Bill Dally (Bill Dally), CTO of NVIDIA, in a recent conference in talking about GPGPU hinted that solutions AMD probably would support an application to CUDA. Dally said: "In the future you will be able to run applications written in C with CUDA extension to a broader range of platforms, so I do not think it will be a major limitation." Although not named specifically Dally AMD, he said: "I know some of the projects, which aim to use CUDA on other platforms."

Dally said that the programming language Brook, which is used in AMD Stream, has a "significant number of limitations." He believes that CUDA will provide "hundreds of applications," while "at AMD has a very small number of items that are suitable for GPU-computing."

CTO NVIDIA is not worried that users can begin to willingly buy a graphics card for graphics chips instead of AMD chips to the outcome of NVIDIA, if AMD chips will support CUDA. "We do not care whether GPGPU applications are limited only by our GPU, or to be able to work on a broader spectrum of platforms," ─ Dally said. ─ "We produce the best GPU, which has ever been, and, subject to a fair competitive environment, people will choose our solutions."