Noticia
Nokia launches commercial AI-RAN platform with a software path to 6G
Nokia has announced a commercial AI-native radio access network platform designed to increase mobile capacity through software, while pilots are planned before a 2027 launch.
Nokia has announced what it describes as the mobile industry’s first commercial AI-RAN platform, bringing artificial intelligence and accelerated computing into the radio access network. The company says the platform is designed to help operators extract more capacity from the spectrum and infrastructure they already own, while creating a software-based route toward future 6G networks.
The announcement was made on July 15, 2026. Nokia’s platform combines its anyRAN software with NVIDIA’s Aerial AI-RAN platform. It is intended for 4G, 5G and future 6G deployments, and Nokia says it will support open, interoperable network architectures through compliance with Open RAN standards.
Why AI is moving closer to the mobile network
Radio access networks connect smartphones and other devices to an operator’s core network. Traditionally, their performance has depended heavily on dedicated hardware, radio upgrades and carefully planned spectrum use. Nokia argues that the growth of generative, agentic and physical AI is changing that balance. AI services can create heavier uplink demand, require lower latency and generate traffic patterns that are less predictable than conventional voice, video and web use.
In response, Nokia is positioning AI-RAN as a programmable platform rather than a fixed collection of network appliances. The company says radio functions and AI workloads can run on a shared accelerated-computing foundation. That approach is intended to let operators improve network performance through software updates instead of waiting for a complete hardware refresh.
Three deployment paths for operators
Nokia says its AI-RAN platform will be available through three hardware options. Existing AirScale customers will be able to add a GPU-powered capacity plug-in unit to current baseband deployments. This option is designed to preserve the existing site footprint while adding AI-accelerated processing.
The second option is a standalone GPU-powered AI-RAN node. Nokia says it can operate independently, in a cluster or alongside AirScale equipment as part of a single logical base station. The company presents this as a way to add capacity while supporting the evolution from 4G and 5G toward 6G.
The third route is a cloud-native deployment using commercial off-the-shelf servers. Nokia says this option is aimed at operators that want more flexible, distributed infrastructure and an open supply chain. All three paths are based on the same anyRAN software foundation and share a roadmap toward 6G.
Efficiency claims still need to be proven in live networks
Nokia says its AI-driven radio innovations have already demonstrated more than 20 percent spectral-efficiency gains. The company is targeting 50 percent gains by 2027 and more than 100 percent by 2028, which it says could double the capacity available from existing spectrum assets.
Those figures are company projections and should not be read as results that every operator or smartphone user will immediately experience. Actual performance will depend on network configuration, spectrum conditions, traffic density, radio equipment, software deployment and the workload being processed. Nokia’s announcement does not provide an independent test protocol or third-party measurements for the broader platform.
For mobile users, the possible benefit is therefore indirect at first. If operators deploy the technology successfully, more capacity could help dense cells handle busy-hour traffic and demanding uplink applications. It could also give networks more room for cloud gaming, video sharing, connected devices and AI services. The announcement does not promise a specific speed increase for consumer smartphones.
Software subscriptions and the 2027 roadmap
Nokia is also introducing a subscription-based commercial model for AI-RAN. Under this approach, operators would receive ongoing access to new AI algorithms, spectral-efficiency improvements, optimization features and other software capabilities as they become available. Nokia says this model is intended to reduce dependence on traditional hardware upgrade cycles.
The timing remains important. Nokia expects pilot deployments to begin at the end of 2026, while commercial availability is planned for 2027. That means the platform is an announced product roadmap rather than a service that mobile customers can use today. Operators will still need to evaluate integration, power consumption, security, standards compliance and the business case for adding accelerated computing to live networks.
What the announcement means for the smartphone market
AI-RAN is infrastructure technology, but it could influence the smartphone experience as operators adapt to AI-heavy traffic. Smartphones increasingly rely on cloud assistants, real-time translation, image processing and other services that exchange data continuously with remote servers. A radio network that can respond more efficiently to changing demand could support those services at scale.
That does not make AI-RAN a phone feature. Users will not install Nokia’s platform from an app store, and its availability will depend on decisions made by individual operators. The immediate significance is that Nokia is proposing a way to modernize mobile networks through software and accelerated computing while keeping existing deployment choices open.
Nokia’s announcement moves AI-RAN from a research and standards discussion toward a defined commercial roadmap. The next meaningful checkpoint will be the planned pilot deployments at the end of 2026. Until those trials produce independently verifiable results, the platform’s efficiency targets should be treated as Nokia’s stated expectations rather than established industry performance.
Source: Nokia’s official announcement and AI-RAN product information.