Open-RAN compliant, high-performance, energy-efficient 5G vRAN solution leverages the NVIDIA GH200 Grace Hopper™ Superchip, accelerating AI-RAN commercialization for public and private wireless networks
SynaXG, a leading provider of AI-RAN solutions, in collaboration with Kyocera and NVIDIA, unveiled the world’s first data transmission and reception over mmWave radios utilizing the NVIDIA GH200 Grace Hopper™ Superchip. This industry-first milestone marks a transformative leap toward AI-native wireless communications, enabling operators worldwide to unlock new AI-based monetization opportunities and achieve significant gains in wireless network performance and efficiency with AI.
Also Read: Combatting the rise in AI-assisted fraud in 2025
Pioneering AI-Native Wireless Infrastructure
The solution combines advanced radio technology (PHY) with network control units (CU and DU) for a complete 5G mmWave system, based on the NVIDIA AI Aerial™ platform. This full-stack AI-RAN solution enables both AI-and-RAN and AI-for-RAN capabilities, running Layer 1 to Layer 3 functions on the NVIDIA GH200 Grace Hopper™ Superchip.
The integration of NVIDIA GH200 into SynaXG’s AI-RAN solution represents a paradigm shift towards AI-native wireless infrastructure, enabling lightning-fast data transmissions, ultra-reliable performance, real-time intelligent operations, and dynamic resource optimization. These innovations coupled with a fully software-defined, O-RAN compliant vRAN solution, help lay the groundwork for the seamless evolution to 6G networks.
Also Read: The Arbitrage Opportunity of Small Language Models: Unlocking AI Efficiency and Performance
Exceptional Energy Efficiency and Performance
In recent laboratory trials, the solution showcased impressive mmWave performance by supporting five carriers, each with 400 MHz of bandwidth, on a single NVIDIA GH200 superchip. This setup achieved an aggregate peak throughput of up to 30 Gbps per GH200. This performance is comparable to custom ASIC-based solutions while offering significantly greater flexibility and upgradability.
These early trials set the stage for creating one of the world’s most energy-efficient AI-RAN solutions, redefining how networks are optimized for cost and sustainability. In particular, the use of AI algorithms in radio signal processing, such as for frequency selection and power optimization, deliver significant gains in cell throughput, cell edge throughput and overall power consumption, as observed through simulation in lab environments. In addition, the ability of mmWave technology to transfer large amounts of data at minimal energy consumption underscores its transformative potential in wireless networks.
“This achievement lays the foundation for AI to improve and run on super-fast mmWave networks,” said Mantosh Malhotra, Chief Business Officer, SynaXG. “Together with Kyocera and NVIDIA, we have demonstrated how these technologies can converge to help redefine AI-native wireless communications and pave the way for smarter, more efficient networks that meet the demands of both carriers and enterprises globally.”
“Kyocera is proud to partner with SynaXG and NVIDIA on this world-first innovation,” said Masaaki Hori, Deputy General Manager, Kyocera Corporate KWIC (Wireless Innovation Center) Division. “With Kyocera’s strengths in mmWave technology, our collaboration with SynaXG and NVIDIA is a testament to the power of teamwork in driving meaningful advancements in telecommunications technology, starting with AI-RAN.”
“AI-RAN represents the next frontier for wireless networks, taking 5G, 5G Advanced, and O-RAN to new heights and paving the way for 6G as the future of connectivity,” said Soma Velayutham, Vice President of Telecoms, NVIDIA. “This implementation by SynaXG and Kyocera using NVIDIA AI Aerial demonstrates that software-defined AI-RAN can deliver high performance, optimized costs, and improved sustainability for next-generation networks.”
[To share your insights with us, please write to psen@itechseries.com ]
