Video telephony (VT) call management techniques are described. The techniques enable a device to cache intra-frame data at a pre-decoder-initialization stage. An example device includes a memory configured to store video data associated with a VT call, a video decoder configured to render a portion of the stored video data; and one or more processors. The processor(s) are configured to determine whether the received video frame data comprises i-frame data, to determine whether the video decoder is in a pre-initialized state or an initialized state, and when the received video frame data comprises the i-frame data and the video decoder is in the pre-initialized state, to store the i-frame data to the memory.