NVIDIA’s AI chip advantage is becoming the target of Google, Intel and Arm

Major tech companies are trying to eliminate the software advantage that has helped Nvidia dominate the artificial intelligence market. According to Reuters, a group including Intel, Google, Arm, Qualcomm, Samsung and other technology companies are developing a set of open source software to prevent artificial intelligence developers from being locked into Nvidia's proprietary technology and allow their code to run on any machine and any chip.

The group, called the Unified Acceleration Foundation (UXL), said the project's technical details should reach "maturity" in the second half of this year, but did not give a final release target. The project currently includes the OneAPI open standard developed by Intel, which aims to eliminate requirements for specific coding languages, code libraries and other tools, so that developers are no longer tied to using a specific architecture such as NVIDIA's CUDA platform.

Last month, NVIDIA became the first chipmaker to reach a $2 trillion market capitalization, growing rapidly after focusing on hardware that powers artificial intelligence models like the H100 and upcoming H200 GPUs. These NVIDIA chips lock developers into using NVIDIA's CUDA architecture and are superior to anything currently produced by other chipmakers, but explosive demand has created scarcity while competitors continue to develop their own alternatives. In the 2023 Computex keynote speech, NVIDIA CEO Jensen Huang said that 4 million developers are using the CUDA computing model.

UXL said the project's initial goal is to open options for artificial intelligence applications and high-performance computing applications, but the group plans to eventually support NVIDIA's hardware and code as well. UXL is seeking assistance from more chip manufacturers and cloud computing companies like Microsoft and Amazon to ensure the solution can be deployed on any chip or hardware. Microsoft is rumored to have partnered with AMD last year to develop alternative artificial intelligence chips to challenge NVIDIA's effective monopoly in the industry.