LITEON and NVIDIA Push AI-RAN at MWC
At MWC 2026, LITEON Technology is showcasing its integration of Open Radio Units with NVIDIA's AI Aerial platform. The demonstration highlights how GPU-accelerated architectures are being used to create scalable, AI-native Radio Access Networks. This reflects a broader industry trend of pushing AI and advanced computation closer to the network edge.
The move towards AI-RAN is rooted in the principles of Open RAN (O-RAN), an industry-wide shift to disaggregate the traditionally monolithic Radio Access Network. This approach uses standardized, open interfaces between components like the Radio Unit (RU), Distributed Unit (DU), and Centralized Unit (CU), enabling operators to mix and match hardware from various vendors and avoid single-vendor lock-in. LITEON's contribution is its O-RAN 7.2x compliant Open Radio Unit (O-RU), a critical hardware component that provides carrier-grade stability for the radio signals. By adhering to open fronthaul standards, LITEON's RUs can be more easily integrated with the virtualized, software-defined components from other ecosystem partners, a key requirement for commercial deployment. NVIDIA's Aerial platform serves as the software-defined core of this architecture, running on GPU-accelerated commercial-off-the-shelf (COTS) servers instead of specialized hardware. The platform includes CUDA-accelerated libraries for 5G signal processing and allows for the dynamic allocation of both 5G and AI workloads on the same GPU, which can boost resource utilization by 2-3x. This collaboration is part of a broader industry ecosystem that includes partners like Supermicro and the AI-RAN software provider SynaXG. At MWC, SynaXG demonstrated a fully software-defined AI-RAN on a single NVIDIA GH200 server, achieving 36 Gbps throughput with under 10 milliseconds of latency. The push for AI-RAN extends beyond this partnership, with the AI-RAN Alliance growing to 132 members, signaling wide industry adoption. Competitors like Nokia are also deeply involved, announcing their own customer integrations with NVIDIA's platform alongside operators such as T-Mobile, BT Group, and Vodafone. This convergence of GPU acceleration and open, virtualized architectures is considered a foundational step in the evolution from 5G towards AI-native 6G networks. Market projections reflect this shift, with the AI-RAN market expected to account for approximately one-third of the total RAN market by 2029.
