OLPC seems like a good example of spending too many "innovation tokens" at once. Very difficult to succeed when making a new system with unique hardware components and new software all at once.
Pixel Qi was always one of those companies I thought should have succeeded but didn't.
This a decent walk through of their timeline in tech history:
https://goodereader.com/blog/electronic-readers/the-rise-and...
Still today I would love to have a minimalist 'typewriter' laptop that I could work with just sitting outdoors in full sun.
From my perspective, the point was the software. Getting that software to markets with unreliable electricity and little money would have been a problem without a new type of hardware device. In the end though, it was the hardware that dominated the consciousness, which was too bad. OTOH, if they hadn't tried, that might mean something like 2+ million kids might have had a worse education (assuming that the OLPC machines were useful to their education), so I doubt it is fair to call it a failure.
It definitely wasn't aimed at US or European mainstream success, nor was it even really offered in those markets, so lack of success in those markets is expected.