We study the design of procurement contracts when the buyer faces uncertainty about both the value of the good and the seller’s cost. The buyer holds a conjecture— based on a model estimated or calibrated from past data—but does not fully trust it. They first identify all worst-case optimal mechanisms, which maximize the buyer’s guaranteed payoff across a set of plausible value and cost specifications. Among these, the buyer then selects the mechanism that maximizes their expected payoff under their conjecture. We show that robustness leads the buyer to procure larger quantities from the least efficient sellers and smaller quantities from those with intermediate costs, relative to the optimal mechanisms under full trust in the model. Finally, we apply our framework to monopoly regulation and identify conditions under which quantity regulation dominates price regulation.