Former Google CEO predicts NVIDIA will win the AI race. Eric Schmidt, CEO of the tech giant in its early years, noted that high demand for GPUs to train language models will be crucial to securing victory.
In a speech to Stanford students, Eric Schmidt shared his perspective on the future of artificial intelligence. The former Google boss said that NVIDIA has an advantage thanks to its hardware, because there is no single competitor that can compete with the GPUs it has produced so far. All companies that develop AI models trust NVIDIA and plan to invest heavily in its technologies.
Eric Schmidt said that it will be difficult to catch up with NVIDIA because it is not only hardware dependent but also software dependent. Most of the open source tools used in AI development are based on CUDA. Although Intel, AMD and other companies are working on an alternative to CUDA, NVIDIA will have a significant advantage.
WHAT IS CUDA?
CUDA is an architecture and technology for GPUs that is available as an add-on to the C programming language from NVIDIA. It is a PathScale-based C compiler and a set of development tools that enable algorithms written in C to run on the GPU.
