Google, Intel, Qualcomm team bid for Nvidia’s AI crown

[Photo by each company]

U.S. chip giants Intel Corp. and Qualcomm Inc. and big tech company Google LLC are teaming up to develop open-source software for artificial intelligence (AI) apps. The collaboration aims to counter Nvidia Corp.’s dominance in the AI semiconductor market by forming a united front.

Reuters reported on Monday local time that the consortium, called the Unified Acceleration (UXL) Foundation, is currently working on an open-source project to rival Nvidia’s CUDA platform.

UXL Foundation, founded in September 2023, aims to address the issue of monopoly on AI applications by Nvidia and a few other chip manufacturers, particularly targeting CUDA.

CUDA is Nvidia’s software platform that supports the development of AI-related applications, with the company also viewed as one of the world’s top AI companies alongside its role in AI chip manufacturing.

Although CUDA is available for free, it runs solely on Nvidia chips, thereby reinforcing Nvidia’s dominance in the chip market. Nvidia’s accumulation of computer code over the past two decades also gives it a unique competitive advantage. Around 4 million developers around the world use the CUDA platform for AI application development.

UXL Foundation plans to develop a range of software and tools that can run various AI accelerator chips and compete with the CUDA platform. The project aims to ensure that computer code operates on all machines regardless of the type of chip or hardware being used.

In addition to the initial participating companies, the UXL Foundation plans to invite other cloud and semiconductor companies, including Amazon.com and Microsoft Corp., to participate in the project.

The UXL Foundation reportedly began receiving technical contributions from third parties interested in open-source technology at its inception in 2023.

“It’s about specifically – in the context of machine learning frameworks – how do we create an open ecosystem, and promote productivity and choice in hardware,” Google’s director and chief technologist of high performance computing Bill Margo told Reuters in an interview.