en.Wedoany.com Reported - Nvidia is promoting a technical route to replace traditional dedicated integrated circuits in 5G base stations with general-purpose GPUs. The AI-RAN Alliance (Artificial Intelligence Radio Access Network Alliance), which it leads, has grown from 11 founding members to 132.

Nvidia believes that the nature of the communication physical layer is highly consistent with AI deep learning operations. By leveraging the CUDA cores and Tensor Cores within GPUs, it can efficiently handle parallel computing while extending the technology from the cloud to low-power distributed sites at the tower level.

In response to Nvidia's move, Nokia has adopted a strategy of full embrace, transferring its RAN chip business originally customized by Marvell to the Nvidia GPU platform and developing its Layer 1 RAN operations within the CUDA software environment. Ericsson, on the other hand, insists on a self-developed ASIC path and has allied with Intel to support the CPU solution. Last year, Ericsson invested $5.2 billion in RAN R&D. The company believes that, against the backdrop of declining operator RAN spending, ASICs offer irreplaceable advantages in cost and thermal control.

Industry insiders believe that the software-defined networking revolution initiated by Nvidia is forcing traditional telecom equipment manufacturers to reassess their development paths. Once GPUs become the default computing foundation for wireless networks, traditional equipment makers may transition from system integrators to hardware assemblers.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com