Machine learning (ML) offers a faster and more efficient approach to simulating spatio-temporal dynamics in weather forecasting. ML weather emulators can generate large forecast ensembles, providing a unique opportunity for studying low-likelihood, high-impact events like heatwaves. This study demonstrates the added value of a huge ML ensemble in resolving extreme heatwave statistics for the summer of 2023. The ensemble provides more robust sampling, better predictions, and quantifies worst-case scenarios of compound risk from extreme heatwaves.