Most language toolchains are held together by small teams of volunteers or corporates (Apple, MS) that sell their own IDEs. They struggle to provide even the basics like solid fast compilers, portability, debuggers that aren't gdb, profilers that work and so on. Look at Go, Rust, Swift. All of them are fighting with basics like performance and language evolution. And now you want them to implement most of an IDE as well? How will they handle that? The only real exception is Kotlin (by JB!), but that was possible because the compiler and IDE plugin are a unified codebase
and because they built on the JVM, so they get top quality compilers and profilers, portability layers etc for free.
The key to JetBrains' success is not any particular software architecture, although having language analysis plugins run in-process with direct access to AST structures is an obvious advantage, it's that they're a company that charges money for their tools, which has kept their cost overheads low (very low relative to never-profitable SV startups that snort VC capital like party hounds), and which are building on an already very productive platform to begin with (Java).
This combination means their feature throughput is very high but more importantly, sustainably high. They've been doing this for 20 years and aren't about to hit a tech debt wall where they suddenly need to rewrite their product from scratch.
Contrast this to their primary competitors, Microsoft and Apple.
Visual Studio has a comparable feature set to JB IDEs. But it's Windows only, totally un-portable, hardly supports web dev outside of the .NET stack, and the plugin model is native. They badly struggled with the transition to 64 bit because of this, which in turn meant users regularly hit its internal architectural limitations like project size limits (cuz of the 32 bit address space). VS doesn't scale well because there's no good way to stop plugin crashes taking out the whole IDE. Even though MS have been pushing .NET for 20 years, VS itself doesn't use it. IntelliJ built on Java which hurt them in the early days, but now Java's perf issues are mostly solved and they benefit every day from portable plugins, a seamless 64 bit transition, cross-platform portability, etc.
XCode is under-funded, has very few plugins and can't compete for any task outside of Apple's own ecosystem. It doesn't even try, really.
VSCode is a text editor with plugins. It is at least trying to compete with JB, but I don't forsee much success in the short term, especially if JetBrains pull their finger out on light mode. It at least uses managed languages. But the most mature JB plugins have been developed over a period of decades by now. The feature-set is overwhelmingly huge, and they have hundreds of employees that do nothing else but add them. How much money is MS willing to throw into the VS Code project? Its lack of revenue model limits them, even with a sugar daddy funding them. It's unlikely they can scale to JB size because a free IDE that competes with their own revenue generating VS product is simply not something they will tolerate long term. VSCode is going to be stuck focusing on the JavaScript/web stack for a long time and won't receive a lot of funding to grow out of that.