SoftBank Corp. (“SoftBank”) has agreed with NVIDIA to open a research facility called “AI-on-5G Lab.”, where various solutions including AI technology can be demonstrated and applied to the business domain in an environment where virtualized radio access networks (vRAN) and multi-access edge computing (MEC) are integrated. The facility, which will feature fully virtualized wireless networks for 5G, will be located in SoftBank's headquarters in 2022. Through various demonstrations at the “AI-on-5G Lab.”, SoftBank will develop end-to-end solutions for private 5G to accelerate the commercialization of fully virtualized private 5G.

AI technologies are now applied to more areas than before, including for the detection of objects using image-recognition technologies, voice recognition/translation, and network optimization/automation operations. At the same time, the virtualization of network equipment that supports mobile communications, such as vRAN, using general-purpose computers is expected to become widespread. Both AI technology and network equipment virtualization are software that NVIDIA graphics processing units (GPUs) can process efficiently.

Since 2018, SoftBank has highly evaluated NVIDIA's GPU-based vRAN and MEC edge computing processing, and has been collaborating with NVIDIA on projects such as “Commercialization of GeForce NOW by GPU video rendering in Japan” (2019), “Evaluation of vRAN accelerator performance” (2020), and “Demonstration of super-resolution by AI-engine “Maxine”” (2021). Based on these experiments, SoftBank concluded that the merits of the widespread adoption of vRAN not only includes cost reductions by configuring communication devices with software on a general-purpose server, but also the flexibility to simultaneously develop a variety of applications and services besides communication. For example, in factories where private 5G is installed, it will be possible to improve productivity by using the information accumulated in the MEC as a computing resource for AI learning during the nighttime when little communication is taking place, thus enabling new value creation.

The “AI-on-5G Lab.” will enable a variety of experiments that integrate both vRAN and MEC functions with GPUs. The “AI-on-5G Lab.” consists of hardware and middleware for base station virtualization and AI processing both provided by NVIDIA, virtualized wireless signal processing software and core network software provided by Mavenir, a network software provider, and physical antennas provided by Foxconn Technology Group on top of private 5G networks provided by SoftBank. This is a fully virtualized platform that combines the software required for private 5G communications with AI MEC applications.

Image of “AI-on-5G Lab.”

“Network of AI” is one the key technologies that SoftBank has released as “12 Challenges for 6G” this July, and will use the “AI-on-5G Lab.” as a place to develop and verify “AI networks”. The “Network of AI” is a concept where AI workloads are placed everywhere in the network so that AI resources can be utilized simultaneously while using the network. This will allow us to freely configure the network to meet the needs of the application and to automate the optimization of the network using AI technology.

• SoftBank, the SoftBank name and logo are registered trademarks or trademarks of SoftBank Group Corp. in Japan and other countries.
• Other company, product and service names in this press release are registered trademarks or trademarks of the respective companies.