You used to be able to buy low-end SPARC systems for $995 (v100, Blade 100) now the cheapest system is much more than that - the same pricing strategy that ruined SGI: as their market share declined they retreated to higher end customers, which led to more market share declines.
Sun used to make really high-quality equipment, but now they are making crappy PCs like everyone else - maybe they should have focused on the customers willing to pay for quality.
Schwartz's task was essentially impossible, and selling to IBM or Oracle was probably the best outcome he could have hoped for.
Companies die if there isn't an entry point that makes sense for customers.
It doesn't have to be at the same price point as the folks eating them from below, but it does have to be some place where customers are willing to enter.
The mid/high-range sun systems are about as expensive as comparable PC based systems. However, the PC systems have a low cost option for development. Suns, not so much.
It doesn't have to cost $500, but it can't cost $10k either.
Which lead to decreased network effects that, in turn, hurt the viability of their whole stack (SPARC+Solaris+apps)
You have to sell something entry-level unless you want to be servicing legacy customers forever.
I believe IBM is making with POWER/AIX exactly the same mistake Sun made. Fortunately for them, their legacy customers are less loyal than Sun's were.
In order to sell servers a year in the future, you must captivate developers right now.