Nokia’s AI-RAN platform was announced on July 15 with the worthy claim of being an industry first. The vendor says the platform, built on its anyRAN software and NVIDIA’s Aerial system, will allow carriers to extract far more capacity from the spectrum they already own, positioning the announcement as one of the most significant wireless architecture changes in decades.
The technical pitch is simple. Nokia says the platform has already shown an increase in spectral efficiency of more than 20%, with the company targeting 50% by 2027 and more than 100% by 2028, at which point carriers will be able to roughly double the capacity of their existing spectrum, according to its own projections. The last two numbers are targets, not results, and Nokia’s own timeline calls for pilots by the end of this year and commercial availability in 2027.
Operators purchase capabilities through software subscriptions rather than hardware updates and choose from three deployment options: GPU-powered plug-in cards for existing AirScale sites, standalone AI-RAN nodes, or cloud server builds offered through partners.
The revival of Nokia’s weakest business
If you understand this announcement only as a product story, you’ll miss why it’s important to Nokia. Radio has been CEO Justin Hotard’s toughest challenge since he took over in 2025. At Nokia’s capital markets day in November, he told investors that the mobile business was not delivering satisfactory returns and said the company had folded it into a new mobile infrastructure unit, alongside further cost reductions.
The partnership with NVIDIA, announced in October 2025, is at the center of that repair effort, with a $1 billion investment from the chipmaker for a roughly 3% stake. By building on NVIDIA’s silicon and CUDA software rather than its own custom chips, Nokia is able to eliminate some of its expensive in-house R&D and redirect it to software. This shift, Hotard describes, is a departure from the traditional hardware model.
Investors rewarded the story. Nokia’s stock price has been significantly revalued into 2026 due to strong AI and cloud momentum, and the AI-RAN announcement comes just days before Q2 earnings. Omdia, whose analyst Remy Pascal is cited in Nokia’s own announcement, predicts that the cumulative opportunity for AI-RAN will exceed $200 billion by 2030. This direction is realistic. The open question is how much of that Nokia can claim as a lead.
Is the Nokia AI-RAN platform really the first?
Be wary of the “industry first” label here. In June, Ericsson began selling subscriptions for its commercial AI-in-RAN software. The company says it delivers up to 20% higher downlink throughput and up to 10% better spectral efficiency in more than 15 live deployments, and more importantly, runs on carriers’ existing baseband silicon without the need for GPUs. In terms of availability, Ericsson is already on the market.
Nokia’s first argument is based on a narrower definition. That means it’s a GPU-accelerated AI-RAN platform, a different architecture with AI capabilities layered on top of existing hardware. Both statements can be true simultaneously, which is precisely why the framework deserves scrutiny rather than simple repetition.
But this discrepancy goes beyond timing.
Nokia has tied its wireless roadmap to NVIDIA, and the company’s chief technology officer, Pallavi Mahajan, has acknowledged that at least some of its Layer 1 software is tied to the underlying hardware. To avoid that dependency, Ericsson took the opposite route by design, making its AI capabilities silicon-agnostic.
Although Nokia points to Marvell’s merchant silicon within its broader ecosystem and describes its platform as Open RAN compliant, the performance case it sells – improving spectral efficiency – is currently performed through NVIDIA’s stack, and no comparable replacement currently exists. Openness in messaging and dependence on NVIDIA for engineering are both hallmarks of the same release.
None of these strategies are wrong. Outsourcing the silicon competition to the industry’s leading AI chip suppliers was a legitimate answer to a business Nokia was struggling to solve on its own, and the subscription model would bring recurring revenue to radio that it didn’t get in the hardware cycle.
But the platform has not yet been commercialized, its leading efficiency figures are still two years away, and at least one major rival reached the market first by a different route. For Nokia, this is not a triumphant victory, but an ongoing resurgence whose trajectory, for better or worse, runs through NVIDIA.
SEE ALSO: AI-native networks are no longer a 6G promise – MWC 2026 proved it
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events such as Cyber Security & Cloud Expo. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
