Chipmaking giant Nvidia recently received FCC approval to build a private, standalone (SA) 5G wireless network at its Santa Clara, California, headquarters in order to test its products with unnamed O-RAN vendors.
The company’s network will run in 3.7GHz-3.8GHz spectrum, which sits in the C-band. Verizon, AT&T and T-Mobile are all using midband C-band spectrum in their 5G networks.
“We would like to apply for a license (3.7GHz to 3.8GHz) on our Nvidia campus (Santa Clara, CA),” the company wrote in its application to the FCC. “We are expanding our 5G Stand Alone (SA) End-to-End system by adding over-the-air (OTA) capability. This indoor, Private 5G system will be used to test the interoperability of NVIDIAs 5G solutions with O-RAN compliant vendors (for example radio unit). The system will include a 5G Core Network, 5G gNB, 5G Radio Unit (O-RU) and 5G SA capable devices (for example smart phone).”
The company’s application is signed by Joseph Boccuzzi. On his LinkedIn profile, Boccuzzi is listed as Nvidia’s principal for 5G systems.
An Nvidia representative couldn’t immediately comment on questions from Light Reading about the company’s new tests.
The context
Nvidia’s efforts appear to touch on a variety of hot-topic issues in the 5G industry. First, the company is building a private wireless network to test its products – private wireless networking remains a nascent but potentially lucrative trend across the industry.
Nvidia is also building a standalone 5G network. Standalone (SA) 5G technology has proven difficult for big, established wireless network operators to implement, but it promises to support a variety of high-end services.
Finally, Nvidia’s interest in O-RAN technology is noteworthy considering the O-RAN Alliance continues to promote open RAN specifications designed to create more interoperability in wireless networks. Already AT&T has promised to embrace open RAN in a way that the operator hopes will open the doors to more vendors and more innovation.
An emerging player
Nvidia’s GPUs have taken the AI market by storm. That’s because they are viewed as essential to running most advanced AI services. Specifically, they’re being deployed in large data centers to help train large language models (LLMs) like OpenAI’s ChatGPT.
However, Nvidia is also considering new silicon options for the radio access network (RAN). Nvidia reckons its Grace Hopper chip can multitask in a way that will support new AI applications at the edge as well as in RAN software.
To stimulate interest in and further development of its AI-in-the-RAN story, Nvidia recently put its marketing muscle behind a new initiative dubbed the AI-RAN Alliance. Members include RAN vendors Ericsson, Nokia and Samsung, as well as UK-based chip designer ARM and operators like T-Mobile and Softbank. Softbank is already investing in Nvidia’s GPUs for its Japanese network.
Nvidia’s Boccuzzi recently posted a video onto social media touting some new demonstrations from the AI-RAN Alliance with technologies from Softbank, Juniper Networks and Fujitsu.
Interestingly, Fujitsu – an up-and-coming RAN vendor – recently showed off the results of its AI work with Nvidia. Fujitsu essentially supercharged its 5G radios by pairing them with Nvidia’s Grace Hopper superchip.
This post was originally published on the 3rd party mentioned in the title ofthis site
