Abstract: In general, techniques are described by which to perform memory efficient patching for computing devices. A server computing device comprising a memory and a processor may be configured to perform the techniques. The memory may store first assets that form an unpatched application and second assets that form a patched application. The processor may virtualize the first assets to obtain a single first virtual asset and obtain a single second virtual asset that represents the second assets. The processor may obtain, based on differences between the single first virtual asset and the single second virtual asset, a patch identifying how to update the single first virtual asset to obtain the single second virtual asset. The processor may next segment the patch into a plurality of segments and output to a user computing device a single segment from the plurality of segments for individual application by the user computing device.