This research offers a theoretical model of brokered services and provides an analysis of their impact on the cloud computing market with risk preference-based stratification of client segments. The model structures the decision problem that clients face when they choose among spot, reserved and brokered services. Although all the three types of services do not indemnify the cloud services client against other kinds of service outages, due to changes in market demand, service interruptions occur most frequently in the spot market, and are lower when brokered services are offered, and no risk of inter-ruption is involved in reserved services. Based on our analysis, we show that the profitability and sus-tainability of a cloud service broker depends on its usage of reserved resources and its capability to mitigate the risk of interruptions. We further enrich our explanation through the consideration of the distribution of clients’ risk preferences and the service vendor’s pricing decisions for reserved and spot services.