Motivated by the need for cities to prepare for and adapt to climate change, we advance the paradigm of safe-to-fail by focusing on the decision dilemmas and the consideration of infrastructure failure consequences in developing infrastructure. Infrastructures are largely designed as fail-safe; that is, they are not intended to fail, and when failure happens, the consequences are severe. Safe-to-fail has been recently presented as the antithesis of fail-safe, without any specific guidance of what the paradigm is or how to apply it. There is an emerging need for stakeholders, including policy makers, planners, engineers, utilities, and communities to understand infrastructure failures, bring their knowledge into the infrastructure development process, and help adapt cities to unpredictable and changing climate risks. We frame safe-to-fail as an infrastructure development paradigm that internalizes the consequences of infrastructure failure in the development process. This framing of safe-to-fail further reveals an emerging "infrastructure trolley problem" where the adaptive capacity of some regions is improved at the expense of others. We demonstrate practical dilemmas in developing infrastructure under nonstationary climate and guide managing trade-offs in the prioritization of different consequences of infrastructure failure.