This letter studies the performance of uplink cross layer design for multi-antenna systems with outdated channel state information (CSI). We consider a multi-user system with one base station (with nR receive antennas) and K mobile users (each with single transmit antenna). The multi-user physical layer is modeled based on information theoretical framework and the cross layer design can be cast as an optimization problem. In view of the high computation complexity in the optimal solution, we propose a low complexity genetic algorithm as suboptimal solution. We found that with outdated CSI, there is significant degradation in the spatial multiplexing and multi-user diversity gain due to potential packet transmission outage as well as misscheduling. To address the poor performance in the presence of outdated CSI, we propose two simple but effective empirical solutions, namely the rate quantization and rate discounting, to tackle the packet outage problem. For instance, it is well-known that rate quantization imposes system capacity loss in systems with perfect CSI. However, we found that rate quantization can enhance the robustness of system capacity with respect to outdated CSI.