NVIDIA shipped native Rust GPU kernel programming in two tracks, one already on stable.

CUDA Rust compiles Rust directly to PTX rather than wrapping C++ or Python. The SIMT track (cuda-oxide) is early alpha and needs a pinned nightly plus custom LLVM; the tile track (cutile-rs) runs on stable Rust 1.89+, is on crates.io, and is already used by Hugging Face's Grout and mistral.rs. Both need CUDA 13.3+ and compute capability 8.0+, and both enforce memory safety at compile time to catch the aliasing bugs behind GPU race conditions. Paired with GitHub's 800k-line Rust port above, the week's Rust story is not incidental.