When a project has absorbed many resources, interrupting it may seem like a further loss. However, the now non-recoverable expenses do not change the comparison between future alternatives. The evaluation should start from what each choice will require and can produce from today.
In a hypothetical example, a purchased software needs more work to become usable. Continuing can be reasonable if the expected benefits outweigh additional costs and risks. The mere desire to justify the previous purchase does not prove this convenience.
It is necessary to distinguish irreversible expenditures from resources that are still recoverable. A sellable asset, a modifiable contract, or reusable work can affect the comparison. Obligations already assumed and consequences of the interruption must also be considered on the basis of the actual conditions.
I propose presenting alternatives with the same departure date, separating past costs and future consequences. This does not erase the responsibility of previous decisions: it allows to evaluate them at a distinct time. The current choice should use the available information to limit further losses or recover value, without being driven solely by the difficulty of recognizing a project that is not working.