Digital twin technology has progressed from a conceptual framework articulated in the early 2000s to a mature engineering discipline underpinning modern manufacturing, energy, healthcare, and smart-city deployments. A digital twin is a synchronised virtual representation of a physical asset, process, or system, continuously updated by sensor data and used for monitoring, simulation, and optimisation. This paper surveys digital twin architectures, the integration of Internet-of-Things sensing with edge and cloud computing, the role of artificial intelligence in twin construction and inference, and representative applications across Industry 5.0 sectors. We discuss interoperability standards, security concerns, and the open research questions that will shape the next decade of cyber-physical systems.