The only other thing might have been ZFS, but IBM already had/has lots of their own midrange and high-end storage software and systems that do basically the same thing.
Indeed, including the publicly known jfs (sometimes called jfs2), which ships as part of AIX. And then GPFS (now known as Spectrum Scale), SanFS (previously known as StorageTank), Seastar/Squid (publicly known as Shark), and I forgot a few other nautical code names (like what was Lodestone). None of those exactly match ZFS, and importantly none are available for free.
If Sun tried going back into business with Sparc-based systems now, they would face all the same problems as before. It is extremely costly and difficult to develop, market and then maintain a non-standard CPU architecture, against a strongly entrenched market standard. IBM got nowhere with POWER, once Apple dropped it,
IBM continues to have three non-x86 architectures, and it makes billions by selling them: Power (which still exists in high-end machines, just not consumers), AS/400 (now known as I series, and running on a new processor that is closely related to Power), and the 360/370-style mainframe (now known as Z). It is perfectly possible to develop and sell CPUs, just not into the super price-conscious consumer segment. One just has to be ruthlessly competent.
And while I can say many bad things about IBM (having spent nearly two decades inside it), they are definitely competent. Unlike a company that had Scott McNealy as CEO (chief lying officer, he's the mold from which Elon Musk was made, including all the verbal self-harm), and Andy Bechtolsheim as chief architect (who famously has the strongest RDF = Reality Distortion Field in Silicon Valley: he actually believed his own nonsense, and forced everyone in Sun to also believe it).
Furthermore there have been many attempts to put ARM variants into both desktop PC's/laptops and servers, the only company that has really succeeded in doing that in my opinion has been
Today, a good fraction of all CPUs at places like Amazon/Google/Microsoft are not x86, but are "home brew", using instruction sets such as ARM, MIPS, RISC-V and so on. When I say "home brew", that usually means co-designed with outside vendors, for example Ampere.
In the consumer world, one gets a very biased and unrealistic view of the microprocessor state of the art. That's because consumer devices (cell phones, small computers, video cards for gamers) are optimized for cost and things like battery life, not for reliability, compute power, IO, or good integration into large systems. In particular, the consumer CPU front was so long dominated by Intel (and its licensee AMD) that the x86 architecture has become synonymous with consumer computers. Our (now sadly deceased) neighbor Nick had a good joke about this. Nick was a famous CPU architect, and when giving the humorous after-dinner-and-drinks talk at the Microprocessor forum, he once needled Andy Grove, who had claimed that he didn't interview CPU architecture job candidates, because he himself didn't understand architecture. The punchline is something like "and it shows".
en.wikipedia.org
Perhaps with Intel waning in recent years there will be more opportunities for other architectures to take some business away from x86.
. Oh well, this is just chewing over history.
