This paper investigates the application of quantum and quantum-inspired machine learning algorithms to cross-sectional stock return prediction. The authors compare quantum neural network (quantum circuit learning, QCL) and tensor network (matrix product state, MPS) models against classical benchmarks including linear regression and feed-forward neural networks. Using a dataset of Japanese stocks over a 10-year period, they construct portfolios based on model predictions and evaluate investment performance. The tensor network model achieves superior risk-adjusted returns, outperforming both linear regression and neural networks. While the quantum neural network underperforms the classical neural network overall, it shows competitive performance in the latest market environment, suggesting its ability to capture non-linear feature interactions. The study provides empirical evidence that quantum-inspired algorithms can be effectively applied to financial prediction tasks, highlighting their potential advantages in capturing complex patterns. The results also indicate that tensor networks, being classically simulable, offer a practical alternative to quantum computers for near-term applications.