Disclosed are various embodiments for facilitating load balancing between central processing units (CPUs) and graphics processing units (GPUs). A request is obtained to execute a first application in one or more computing devices. In one embodiment, a second application associated with the first application is assigned to be executed in GPUs of the one or more computing devices instead of CPUs of the one or more computing devices when a resource usage profile associated with the first application indicates that the first application is relatively more CPU intensive than GPU intensive. Conversely, the second application is assigned to be executed in the CPUs instead of the GPUs when the resource usage profile indicates that the first application is relatively more GPU intensive than CPU intensive.