Nokia bangs louder on the AI-RAN drum

• With Nvidia on board as an investor, Nokia is pushing the AI-RAN agenda hard
• The vendor recently proclaimed the “industry’s first AI-native RAN platform”
• It is engaging with ever increasing numbers of telcos for AI-RAN tests and trials

As keen as ever to pitch itself as the frontrunner in AI-RAN engagements, Nokia has provided an updated list of telcos with which it is engaged in proofs of concept (PoCs) and/or trials for its radio access network (RAN) system that is based on the Aerial AI-RAN platform developed by Nvidia. 

The announcement is noteworthy (for reasons we will come to shortly), though it doesn’t offer much that is new other than to identify a number of telcos that are putting Nokia’s AI-RAN solution through its paces. Those operators are: Austria’s A1 Group (Telekom Austria); Taiwan’s Chunghwa Telecom; du and e& in the United Arab Emirates (UAE); Mobily, STC and Zain Saudi in Saudi Arabia; and TPG Telecom in Australia. 

The vendor also notes that trials are continuing with other operators, including Japan’s NTT Docomo, and references its ongoing evaluations with T-Mobile US, Japan’s SoftBank Corp. and Indosat Ooredoo Hutchison (IOH) in Indonesia, operators with which Nokia has already completed “successful functional tests of GPU-accelerated AI-RAN”.

And while Nokia didn’t this time around namecheck the UK’s BT Group, Finland’s Elisa and Vodafone – all of which were highlighted as being engaged in AI-RAN activities earlier in the year – Nokia says they are still very much on board and that more about those engagements will be shared later in the year.  

What all of this reflects is a level of momentum for Nokia’s AI-RAN proposition, though the vendor’s claim that such momentum “demonstrates how AI-RAN is rapidly moving from industry vision to commercial reality” is somewhat jumping the gun.

But it is at least more promising for Nokia and Nvidia than in early 2025, when many operators that TelecomTV spoke to on background were dismissive of the immediate interest of AI-RAN platforms, mainly because of the likely cost and power consumption implications, though such concerns were always batted away by Nvidia. 

But Nokia and Nvidia have been working on this, have deployed a more RAN-appropriate GPU chip in the AI-RAN system, and recently made some very bold claims about projected spectral efficiency gains to be had from the use of the platform. 

In July, Nokia announced it had integrated its anyRAN software with Nvidia’s Aerial AI-RAN platform, which was first unveiled about two years ago, to create what the Finnish vendor called the “industry’s first AI-native RAN platform”, which is due to become commercially available next year. (As a reminder, Nvidia is now at the heart of Nokia’s RAN strategy thanks not only to its technological prowess and clout but also because the AI technology giant (currently with more than $5tn) is a shareholder in the Finnish vendor, having pumped $1bn into Nokia in October last year.)

Specifically, it boasted that the AI-RAN platform enables mobile operators to “unlock significantly more uplink and downlink capacity from the spectrum and radio infrastructure they already own”, adding that the AI-RAN platform “has already shown more than 20% spectral efficiency gains through [unspecified] AI-driven radio innovations… [and] is on track to deliver 50% spectral gains by 2027 and more than 100% by 2028” – effectively doubling the capacity of existing spectrum assets. (TelecomTV understands that those spectral efficiency figures are based on the vendor’s internal testing, validation and roadmap projections, with the results based on controlled conditions: Ultimately, real-world performance will have to be validated with operators during trials.)

These are the sorts of advances mobile operators want to hear about and check out as they advance their 5G deployments and plan for the 6G era, when spectral efficiency and AI-enabled automated operations and new services supported by network edge processing look set to play a key role in delivering advanced services, such as integrated sensing and communications (ISAC). 

And it seems likely that more operators will sign up as the technology evolves further. 

Remy Pascal, practice leader for mobile infrastructure at research firm Omdia, noted: “The breadth of operator engagement announced by Nokia shows that AI-RAN is evolving from a research topic into a strategic priority for a growing number of service providers. What’s notable is the diversity of operators across regions, reflecting an emerging industry consensus that AI will be essential to continue improving network performance and operational efficiency, and to the evolution towards 6G. As deployments mature, collaboration between operators, vendors and the broader ecosystem will be critical to realising AI-RAN’s full potential.”

That’s all very well, but ultimately what the industry needs is some quantifiable, independent proof points that can back up the claims being made for AI-RAN architectures – that would surely lead to much greater operator engagement. 

- Ray Le Maistre, Editorial Director, TelecomTV