But does it happen that startups fail because their technology is poor?
Yes, sure, happens all the time. Healthcare and biotech are especially hurt by this (nobody wants a medicine that doesn't work). Software companies get hit by this too, though the effect is often less obvious.
Cuil, the much hyped Google killer, is dead now IMHO. They hyped themselves to the stars, and then delivered a crappy first release, making everyone feel let down and negative towards their brand. Arguably it's not a 100% pure technology failure, it's also about wrong marketing, but its still a technology-driven failure.
I see plenty of articles that say "ship first, no matter how bad".
No, they don't say that (or they shouldn't say that). They say smartly define a Minimum Viable Product, and then iterate on it. Release the first version in a state where you still have some work to do, but target your marketing so that the only people who see the initial release are early adopters who are somewhat forgiving of problems. In other words you compromise, but mostly in the areas of feature scope and marketing, and only a little on quality (which you then aim to rapidly improve).
Does this imply there's no real place in a startup for proper software engineering principles that may reduce the time to market, but make the solution easier to grow?
No. But getting the compromise just right is damn hard, you don't want premature optimization either.