Approaches presented herein enable optimization of a plurality of resources connected within an Internet of Things (IoT) network. More specifically, usage of a first and a second resource of the plurality of resources connected within the IoT network is tracked. This usage comprises a task and a time at which the task is performed for each of the two resources. An anticipated task at an anticipated time based on the tracked usage of the first resource is identified. An anticipated capacity at substantially the same anticipated time is also identified on the second resource based on the tracked usage of the second resource. Excess capacity is generated on the first resource by transferring the anticipated first task to the second resource for the anticipated time. The excess capacity is designated for use by an arm's-length party. A load from the arm's-length party is then assigned to the excess capacity.