Nokia launches AI-native RAN platform built on NVIDIA silicon
Nokia has announced what the company positions as the first commercial AI-native radio access network platform, developed in partnership with NVIDIA and targeting a substantial improvement in the spectral efficiency of existing mobile infrastructure. The platform, which combines Nokia's anyRAN software with NVIDIA's Aerial AI-RAN compute stack, is scheduled to enter pilot deployments by the end of 2026 and reach commercial availability in 2027.
The announcement marks a significant architectural shift for Nokia's RAN portfolio. Rather than relying on dedicated hardware refresh cycles, the platform introduces a software subscription model through which operators can receive new AI algorithms, spectral efficiency enhancements and network optimisation capabilities as ongoing updates. Nokia says the platform has already demonstrated more than 20% spectral efficiency gains in testing, and is projecting 50% gains by 2027 and in excess of 100% by 2028.
Three adoption paths
Nokia has structured the platform around three hardware deployment routes to accommodate the diversity of operator installed bases. Existing AirScale customers can adopt a new GPU-powered plug-in capacity card, preserving prior infrastructure investment. Operators seeking maximum flexibility can deploy a new standalone GPU-powered AI-RAN node, which supports 4G, 5G and future 6G workloads on a common platform. A third path targets operators pursuing cloud-native architectures, offering COTS server solutions delivered through ecosystem partners. All three share a common software-defined architecture and are designed to be compliant with Open RAN interoperability standards.
Chief executive Justin Hotard described the launch as the biggest radio innovation in decades, arguing it extends AI into the physical world and provides a software-upgrade path toward 6G. NVIDIA founder and CEO Jensen Huang added that bringing CUDA and AI into the baseband effectively turns the RAN into a "planet-scale AI computer," a framing that underscores how far the partnership reaches beyond conventional radio hardware.
Market and competitive context
The AI-RAN category has attracted significant attention from a number of infrastructure vendors, hyperscalers and semiconductor companies, all seeking a position in what is expected to be a large wave of operator capital expenditure as 5G densification continues and 6G planning accelerates. Ericsson, Samsung and a range of Open RAN specialists are all pursuing AI-accelerated baseband approaches, though Nokia's claim of being first to commercial availability with a GPU-native architecture will be tested as competing roadmaps mature.
NVIDIA's involvement is notable. The company's accelerated compute platforms have reshaped server and data-centre economics; applying that model to the mobile baseband, where per-bit cost reduction is an enduring operator priority, could have significant implications for incumbents whose revenues depend on hardware cycle refresh rates. Marvell is also cited as a contributor of AI-accelerated merchant silicon, indicating Nokia is managing its supply chain across multiple chip partners rather than relying solely on NVIDIA.
Industry analyst firm Omdia's Rémy Pascal noted in Nokia's release that the combination of AI-accelerated compute, a software-defined architecture and a clear product roadmap represents a credible step from industry vision to commercial reality, though he stopped short of endorsing the specific efficiency projections.
Regulatory and standards considerations
Open RAN compliance, which the platform claims across its full portfolio, has been a growing requirement in several major markets. Both the UK and the US have issued policy frameworks encouraging multi-vendor interoperability in mobile networks to reduce strategic dependence on single suppliers. Nokia's decision to position the entire AI-RAN line as ORAN-compliant addresses these policy pressures directly and may ease procurement in markets where open-interface mandates are tightening.
The 2028 timeline for the 100%-plus spectral efficiency target extends well into the phase where EU and UK 6G spectrum policy frameworks are expected to be under active consultation. Operators deploying the platform will be watching closely for independent validation of Nokia's efficiency claims before committing to large-scale rollouts.