In this paper, we analyze the performance of the uplink communication of massive multi-cell multiple-input multiple-output (MIMO) systems under the effects of pilot contamination and delayed channels because of terminal mobility. The base stations (BSs) estimate the channels through the uplink training, and then use zero-forcing processing to decode the transmit signals from the users. The probability density function (PDF) of the signal-to-interference-plus-noise ratio is derived for any finite number of antennas. From this PDF, we derive an achievable ergodic rate with a finite number of BS antennas in closed form. Insights of the impact of the Doppler shift (due to terminal mobility) at the low signal-to-noise ratio regimes are exposed. In addition, the effects on the outage probability are investigated. Furthermore, the power scaling law and the asymptotic performance result by infinitely increasing the numbers of antennas and terminals (while their ratio is fixed) are provided. The numerical results demonstrate the performance loss for various Doppler shifts. Among the interesting observations revealed is that massive MIMO is favorable even in channel aging conditions.