A solution for server-side rendering includes, at a server configured to store a video images representing states of users in a computer application, identifying future user actions based at least in part on a state of a user in the computer application. The solution also includes, responsive to the identifying, rendering video images for sending to a user device associated with the user. At the user device, a state of the user in a computer application is sent to the server. Responsive to the sending, video images are stored, each of the video images representing a future state of the user after the user performs a future action. Responsive to a user action, one of the video images is selected for display on a user display of the user device. According to one aspect, the future user actions identified by the server are limited to less than a possible number of user actions for users having the state.