Some other companies do sponsor work at times as well. A recent hot topic of discussion is “does a Rust foundation make sense”, there are good arguments for and against. We’ll see.

Oh, and note that that number is people who are decision makers in the project; about 120~150 distinct people contribute to each six-week release. We’ve had 2786 distinct people contribute. And that’s only to rust-lang/rust; not Cargo, not any of the libraries the team maintains, not any of the other kinds of work that doesn’t produce a commit.

To put this into more context: a lot of the code shipped with the Rust toolchain does not live in rust-lang/rust.

Some parts of the standard library (HashMap, HashSet, C FFI, intrinsics, etc.) are used as dependencies. All of the tools (the linter, language server, rustfmt, rustup, cargo, etc.) are dependencies. Many compiler components like miri (constant evaluator), chalk (type queries), polonius (borrow checking), etc. are also dependencies. The Cranelift code generation backend can be dynamically used and does not live in rust-lang/rust, etc. Then there is the RFC repository, and probably other things.

Counting contributors for such a distributed infrastructure isn't easy, but the number of contributors to rust-lang/rust isn't super meaningful.

One might be able to do better by finding commits to rust-lang/rust that update "Rust" submodules (e.g. not LLVM), and adding the contributors to those repos from the last update. One might also want to find commits that update versions of rust-lang/rust dependencies, and for those that live in the rust project (or that have dependencies there), count the contributors from the last version, etc.

The Rust project has seen significant modularization over the last couple of years, and the more modular it gets, the less meaningful "contributors to rust-lang/rust" becomes, and the less accurate the "Thanks!" pages of each release become.