AbstractThis article examines disruption risks at fulfillment centers and develops risk mitigation strategies based on inventory stocking and delivery decisions. It considers a Fortune 150 firm whose delivery operations are designed to fulfill the orders from contracted business customers within the next day. The firm promises its customers that the probability of late deliveries exceeding a certain threshold will be limited. We coin this requirement as the service‐at‐risk (SaR) constraint. The firm proactively determines the inventory amount to be kept in each fulfillment center. If a disruption occurs, the firm determines the best way to deliver orders from its operational fulfillment centers and vendors under disruption length and demand uncertainty to minimize additional costs and satisfy the SaR constraint. This article makes four main contributions. First, we find a surprising result that total inventory commitment can decrease with risk aversion when there exists a disruption possibility that impacts two nearby facilities together. Using actual data from the motivating firm, the numerical analysis demonstrates that this phenomenon exists in practice. Second, we define a new metric: The Risk Dispersion Index (RDI), which measures the dispersion in risk exposure across fulfillment centers. It leads to a lower and more balanced risk exposure in the firm's delivery operations. Third, we find that a facility may elect to abandon its own customers to serve the customers of a disrupted facility; this behavior becomes more prominent under risk aversion. Fourth, the introduction of demand uncertainty leads to a smaller inventory commitment for a risk‐neutral retailer.