NVIDIA Rolls Out AI Coding Tools to 30,000 Engineers, Tripling Code Output

IT Home reported on February 8 that NVIDIA—one of the key forces behind today’s AI infrastructure and developer boom—is now seeing major internal gains from the same wave it helped ignite.

According to Cursor’s official blog, NVIDIA has rolled out generative AI tools across its organization for as many as 30,000 engineers. Through a partnership with San Francisco–based Anysphere Inc., NVIDIA gained access to a customized version of the Cursor IDE, built specifically around AI-assisted code design.

The impact has been striking: NVIDIA engineers are reportedly producing roughly three times more code than under previous development workflows. In other words, many NVIDIA products and services you use today may well be created through a “human-led, AI-assisted” engineering process.

NVIDIA’s software and platform stack is mission-critical, spanning everything from GPU drivers that power everyday gaming to large-scale AI training and inference. That reality suggests the company applies rigorous standards to any newly generated code, including extensive testing and validation before production deployment.

This also isn’t NVIDIA’s first step into AI-assisted workflows. The company has long operated dedicated supercomputers and spent years refining deep learning super sampling (DLSS). Some aspects of chip design have also been optimized using internal AI tools.

NVIDIA and Cursor add that even with output tripling, bug rates have remained stable—an outcome they attribute to enterprise-focused optimizations tailored for real-world production environments.