Electric vehicle aggregators (EVAs) play an indispensable role in the integration of EV resources and load management. In the electricity spot market, EVAs can leverage the flexibility of EV charging to reasonably schedule EV loads and reduce their operational costs. In general, the electricity spot market consists of the day-ahead and real-time markets. In the day-ahead bidding stage, EVAs ought to submit electricity bids in advance to lock in the EV charging requirement for the next day. As the day-ahead bidding process accounts for the majority of the EVA’s operational costs, it is crucially important. Nevertheless, the randomness of EV charging and the uncertainty of spot market operation would pose substantial challenges to the day-ahead bidding decision of EVAs. Most existing studies adopt a scenario-based approach to address uncertainty and derive day-ahead bidding strategies of EVAs to minimize expected costs or maximize expected profits. However, such aggressive bidding strategies may be insufficient to hedge against operational risks caused by market price volatility, exposing EVAs to considerable financial losses. To tackle the above issues, this chapter introduces Conditional Value-at-Risk (CVaR) as a risk metric to cope with the uncertainty in the real-time clearing market, optimize the CVaR of EVA bidding, and mitigate potential market risks. On this basis, a risk-averse day-ahead bidding method for EVAs is proposed. First, a day-ahead bidding model for EVAs under deterministic information is established, and the corresponding nonlinear programming model is linearized. Furthermore, to capture the stochastic nature of EV charging, a probabilistic model is employed to simulate the charging behavior of the EV fleet for the next day. Aiming at the uncertainty of the real-time electricity market, a risk-averse decision-making approach is utilized to model potential real-time clearing scenarios, and a day-ahead bidding model with the objective of minimizing CVaR is formulated. Finally, the proposed risk-averse bidding strategy is compared with the risk-neutral one, and the superiority of the risk-averse bidding strategy in risk management is demonstrated. Results show that higher confidence levels (e.g., 0.95 or 0.99) effectively constrain costs within predefined intervals with minimal variation in optimized CVaR values. Furthermore, the linearized model significantly outperforms the original non-linear formulation in computational efficiency, especially for a large EV fleet. This chapter concludes that the risk-averse bidding approach can provide superior risk management for EVAs, though future research should address real-time scheduling complexities and computational speed for practical implementation.