There were always those legacy customers from years before who had there stuff running on some boxes in an old datacenter that only a few of the original staff knew about, and required continuous care and attention. Ihere's a business lesson to learn there - if you sell a "lifetime" subscription for a product that will become obsolete, you'll pay a lifetime PR cost in the future when you can't continue with that product.
Coming from a pretty long Linux background, Solaris took some getting used to, but once I figured it out, it really was a better, more solid, better engineered solution. But Linux has a huge community - the people that loved Solaris are the old Unix graybeards, and that community wasn't growing.
Before my time, there were some debacles, such as a storage platform product that ended up losing data, but by the time I was there, they had learned the lessons, and I'll still say their core stack is still superior to any Linux based stack used in the datacenter today. It's been 4 years since I worked there.
In their shoes I think the lesson I would learn is "don't go back on a deal". They could have returned the money, transitioned people to another product, shifted people to a different vendor, or offered people the option of any of the three. I don't think it was initially a PR problem; it was bad behavior, unilaterally breaking a promise, that got them in hot water.
I'm sure he's a bright and likable guy, and from his resume I figure he must be an inspiring leader. But for me, being a great technologist is all about things actually working, about users actually getting served. Sure, one should fail a lot in the lab, and sure, startups should be trying enough things that some just don't have product/market fit. But that's very different to me than promising the moon and delivering stuff that doesn't actually work.