Nvidia's Jensen Huang: Without GeForce, there would be no AI

Tech News 3/13: The technology media Wccftech published a blog post yesterday (March 12), reporting that while celebrating the 25th anniversary of the GeForce 3 graphics card release with the GeForce team, NVIDIA CEO Jensen Huang stated that the gaming business has initiated a modern AI revolution, and directly said, “Without GeForce, there is no CUDA; without CUDA, there is no AI.”

Huang recalled that early graphics cards like Riva 128 and TNT used fixed-function accelerators, which lacked flexibility, making all game visuals in the late 1990s look monotonous.

To break this deadlock, NVIDIA introduced programmable vertex and pixel shader architectures on the GeForce 3. This technology granted developers significant control, allowing them to inject artistic expression into every game. IT Home has included the evolution of related game graphics as follows:

Source of the above images: video screenshots

To achieve the leap from fixed hardware pipelines to programmable pipelines, NVIDIA also officially transformed into a computing company. Huang said that this shift to a new programmable architecture ultimately paved the way for the birth of CUDA, empowering GPUs with powerful parallel computing capabilities.

Huang believes that NVIDIA’s bold bets on ray tracing technology are equally significant. Due to the extremely high computational cost of ray tracing, relying solely on raw power is no longer sufficient. NVIDIA launched the RTX architecture and DLSS, among other image super-resolution technologies. DLSS uses neural rendering technology to successfully introduce generative capabilities into the field of computer graphics.

The continuous progress in computer graphics has made NVIDIA deeply aware that the potential of GPUs extends far beyond rendering workloads, which directly led to the development of modern AI computing today.