The complexity and size associated with industrial applications such as the solution of wind-farm flows present a major challenge in the use of high-fidelity simulation data in operation control technologies. Conventional, high-fidelity methods have matured and become an integral part of the design process. However, they are typically characterised by prohibitively large computational cost, unfit when considering parametrised studies and time-critical applications. An alternative, in the form of reduced order models (ROMs) aims to overcome this bottleneck by approximating the complex physical system both accurately and efficiently. This work explores the applicability and resulting simulation speed-up of a hybrid ROM strategy in wind-farm applications. Initially, it explores its viability for the efficient solution of turbulent flow around wind turbine blades. The same methodology is then adapted for geometric parametrisation of the blade profile, and finally, it formulates a strategy for the construction of a ROM wind-turbine component that would replace its high fidelity counterpart in the wind farm, resulting in a nested chain-ROM set-up, where ROM components interact with each other. The hybrid ROM combines data-driven and projection-based techniques. Velocity and pressure are approximated using the proper orthogonal decomposition (POD) with Galerkin projection, while eddy viscosity is approximated using POD plus interpolation. When adapted for the chain-ROM, special care is required for the reduction of the parametrisation complexity and the handling of interfaces/nested boundaries. Finally, various industrially-relevant examples are chosen to show the capabilities of the proposed ROMs.