SoftBank Corp.'s “AITRAS” Deployed
at NVIDIA Headquarters to Enable Development
of New Edge AI Applications

SoftBank Corp. (“SoftBank”) announced its “AITRAS”^* converged AI-RAN solution has been deployed at NVIDIA headquarters in Santa Clara, California, USA. The live implementation will facilitate innovation and advanced AI-and-RAN and AI-for-RAN experiments, as well as enable the development of new AI-on-RAN applications.

The SoftBank and NVIDIA collaboration aims to utilize NVIDIA accelerated computing platforms, including, but not limited to, the NVIDIA GH200 Grace Hopper Superchip in building next-generation data centers for AI-powered wireless 5G/6G networks with the goal of introducing GPU-based solutions in the telecom sector. In November 2024, SoftBank announced its AITRAS converged AI-RAN solution that leverages NVIDIA AI Aerial-based accelerated computing platforms and fully software-defined vRAN (virtual Radio Access Network). The first tangible outcome of the strategic collaboration between the two companies was realized by an outdoor field trial.

The deployment of AITRAS at NVIDIA's headquarters marks the first implementation outside Japan, representing a significant step forward in demonstrating the potential of SoftBank's technology to the global market. Moving forward, both companies will continue collaborating to advance mobile edge AI applications and work with innovative third-parties and NVIDIA ecosystem partners to drive broader adoption.

As an initial step toward building an edge AI applications ecosystem, SoftBank, in collaboration with NVIDIA developed a “porter robot” utilizing an LLM and began a demonstration experiment at NVIDIA's headquarters. Using an LLM running on AITRAS, the robot carries luggage with flexibility and follows visitors based on verbal instructions about their clothing. In a low-latency environment with a 20ms round-trip for uplink and downlink communication, it operates stably and demonstrates higher followability compared to conventional robot control.

Both companies aim to realize an unprecedented, innovative AI service with low latency and high security on the mobile edge of AITRAS, while also building their own ecosystem. SoftBank and NVIDIA will accelerate joint development efforts, including the development and validation of mobile edge AI services, and the optimization of GH resources through collaboration with the AITRAS orchestrator.

Ryuji Wakikawa, Vice President, Head of the Research Institute of Advanced Technology at SoftBank, said: “SoftBank and NVIDIA are key partners in advancing AI-RAN and driving innovation in the telecommunications industry. Through the deployment of AITRAS at NVIDIA headquarters, this partnership will be further strengthened, accelerating our joint efforts toward realizing commercial AI-RAN development and innovation.”

Soma Velayutham, Vice President of Telecoms at NVIDIA, said: “AI-RAN is paving the way for next-generation AI-Native wireless networks. SoftBank's AITRAS, leveraging NVIDIA AI Aerial, will now become accessible to a wider ecosystem of partners through its implementation at NVIDIA headquarters, and will accelerate innovation across all dimensions of AI-RAN, including new AI applications on RAN.”

[Notes]

1. *
   For more details on AITRAS, please refer to the press release dated November 13, 2024: “SoftBank Corp. Announces Development of “AITRAS,” a Converged AI-RAN Solution”

• 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.