Tilmann Scheller, a compiler engineer for Samsung's Open-Source Group / Samsung Research UK, presented at this weekend's FOSDEM conference in Brussels about where LLVM is being used today. This isn't anything entirely new, and much of it has been covered on places like Phoronix in the past with all of the innovative LLVM use-cases, but his PDF slide deck does a nice job of being a concise overview and highlighting the many areas where LLVM is utilized.
Companies from Microsoft to NVIDIA and Google rely upon LLVM and contribute upstream while Apple has been traditionally the one most well known for backing the project. AMD, Sony, Intel, Linaro, and IBM are among the many others.
Clang is well known as the C/C++ front-end to LLVM, but there is also language support for Swift, Haskell, Ruby, Python, Lisp, D, Go, Rust, Julia, and other languages. LLVM is also used for language runtime systems by projects like Microsoft's LLILC .NET compiler, Mono, and OpenJDK.
LLVM gets GPU exposure via NVIDIA's CUDA, Mesa LLVMpipe, LunarGLASS, the AMDGPU open-source driver stack, SPIR / SPIR-V, and a majority of the OpenCL implementations in the world. Web projects around LLVM include Google's Portable Native Client (PNaCl), WebKit FTL JIT, EmScripten, and WebAssembly, among others.
LLVM has also seen integration with various sanitizers for catching bad/faulty code, integration with development IDEs like Xcode and KDevelop and Qt Creator, and more. There is also LLVM implementations surrounding binary translation, symbolic execution, and more.
See Tilmann Scheller's slide deck for more examples of the wide ranging world of LLVM.