Since the first detection of gravitational waves emitted during the merger of two black holes, understanding the formation channels of such systems has become one of the most pressing questions in theoretical astrophysics. In the 19 pioneering Dragon-II simulations published this year and which I will present, we have demonstrated that the black holes the merge in our simulations can fit the observed gravitational wave sources very well. This has many implications for the entire field of computational astronomy, but also astrophysics in general. Unlike observations, we can resolve the entire lifecycle of the black holes (and other compact objects) that are gravitational wave sources from birth to eventual demise. This helps us to unravel one of the fundamental questions in astrophysics of how massive black holes form and gets us close to the question where supermassive black holes originate from. I will also highlight results from an as of yet mostly ignored parameter in such simulations, which is initial star cluster rotation and which yields exciting phenomena. Lastly, I will provide insights into the evolution of the first (Population-III) star clusters in the Universe and first results from growing actual supermassive black hole seeds within star clusters.