My personal view is that Solaris is now, like Oracle's database products, officially irrelevant to the future. While they will continue to be purchased for large sums of money by people who must, like COBOL, they will not shape anyone's idea of the future. ZFS had a huge impact on file system development, but I suspect no such future innovation is in store for Solaris.
I will always have a soft spot for Solaris because, in the 90s I wrote and maintained threaded server code- Solaris had the most performant and least problematic threading implementation that we ran on at the time. Really though, why would anyone use Solaris now? It doesn't have any real strengths over the now much improved free un*x variants. Most of the 'crown jewels', dtrace, zfs etc are available some place else.