Ask HN: Projects failed because of TDD / too much testing

I recently thought about this interesting question: there are some discussions pro/con TDD and such, some with examples how testing can be made bad(sure, you can apply every tool the wrong way) and why 100% coverage are stupid (I partly agree because there might reasons why some parts of your software are very unlikely to fail and super hard to write tests for - and 100% coverage doesn't mean you considered all edge cases yet). But I don't know of any real life reports of projects that actually failed because people did TDD there, or in general wrote too many tests, a project became unmaintainable because of bad tests or something like that. In my personal experience there are some examples of the contrary where lack of tests made a project unmaintainable - or having them made me think much better about the problems at hand and write better code in turn.

So, do you have any concrete experiences where you say it would have rescued a project or made it more maintainable if the wouldn't have done TDD / write less or no tests?