Robustness analysis can support long-term planning, design and operation of large-scale water infrastructure projects confronting deeply uncertain futures. Diverse actors, contextual specificities, sectoral interests, and risk attitudes make it difficult to identify an acceptable and appropriate robustness metric to rank decision alternatives under deep uncertainty. Here, we contribute an exploratory framework to demonstrate how methodological choices affect robustness evaluation. The framework is applied to a multi-actor, multi-sector Inchampalli-Nagarjuna Sagar (INS) water transfer megaproject in Southern India. We evaluate a suite of dynamic adaptive water transfer strategies discovered using evolutionary multi-objective direct policy search (EMODPS), a status quo strategy of no water transfer, and a strategy proposed by regional authorities. We evaluate robustness across wide-ranging scenarios that capture key uncertainties in potential future changes in reservoir inflows and water demands in the basins. Results show that the priorities of different actors, sectoral perspectives, and risk attitude significantly affect robustness rankings of strategies. We found that compromise strategies obtained from EMODPS are better able to balance the diverse performance requirements across various actors and sectors when compared to the no-transfer and proposed transfer. We reveal a key robustness tradeoff between the donor basin’s ecological requirements and the recipient basin’s socio-economic requirements. While robustness analysis is central to water infrastructure planning, we show why exploratory robustness analyses that engage with conflicting stakeholder objectives is vital for long-term sustainability. Furthermore, the selection of compromise solutions should be guided by an explicit understanding of how assumed risk attitudes shape stakeholders’ understanding of consequential vulnerabilities.