rate, the ultimate recovery rate and the depletion parameters show the largest influence on the results, while learning parameters dominate in the case of in-situ SCO.