This article proposes a computation offloading method to improve the revenue of a virtual power plant (VPP). The cloud-edge collaboration framework is developed to facilitate the computation offloading task, while handling the communication and computational delays encountered in the offered frequency regulation. The impact of such time delays on a VPP’s frequency regulation is worth further investigation. To this end, this study developed a revenue optimization model for a VPP based on the related regulation performance. An evaluation model was also built to account for the impact of time delay on a VPP. Both models are used to quantify the benefits of computation offloading to boost the revenues of the VPP. Moreover, a necessary condition for using the computation offloading technique is proposed. A greedy algorithm is designed to make the offloading decisions. Numerical studies demonstrated the proposed offloading method could improve VPP revenue by more than 31% compared to the method without computation offloading.