Abstract-Wireless Power Transfer (WPT) is expected to be a technology reshaping the landscape of low-power applications such as the Internet of Things, RF identification (RFID) networks, etc. To that end, multi-antenna multi-sine waveforms adaptive to the Channel State Information (CSI) have been shown to be a promising building block of WPT. However, the current design is computationally too complex to be applied to large-scale WPT, where the transmit signal is sent across a large number (tens) of antennas and frequencies. In this paper, we derive efficient singleuser and multi-user algorithms based on a generalizable optimization framework, in order to design transmit waveforms that maximize the weighted-sum/minimum rectenna DC output voltage. The study highlights the significant effect of the nonlinearity introduced by the rectification process on the design of waveforms in single/multi-user systems. Interestingly, in the single-user case, the optimal spatial domain beamforming, obtained prior to the frequency domain power allocation optimization, turns out to be Maximum Ratio Transmission (MRT). On the contrary, in the general multi-user weighted sum criterion maximization problem, the spatial domain beamforming optimization and the frequency domain power allocation optimization are coupled. Assuming channel hardening, low-complexity algorithms are proposed based on asymptotic analysis, to maximize the two criteria. The structure of the asymptotically optimal spatial domain precoder can be found prior to the optimization. The performance of the proposed algorithms is evaluated. Numerical results confirm the inefficiency of the linear model-based design for the single and multi-user scenarios. It is also shown that as nonlinear modelbased designs, the proposed algorithms can benefit from an increasing number of sinewaves at a computational cost much lower than the existing method. Simulation results highlight the significant benefits of the large-scale WPT architecture to boost the end-to-end power transfer efficiency and the transmission range.