Abstract Hybrid ground-coupled heat pump systems can efficiently heat and cool buildings by exchanging heat with geological materials. Their sizing is, however, complex and their financial profitability is hard to establish during the design phase due to uncertainties that taint important design parameters. Historically, impacts of uncertainties are assessed after sizing completion using sensitivity analyses. Unfortunately, these analyses cannot mitigate design risks, should the consequences of uncertainties be significant. Here, we show how the net present value-at-risk, a stochastic financial indicator inspired by a metric used in the financial sector, can weave the impacts of uncertainties throughout the design phase. The net present value-at-risk is compared to the net present value in a case study that considers uncertainty of construction costs, building’s heat load, energy tariffs and ground thermal conductivity. Results show that this financial indicator, although it comes at a higher computational price than traditional net present value indicator, leads to shorter payback periods, greatly reduces the risks of unforeseen financial losses and does not require further sensitivity analysis. By describing uncertain parameters with statistical distributions during sizing, the proposed designs are more conservative, still efficient and financially viable while being shielded from worst case scenarios.