Key Highlights
- NVIDIA says its AI Aerial platform improves radio access network (RAN) performance using AI.
- The platform is designed to increase spectral efficiency and boost network throughput.
- The technology supports telecom operators as they prepare for AI-native 5G Advanced and future 6G networks.
NVIDIA is expanding its AI capabilities into wireless communications with AI Aerial, a platform designed to improve the performance of radio access networks (RAN) by combining GPU acceleration with artificial intelligence.
The company said AI Aerial helps address long-standing challenges associated with Massive MIMO technology, allowing operators to use wireless spectrum more efficiently while improving network capacity and reliability.
Massive MIMO has been viewed as one of the key technologies behind modern mobile networks. However, many deployments have struggled to deliver their expected performance because of limitations such as inefficient user pairing, signal tracking, and processing bottlenecks.
AI Platform Targets Higher Network Efficiency
According to NVIDIA, AI Aerial replaces traditional CPU-heavy processing with an AI-native architecture built around GPUs. This enables advanced Layer 1 and Layer 2 network algorithms to process large volumes of data more efficiently and respond faster to changing network conditions.
One of the platform’s key capabilities is AI-powered beamforming, which directs wireless signals more accurately toward connected devices.
NVIDIA said internal testing showed beamforming delivered up to a 1.62x improvement in spectral efficiency in a 64T64R Massive MIMO system using 32 transmission layers compared with conventional techniques.
The company also highlighted deep reinforcement learning (DRL) for link adaptation, which continuously adjusts modulation and coding schemes based on real-time network feedback. According to NVIDIA, the technology improved throughput by around 1.3 times for users located at the edge of a network cell, where signal quality is often poorer.
Growing Demand for AI-native Telecom Networks
Efficient spectrum usage has become increasingly important as telecom operators continue investing in 5G infrastructure while preparing for future 6G deployments.
Wireless spectrum remains one of the industry’s most valuable assets, and operators are looking for technologies that can improve network performance without requiring additional spectrum purchases.
NVIDIA said recent trials conducted with SoftBank demonstrated up to a threefold improvement in spectral efficiency during outdoor testing using AI-RAN technologies.
The broader telecom industry is also moving toward AI-driven network management. Companies are increasingly using artificial intelligence to optimize scheduling, traffic management, and network resource allocation across large-scale 5G deployments.
NVIDIA said its collaboration with Nokia reflects the industry’s growing focus on AI-native network infrastructure for next-generation wireless communications.
As commercial AI-RAN deployments move closer, the company believes platforms such as AI Aerial could help operators improve throughput, make better use of existing infrastructure, and support new edge AI services while preparing networks for the demands of future 6G technologies.
