⚡NVIDIA announced the introduction of CUDA Rust, allowing developers to write GPU kernels in Rust.

What it means

NVIDIA has officially announced CUDA Rust, a new tool/library that enables developers to write GPU kernels using the Rust programming language instead of the traditional CUDA C/C++. This bridges Rust's memory safety guarantees with NVIDIA's GPU computing platform, likely offering two distinct tracks (e.g., a high-level and a low-level approach) for kernel development. The announcement positions Rust as a first-class language for GPU programming in the CUDA ecosystem.

Accuracy & sourcesGrounded

The claim is based directly on NVIDIA's official developer blog, making it a primary source and authoritative for the announcement itself. The blog post is the original disclosure, so the accuracy of the stated fact—that NVIDIA is introducing CUDA Rust—is near-certain, though the specific details of the two tracks are inferred from the headline and haven't been independently verified here.

How new

Genuinely new: the announcement was published on September 8, 2026, with no indication of older material being resurfaced.

Go deeper

• NVIDIA Developer Blog: Introducing CUDA Rust