SynaXG Achieves Carrier-Grade Performance with AI-RAN
SYDNEY — SynaXG has demonstrated significant advancements in AI-native radio access networks (AI-RAN) on the NVIDIA AI-RAN platform, the company announced this week. Conducted in Barcelona and Singapore, the tests showcased the concurrent operation of 5G FR1, 5G FR2, and AI workloads with real-time, policy-driven GPU orchestration, according to SynaXG.
The company reported achieving carrier-grade 5G FR1 performance on a single NVIDIA GH200 platform, operating 20 x 100MHz 5G NR cells and delivering over 36Gbps of aggregated throughput with sub-10 millisecond latency. These results indicate that a fully software-defined AI-RAN can match the performance of top-tier commercial 5G deployments while supporting concurrent AI workloads, the company confirmed.
Technical Breakthroughs and Future Plans
In addition to FR1, SynaXG revealed a world-first carrier-grade FR2 virtualized RAN implementation running alongside FR1 and AI workloads on a shared GPU platform. This achievement delivered high throughput with end-to-end latency as low as 5 milliseconds, meeting the demands of dense urban and enterprise deployments. SynaXG’s orchestration software dynamically allocates GPU resources based on live RAN and AI performance indicators, ensuring uninterrupted service and optimal infrastructure utilization.
According to Xin Huang, CEO of SynaXG, the AI-RAN system operated continuously under sustained load, confirming its stability and readiness for commercial environments. Huang stated, “Our software-defined approach enables operators and enterprises to deploy scalable, AI-native networks ready for commercial deployment.” Soma Velayutham, VP of AI and Telecoms at NVIDIA, emphasized the importance of software-defined architecture for next-generation networks, highlighting SynaXG’s achievements as critical for maintaining carrier-grade performance and flexibility.
The announcement comes as the industry seeks more efficient and flexible network solutions. SynaXG plans to extend its AI-RAN platform to other NVIDIA infrastructures, demonstrating the portability and scalability of its architecture. The company also aims to enhance AI-for-RAN capabilities, focusing on network optimization, spectral efficiency, and automation.
Source: newshub.medianet.com.au
