The field of swarm robotics has attracted considerable interest for its capacity to complete intricate and synchronized tasks. Existing methodologies for motion planning within swarm robotic systems mainly encounter difficulties in scalability and safety guarantee. To address these two limitations, we propose a Risk-aware swarm mOtion planner using conditional ValuE at Risk (ROVER) that systematically modulates the safety and conservativeness and navigates the swarm to the target area through cluttered environments. Our approach formulates a finite-time model predictive control (FTMPC) problem predicated upon the macroscopic state of the robot swarm represented by Gaussian Mixture Model (GMM) and integrates conditional value-at-risk (CVaR) to avoid collision. We leverage the linearized Signed Distance Function for the efficient computation of CVaR concerning the proximity between the robot swarm to obstacles. The key component of this method is implementing CVaR constraint under GMM uncertainty in the FTMPC to measure the collision risk that a robot swarm faces. However, the non-convex constrained FTMPC is nontrival to solve. To navigate this complexity, we develop a computationally tractable strategy through 1) an explicit linear approximation of the CVaR constraint; and 2) a sequential quadratic programming formulation. Simulations and comparisons with other approaches demonstrate the effectiveness of the proposed method in flexibility, scalability, and risk mitigation.