Nokia recently highlighted major advancements in its AI‑RAN strategy in partnership with Nvidia at MWC26 Barcelona. The event helped the company to showcase operator integrations and successful functional tests of its GPU-accelerated AI-RAN.
At a media and analyst conference on the eve of MWC26 Barcelona, the vendors noted the collaboration pushes the shift toward AI‑native 6G, with operators across the US, Europe and Asia validating the potential of GPU‑accelerated RAN.
Nokia completed tests of its AI and RAN workloads on Nvidia’s GPU-accelerated AI-RAN, to demonstrate the scalability and efficiency of the platform.
The two vendors stated the tests validate the readiness of AI-RAN for real-world deployments and its ability to deliver enhanced user experiences and unlock new monetisation opportunities through advanced services.
Nokia touted new customer integrations and successful functional tests of its anyRAN software on Nvidia’s GPU AI-RAN platform with T-Mobile US, Indosat and SoftBank Corp.
At T-Mobile’s AI-RAN Innovation Centre in its US headquarters, Nokia’s AirScale Massive MIMO radio operating in the 3.7GHz band supported equipment running use cases such as video streaming, generative AI queries and AI-based video captioning tests, running on the Nvidia Grace Hopper 200 platform.
Pallavi Mahajan, chief technology and AI officer at Nokia, stated during a panel the trial with T-Mobile “is a good proof point of showing the world” where the vendor and ecosystem is headed.
“Now, 2026 is where we will do our first commercial trial, and 2027 is where we will have our first commercial release out”, she said. “This is stuff that actually excites us a lot”.
Nokia and Indosat Ooredoo Hutchison (IOH) achieved what they claim is Southeast Asia’s first AI RAN–powered Layer 3 5G call. The call used IOH’s open, cloud native network with Nokia’s AirScale remote radio heads (RRHs) and RAN software accelerated by Nvidia’s GPUs.
The Finnish vendor stated the milestone proves AI and RAN workloads can run simultaneously on shared GPU infrastructure in a live operator environment, paving the way for distributed AI intelligence which makes 5G networks more efficient, intelligent and sustainable.
Lastly, Nokia and SoftBank Corp showed how spare AI-RAN compute capacity, identified by SoftBank’s Aitras orchestrator, can be used to run third-party AI tasks.
Nokia explained the integration represents a major advance in evolving the RAN into an AI‑enabled platform which can host third‑party AI workloads, unlocking new service and revenue opportunities beyond traditional connectivity while adhering to AI‑RAN Alliance architectural principles.
In addition to those operators, BT Group, Elisa, NTT Docomo and Vodafone are engaged in trials with Nvidia and Nokia to confirm AI‑RAN’s readiness for real‑world deployments.
Nokia also expanded its ecosystem of AI-RAN partners, who are building the servers and infrastructure for AI-RAN. The new partners include Quanta and SuperMicro, who join Dell Technologies and build on Red Hat OpenShift for orchestration.
A call to AI arms
During the event, CEO and president Justin Hotard stated AI already generates massive traffic with 1.3 trillion annual sessions, over 100 trillion tokens daily, and 77 exabytes of monthly traffic, predominantly mobile-based, with human-to-machine interactions currently dominant.
He noted unlike previous linear traffic patterns in voice, data, and video, AI traffic is highly variable and dynamic, requiring networks to support fluctuating uplink and downlink demands with real-time adaptation.
Hotard explained the industry needs to collaborate on building AI-native networks instead of building closed architectures. “What we believe is that no single company can lead the transformation of AI networks. It needs rich and broad industry partnership”, he said.
Ericsson, Nokia team
On that note, long-time rivals Nokia and Ericsson teamed to advance adoption of open standards and promote compatibility across their rApp portfolios.
They jointly stated the collaboration supports multivendor SMO ecosystems, giving operators greater choice and “accelerating the transition towards autonomous networks”
Hotard noted partnerships need to extend “from silicon providers to system players like ourselves, to cloud providers, to application platforms, to expert services like security, and ultimately to operators and device manufacturers”.
He stated Nokia is focused on building open interfaces and providing interoperability and extensibility.
Speaking on the panel, John Saw, president, technology and CTO for T-Mobile, said advanced 5G and 6G networks are the connective tissue to serve as the nervous system for physical AI. “And physical AI, many believe, is going to be even bigger than generative AI”, he said.
The shift from digital-only AI to physical AI involves systems interacting with the real economy through devices like robots and vehicles. Saw said this requires networks to handle kinetic tokens for dynamic data requiring low latency and space-time coherence. “I think mobile operators have their licence to play in physical AI”, Saw said.
Nvidia’s Ronnie Vasishta, SVP for telecom at Nvidia, said to support AI workloads alongside traditional telecom functions, networks must evolve into software-defined, open compute fabrics.
“This is now a realisation that the underlying fabric can both enable AI applications to run, but also how AI can improve the radio access network itself”, he said during the panel. “This is kind of key as you start to think about the compute fabric that’s needed”.
Source: Mobile World Live
Image Credit: Nvidia & Nokia
