Nokia RAN platform is at the center of this update. Nokia’s AI-RAN platform is not just another product launch. It is a bet on which version of the telecom AI future will win: NVIDIA-backed acceleration and recurring software revenue, or a more conservative path that keeps operators closer to their existing silicon.
The company says its new platform, built on anyRAN software and NVIDIA’s Aerial system, is the industry’s first AI-native RAN platform. Ericsson, meanwhile, is already selling a commercial AI-in-RAN software subscription that runs on existing baseband silicon. That contrast is the story.
Nokia wants radio to look more like software, but it chose NVIDIA to get there
Nokia says the platform has already shown more than 20% spectral efficiency gains, with targets of 50% by 2027 and more than 100% by 2028. Those later figures are targets, not results. The company also says pilots are planned for the end of this year and commercial availability in 2027.
The architecture matters. Nokia is offering operators a software subscription and multiple deployment paths, including a GPU-powered plug-in card, a standalone AI-RAN node, and a cloud-server build through partners. That is a different play from a simple software patch on existing radio gear.
Ericsson is selling a different answer: AI without a GPU dependency
Ericsson’s approach, as described in the source, is to run AI features on existing baseband silicon without requiring GPUs. That makes its pitch less ambitious in one sense, but also less dependent on a single chip supplier.
This is the strategic split inside the AI-RAN market. Nokia is leaning into NVIDIA’s compute stack and the economics of a subscription model. Ericsson is leaning into compatibility and deployment simplicity. Both can be rational. They are just optimized for different buyers and different risk tolerances.
Justin Hotard’s repair job is bigger than one launch
The source places the launch inside Nokia chief executive Justin Hotard’s effort to fix the mobile business, which he has said has not delivered acceptable returns. Nokia has folded that business into a new Mobile Infrastructure segment and is cutting costs.
That makes the AI-RAN platform more than a technical announcement. It is part of a corporate reset: move away from a legacy hardware model, reduce in-house R&D pressure, and try to build a more durable revenue stream around software.
What operators may gain, and what still needs proving
If Nokia’s approach works, operators could squeeze more capacity from spectrum they already own and buy upgrades through software rather than a full hardware refresh. That is the promise.
The uncertainty is commercial, not just technical. Nokia’s platform is not yet broadly available, the strongest efficiency numbers are still future targets, and Ericsson is already in market with a competing model. The question is not whether AI will enter the radio network. It is which architecture operators will trust to carry it.
Editorial analysis only; not investment advice.
Related coverage: AI Chronicle analysis and updates.
Sources and further reading
- http://Nokia%20newsroom%20announcement%20on%20its%20AI-native%20RAN%20platform
- https://nvidianews.nvidia.com/
- https://www.reuters.com/technology/artificial-intelligence/
Related Articles
Apple Launches Siri AI Powered by Google’s Gemini, Leaving Major Markets Locked Out
Bluesky Enhances Moderation System with New Reporting Categories and Strikes
Salt Typhoon Hacking Group Targets Global Telecom Giants in Widespread Cyberattacks
Study Reveals AI Benchmarks Are Flawed Yet Persistently Used by Industry
