Nokia unveiled an artificial intelligence-based radio access network platform on July 15, with pilot deployments planned by the end of 2026 and commercial availability scheduled for 2027.
The Finnish telecom equipment maker described it as the industry’s first commercial AI-native RAN platform, designed for existing 4G and 5G networks and future 6G systems.
How the AI-RAN platform works
The platform combines Nokia’s anyRAN software with NVIDIA’s Aerial accelerated-computing technology, allowing radio network functions and AI workloads to run on the same infrastructure. Nokia said it could help operators increase capacity and automate network optimization without relying solely on major hardware upgrades.
Nokia aims to double spectral efficiency by 2028
Nokia reported spectral-efficiency gains of more than 20% and said further improvements could reach 50% in 2027 and surpass 100% a year later, allowing operators to carry more traffic over existing spectrum.
“AI-RAN is the biggest innovation in radio in decades,” Nokia Chief Executive Justin Hotard said, adding that it could help operators “get more from their existing infrastructure.”
NVIDIA Chief Executive Jensen Huang said “the radio access network is the next AI infrastructure,” calling the shift “a generational shift for operators.”
Nokia offers operators three paths to AI-RAN
Nokia will offer operators three ways to deploy the platform. The first allows existing AirScale customers to add a GPU-powered plug-in unit to their current equipment.
The second uses a standalone AI-RAN node that can operate independently or alongside AirScale, while the third runs the technology on commercial off-the-shelf servers in a cloud-based environment.
Nokia said the platform will comply with Open RAN standards, while the AirScale upgrade option will also be supported by AI-accelerated merchant silicon from Marvell.
Subscription model reduces reliance on hardware upgrades
The platform will be sold through a software subscription model, giving operators access to new algorithms, network-management features and performance updates without waiting for full hardware replacement cycles.
Nokia said the approach could lower the cost of carrying data and help operators respond to rising uplink demand from generative AI, autonomous systems and other applications that require more processing at the network edge.
Nokia’s comeback is taking shape beyond phones
Nokia largely vanished from the consumer spotlight after selling most of its handset business to Microsoft in 2014, but remained a supplier of telecom networks.
Its push now centers on AI infrastructure, including a $1 billion NVIDIA investment and partnership, AI-RAN development, autonomous-network projects with AWS and Databricks, and agentic network APIs with Google Cloud.
Nokia has also explored blockchain through its Data Marketplace, launched in 2021 for secure data and AI-model exchange, while Bell Labs researches decentralized wireless networks and zero-knowledge proofs.
The strategy positions Nokia less as a revived phone brand than a growing force in AI-era connectivity infrastructure.
