Efforts to challenge Nvidia's AI dominance by focusing on software

Nvidia has built a critical role in the generative AI industry, supplying chips to everyone from small firms to tech giants like Microsoft and Google, leading to its market value of $2.2 trillion.

The company's computer code from the past 20 years is as important as its hardware, making it very hard for others to compete with them. Nvidia's CUDA software platform is used by over 4 million developers worldwide to create AI and other applications.

A group of tech giants, including Qualcomm, Google, and Intel, plans to challenge Nvidia by targeting the software that ties developers to using Nvidia chips. They are among the growing number of investors and businesses that are undermining Nvidia's hegemony in artificial intelligence.

Executives involved say the UXL Foundation, formed by tech companies, aims to create software that can support different AI accelerator chips, beginning with Intel's OneAPI technology. Open-source software allows computer code to run on any machine, no matter the hardware or chip.

Bill Magro, a top technologist for Google, said in an interview that their goal is to create an open and productive environment that offers choice in hardware for machine learning. According to Magro, one of the original members of UXL, Google contributes to deciding the project's technical course.

In order to finalize technical requirements in the first half of this year, UXL's technical steering committee is getting ready. Executives stated that engineers intend to bring the technological elements to a "mature" condition by year's end. The CEOs highlighted the need to build a strong foundation supported by various companies, compatible with all types of chips and hardware.