We know x86 has become much more successful than SPARC over the past 20 years, even on the server side. This in turn translates into more budget in R&D on x86. These external factors aside, is there any inherent reason why SPARC has failed? Before we begin, it would help to read John Gustafson's answer to the question about their difference (https://www.quora.com/What-are-the-differences-between-Sparc-and-Intel-architecture). In short, there's not much difference nowadays, including RISC vs CISC. He thinks pretty much the only difference is in endianness. So, the question is, why is x86 a success but SPARC is not, from a pure technical point of view? If there isn't any, that's perfectly fine, because it's not uncommon for a technically superior technology to fail in business.
SnapOverflow
You can't reduce this to just a technical difference in the architecture. There is also politics and economics and marketing.
If more market share and revenue for x86 makes a good product, why do a few clouds have ARM instances, and Apple Silicon is transitioning away from x86? A complete and useful instruction set is a necessary but not sufficient condition for a CPU to be successful. Also necessary: leadership committed to the multi-year project, investment in talented people, access to the best fabrication facilities, and marketing the resulting chips with a clear message. Intel and AMD execution here is quite different circa 2020, and they both are x86.
Back to SPARC, is an architecture a failure if it powered a system on the TOP500 list for 8 years (K computer)? Or if the open instruction set was implemented in radiation hardened chips that went to space? And Fujitsu is still hacking on SPARC.
Admittedly, these use cases do you no good if you want commodity Linux servers. RHEL already has x86, ARM, and POWER, with IBM championing POWER I don't what justifies spending effort to add SPARC. Debian's sparc64 port shows it is possible to build it, but the list of supported hardware is short and aging. Needs an organization to champion the effort, again this as much politics as it is engineering.