LITEON and NVIDIA Team Up on AI-RAN Tech
At MWC Barcelona, LITEON Technology showcased its latest AI-RAN advancements, which integrate its Open Radio Units with NVIDIA's AI Aerial platform. The collaboration is designed to accelerate the commercialization of GPU-accelerated, AI-native radio access network architectures.
This partnership represents a critical step in the evolution of mobile networks, moving beyond traditional, hardware-specific setups to a more flexible, software-defined model. AI-RAN technology allows network operators to run both cellular network functions and artificial intelligence applications on the same general-purpose hardware, increasing infrastructure efficiency. LITEON's contribution centers on its O-RAN compliant Radio Units (O-RUs) for both sub-6 GHz and millimeter wave 5G spectrums. These units are engineered to provide the stable, carrier-grade performance necessary for reliable AI-RAN operations, ensuring predictable fronthaul behavior for processing radio signals. NVIDIA's AI Aerial is the software and accelerated computing platform that serves as the system's brain. It leverages the company's powerful GPUs, including the Grace Hopper Superchip, to handle the intense computational demands of both 5G signal processing and concurrent AI workloads, from machine learning to generative AI. By pre-integrating LITEON's hardware with the NVIDIA platform, the collaboration aims to lower deployment barriers for telecom operators. This validated compatibility is designed to reduce integration complexity and accelerate the time-to-market for commercial AI-RAN networks, a key step in moving the technology from trials to widespread adoption. This shift enables new, low-latency AI applications to run directly at the network edge. Use cases include real-time video analytics for security, intelligent traffic management for smart cities, and factory automation with robotics, creating new monetization opportunities for network providers beyond simple connectivity. The development of AI-RAN is also seen as a foundational building block for future 6G networks. The AI-native architecture being developed by LITEON, NVIDIA, and other ecosystem partners like Supermicro and SynaXG is designed to be inherently more adaptive and scalable to meet future data-intensive demands.
