In this paper, the fundamental limits of simultaneous information and energy transmission (SIET) are studied in the non-asymptotic block-length regime. The focus is on the case of a transmitter simultaneously sending information to a receiver and energy to a harvester through binary symmetric channels. Given a finite number of channel uses (latency constraint) as well as tolerable average decoding error probability and energy shortage probability (reliability constraints), two sets of information and energy transmission rates are presented. One consists in rate pairs for which the existence of at least one code achieving such rates under the latency and reliability constraints is proved (achievable region). The second one consists in a set whose complement contains the rate pairs for which there does not exist a code capable of achieving such rates and satisfying both latency and reliability constraints (converse region). These two sets approximate the information-energy capacity region, which allows analyzing the trade-offs among performance, latency, and reliability in SIET systems.