The paper discusses the ongoing efforts to port the Quantum ESPRESSO software, a widely used open-source tool for quantum-mechanical materials modeling, to heterogeneous architectures based on hardware accelerators. The goal is to overcome energy constraints and enable exascale computing. The authors describe the history and current status of Quantum ESPRESSO, highlighting its features and capabilities. They then address the challenges and opportunities posed by new heterogeneous architectures, emphasizing the need for performance portability and sustainable development models. The paper outlines the strategy of refactoring Quantum ESPRESSO into multiple layers of modules and libraries to enhance maintainability and portability across different architectures. It also details the development of low-level system libraries and domain-specific mathematical libraries to support performance portability. The paper concludes with a discussion of the GPU-enabled version of Quantum ESPRESSO, its performance, and future plans for porting to other accelerated architectures.The paper discusses the ongoing efforts to port the Quantum ESPRESSO software, a widely used open-source tool for quantum-mechanical materials modeling, to heterogeneous architectures based on hardware accelerators. The goal is to overcome energy constraints and enable exascale computing. The authors describe the history and current status of Quantum ESPRESSO, highlighting its features and capabilities. They then address the challenges and opportunities posed by new heterogeneous architectures, emphasizing the need for performance portability and sustainable development models. The paper outlines the strategy of refactoring Quantum ESPRESSO into multiple layers of modules and libraries to enhance maintainability and portability across different architectures. It also details the development of low-level system libraries and domain-specific mathematical libraries to support performance portability. The paper concludes with a discussion of the GPU-enabled version of Quantum ESPRESSO, its performance, and future plans for porting to other accelerated architectures.