Nokia is making a bold play in the competitive telecommunications infrastructure market with the launch of its AI-RAN platform. Unveiled on July 15th, the Finnish giant is touting it as the industry’s first “AI-native” radio access network (RAN) solution. This ambitious claim, built upon Nokia’s existing anyRAN software and NVIDIA’s powerful Aerial system, signals a significant architectural shift, aiming to unlock substantial capacity gains from existing spectrum holdings for mobile operators.
The core technical proposition is compelling: Nokia claims its platform has already demonstrated spectral efficiency improvements exceeding 20%. The company has set aggressive targets, aiming for a 50% leap by 2027 and a remarkable over 100% increase by 2028. If realized, this would effectively allow operators to double the capacity of their current spectrum assets without additional infrastructure investment. While these ambitious figures represent future projections, Nokia’s immediate roadmap includes pilot deployments by the end of this year, with commercial availability slated for 2027.
Nokia is structuring the offering as a software subscription, a departure from traditional hardware upgrade cycles. Operators will have flexibility in deployment, choosing between a GPU-enhanced plug-in card for existing AirScale sites, a dedicated AI-RAN node, or a cloud-based solution delivered through strategic partners. This move towards a software-centric model is designed to provide a more agile and recurring revenue stream for Nokia’s radio business.
This AI-RAN platform represents a critical pivot for Nokia, particularly in its radio infrastructure segment, which has been a persistent challenge for CEO Pekka Lundmark. At a November capital markets day, Lundmark acknowledged that the mobile business had not met acceptable return thresholds, prompting a strategic realignment and cost-cutting measures. The collaboration with NVIDIA, formalized in October 2025 with a substantial $1 billion investment from the chipmaker for approximately a 3% stake, is central to this turnaround strategy. By leveraging NVIDIA’s cutting-edge silicon and CUDA software, Nokia is strategically offloading a significant portion of its costly in-house R&D, redirecting resources towards software development—a fundamental shift away from its legacy hardware-centric approach.
This strategic pivot appears to be resonating with investors. Nokia’s stock has experienced a notable re-rating throughout 2026, buoyed by momentum in its AI and cloud initiatives. The AI-RAN launch conveniently preceded its second-quarter financial results, further amplifying market optimism. Analysts at Omdia, whose insights are cited in Nokia’s announcement, project the cumulative AI-RAN opportunity to surpass $200 billion by 2030, underscoring the significant market potential. The industry’s trajectory is undeniably moving towards AI-enhanced networks, and the critical question for Nokia is how much of this burgeoning market it can effectively capture as a leader.
**Navigating the “Industry’s First” Claim**
The “industry’s first” designation, while attention-grabbing, warrants a closer examination. In June, rival Ericsson announced the commercial rollout of its own AI-in-RAN software subscription. Ericsson claims its solution delivers up to 20% higher downlink throughput and a 10% improvement in spectral efficiency across more than 15 live deployments. Crucially, Ericsson’s offering operates on existing baseband silicon, eliminating the need for specialized GPUs. This means Ericsson is already present in the market with an AI-enhanced RAN solution.
Nokia’s claim to originality hinges on a more specific definition: a GPU-accelerated AI-RAN platform. This signifies a distinct architectural approach compared to layering AI functionalities onto existing hardware. Both claims can coexist, highlighting the nuanced nature of market leadership and the importance of scrutinizing technological distinctions beyond broad pronouncements.
The divergence in approach runs deeper than mere timing. Nokia’s radio roadmap is intrinsically linked to NVIDIA’s ecosystem, with Chief Technology Officer Pallavi Mahajan acknowledging that certain Layer 1 software components are tied to the underlying hardware. Ericsson, conversely, has deliberately pursued a silicon-agnostic strategy for its AI features, aiming to avoid vendor lock-in.
While Nokia emphasizes its adherence to Open RAN principles and its use of merchant silicon from companies like Marvell within its broader ecosystem, the performance gains it is currently marketing—the spectral efficiency improvements—are directly enabled by NVIDIA’s proprietary stack. There is currently no readily available alternative to NVIDIA’s solution for achieving these specific performance metrics. This duality, where Nokia promotes openness in its messaging while engineering its high-performance solutions on a specific vendor’s proprietary technology, is a defining characteristic of this launch.
This strategy is not inherently flawed. Outsourcing the complex and capital-intensive silicon development race to NVIDIA, the undisputed leader in AI chips, is a pragmatic response to a business segment Nokia has struggled to revive independently. The subscription model is also a significant advantage, promising the recurring revenue that Nokia’s traditional hardware sales have historically lacked.
However, it’s crucial to note that Nokia’s AI-RAN platform is not yet commercially available, and its most impressive efficiency figures are still two years away. Furthermore, a major competitor has already entered the market with a different, albeit less GPU-dependent, AI-enhanced RAN solution. For Nokia, this represents a comeback in progress, not a victory already secured. The company’s future trajectory in this vital market is now, for better or worse, inextricably linked to its partnership with NVIDIA.
Original article, Author: Samuel Thompson. If you wish to reprint this article, please indicate the source:https://aicnbc.com/23749.html
