Purpose. The 2010 to 2016 cohort of United States utility-scale and community solar plants is reaching the age at which central inverters leave warranty, manufacturers withdraw support and parts become scarce; the paper determines which of repowering, repair and run to failure minimises expected life-cycle cost and how the funding route constrains the choice. Design. A critical review with framework synthesis of 270 sources is condensed into five equations, a decision matrix, a four-gate flowchart and a reproducible worked example for a hypothetical 5 MWdc plant with a stated parameter set and a sensitivity analysis. Findings. Published fleet studies predate the wear-out phase, so hazard must be extrapolated from mission-profile models. In the example run to failure is cheapest only for remaining terms of three years or less, repair for four to ten years and repower from eleven years; the switch points move by several years with the discount rate, failure probability, price and repower cost, and the repair window closes when failure probability is halved. For community solar gardens the funding gate is reached before the engineering gate. Originality. To the authors’ knowledge no prior study combines the field reliability evidence, maintenance economics, investment tax credit recapture and retrofit rules and project-finance covenants in one decision framework; the nearest models omit the funding and tax terms.