Nokia Launches AI-RAN Platform with NVIDIA Support Amidst Market Challenges
### Nokia’s AI-RAN Platform: A Strong Reentry Fuelled by NVIDIA
Nokia officially launched its AI-RAN platform on July 15, aiming to revolutionize radio architecture with claims that it is a world first in the industry. Built on its anyRAN software and NVIDIA's Aerial system, this platform promises to help operators extract significantly greater capacity from their existing spectrum. The launch is framed as a monumental shift in radio technology that could reshape the sector for decades to come.
The technical advantages are compelling. According to Nokia, the platform has already achieved over 20% in spectral efficiency gains, with ambitious targets set at 50% by 2027 and over 100% by 2028. These figures suggest that operators could potentially double the capacity of their spectrum. However, Nokia's timeline indicates that pilot programs will begin by the end of this year, with full commercial availability not expected until 2027.
Operators will have the flexibility to access this capability through a software subscription model, eschewing traditional hardware upgrades. They will be able to choose from three deployment methods: a GPU-powered plug-in card compatible with existing AirScale sites, a standalone AI-RAN node, and a cloud-server option through partnerships.
This launch is significant not only as a technical advancement but also for Nokia's position in the market. The radio segment has presented major challenges for CEO Justin Hotard since he took over in 2025. During an investor meeting in November, he acknowledged that the mobile division had not been providing satisfactory returns and merged it into a new Mobile Infrastructure segment alongside plans for further cost reductions.
Nokia's collaboration with NVIDIA, which was announced in October 2025 with a $1 billion investment for approximately a 3% stake, plays a crucial role in revitalising its radio business. By leveraging NVIDIA's advanced silicon and CUDA software instead of relying on custom chips, Nokia aims to decrease expensive in-house research and development, transitioning from a legacy hardware-centric approach.
Investors have reacted positively to these developments, resulting in a notable increase in Nokia's stock value throughout 2026, especially as the AI-RAN announcement came just days before its second-quarter results. Analyst Rémy Pascal from Omdia has stated that the cumulative market opportunity for AI-RAN could exceed $200 billion by 2030, underscoring the significance of this technology.
However, questions arise regarding Nokia's claim of being the "first" in the industry to launch such a platform. In June, Ericsson began offering a commercial AI-in-RAN software subscription that reportedly provides up to a 20% increase in downlink throughput and 10% better spectral efficiency across over 15 live deployments, operating on existing baseband silicon without the need for GPUs.
Nokia's assertion of being first hinges on a specific definition: it focuses on a GPU-accelerated AI-RAN platform rather than AI features layered onto existing infrastructure. While both claims can coexist, careful scrutiny of the context is essential.
Furthermore, Nokia has committed its radio roadmap to NVIDIA, with Chief Technology Officer Pallavi Mahajan admitting that certain aspects of the Layer 1 software rely on the underlying hardware. Conversely, Ericsson has strategically opted to maintain silicon independence for its AI features to sidestep hardware reliance.
Nokia highlights its use of merchant silicon from Marvell and positions its platform as Open RAN-compliant. However, the performance metrics it promises, particularly spectral efficiency gains, currently depend on NVIDIA's technology, for which there are no direct competitors available.
This strategy is not inherently flawed; relying on the industry leader in AI chip production allows Nokia to tackle challenges in its radio division. Moreover, the subscription model could generate recurring revenue unlike traditional hardware sales. Nevertheless, the platform is still not commercially available, and its high-efficiency targets lie two years in the future, while at least one prominent competitor has captured the market first through a different strategy. For Nokia, this marks a work-in-progress comeback rather than a completed victory, with its journey tightly interwoven with NVIDIA's advancements.
