NVIDIA bets big on AI-native networks at MWC 2026

NVIDIA arrives at MWC 2026 with live AI-RAN field results, a global 6G coalition, and new open-source tools for autonomous networks, signaling the telco AI buildout is no longer a roadmap item.

NVIDIA used Mobile World Congress Barcelona as the stage for a sweeping push into the future of telecommunications, announcing a coalition to build 6G on AI-native platforms, unveiling new tools for autonomous network operations, and releasing field trial results that show its AI-RAN architecture is ready for commercial deployment.

Taken together, the announcements mark a significant step in NVIDIA's bid to position itself at the heart of the world's next-generation wireless infrastructure.

A global coalition for AI-native 6G

In arguably the most headline-grabbing announcement of the week, NVIDIA secured commitments from more than a dozen global telecom and technology heavyweights–including BT Group, Deutsche Telekom, Ericsson, Nokia, SK Telecom, SoftBank, T-Mobile, Cisco and Booz Allen–to build 6G on open, secure and AI-native software-defined platforms.

The initiative is underpinned by a key argument: legacy wireless architectures were not designed for a world where billions of autonomous machines, vehicles, sensors and robots depend on the network to see, sense and act. By embedding AI across the radio access network (RAN), edge and core, Nvidia and its partners believe 6G must be rebuilt from the ground up as intelligent infrastructure.

"AI is redefining computing and driving the largest infrastructure buildout in human history–and telecommunications is next," said Jensen Huang, Nvidia's founder and CEO. The company is participating in 6G initiatives across the US, UK, Europe, Japan and Korea, including the FutureG Office-led OCUDU Initiative and the AI-RAN Alliance, which now counts over 130 member companies.

AI-RAN moves from lab to live

Beyond the policy commitments, Nvidia and Nokia announced new AI-RAN collaborations with operators across Europe, Asia and North America–and crucially, real-world results to back them up.

T-Mobile US demonstrated concurrent AI and RAN processing on Nvidia's AI-RAN platform using Nokia's CUDA-accelerated RAN software, running video streaming, generative AI and AI-powered captioning applications alongside 5G over the air.

SoftBank achieved an industry-first 16-layer massive MIMO using fully software-defined 5G on Nvidia hardware–a meaningful technical milestone on the path to AI-RAN commercialization.

Indonesia's Indosat Ooredoo Hutchison (IOH) took things further, showcasing Southeast Asia's first AI-powered 5G call at MWC–including remote control of a robotic dog over a live 5G network–as it moved from proof of concept to pre-commercial field validation.

Meanwhile, SynaXG set a new benchmark by running 4G, 5G sub-6GHz and millimeter wave bands alongside agentic AI workloads on a single NVIDIA GH200 server, achieving 36 Gbps throughput and under 10 milliseconds latency. It was also the world's first AI-RAN implementation on FR2 millimeter wave bands.

MWC 2026 will see triple the AI-RAN innovation showcased compared to last year, with 26 of 33 AI-RAN Alliance demos built on NVIDIA AI Aerial and a software-defined architecture.

Agentic AI comes to network operations

On the operations side, NVIDIA announced a suite of tools designed to help telecom operators move toward fully autonomous network management–networks that don't just automate pre-defined tasks, but understand operator intent and make their own decisions.

Central to this is a new open source, 30-billion-parameter Nemotron Large Telco Model (LTM), developed in collaboration with AdaptKey AI and built on Nvidia's Nemotron 3 foundation model family.

Fine-tuned on open telecom datasets including industry standards and synthetic logs, the model is designed to reason through workflows like fault isolation and remediation planning. Crucially, it gives operators full transparency into training data and supports secure on-premises deployment.

NVIDIA has also published–alongside Tech Mahindra–an open source guide on building AI agents that reason like network operations center (NOC) engineers, using the Nvidia NeMo-Skills pipeline to fine-tune models on structured reasoning traces.

New Nvidia Blueprints for RAN energy efficiency and network configuration round out the offering, with the energy blueprint integrating VIAVI's simulation platform to allow operators to safely test energy-saving policies without touching live configurations.

Early adopters are already putting the network configuration blueprint to work. Cassava Technologies is deploying it as the basis of an autonomous network platform for Africa's multi-vendor mobile environment, while NTT DATA is using it to help a tier 1 Japanese operator manage traffic surges after outages.

Norwegian operator Telenor Group will be the first to adopt an enhanced version with partner BubbleRAN for its maritime connectivity arm.

What it means for the channel

For the partner ecosystem, Nvidia's MWC push signals both opportunity and urgency. The company is actively expanding the hardware ecosystem around its AI-RAN platforms: QCT is releasing commercial on-the-shelf AI-RAN products, Supermicro is extending support across the full NVIDIA AI-RAN portfolio, and WNC has introduced a new AI-optimized radio unit supporting 5G Advanced and 6G use cases.

According to NVIDIA's own State of AI in Telecom report, 77% of respondents anticipate a significantly faster deployment timeline for AI-native wireless network architecture than for previous network generations.

With field trials clearing technical milestones and a growing coalition locking in on open, software-defined 6G foundations, the window for partners to establish themselves in this space is opening–and it appears to be opening fast.