At MWC2026 Operators and Vendors Race to Turn Networks into AI Platforms — and to Claim 6G Turf

At MWC2026 vendors portrayed mobile networks as the foundational platform for AI, pushing 5G‑Advanced and early 6G work while unveiling edge compute to host large models. Huawei, Ericsson and Qualcomm outlined complementary strategies — integrated super‑nodes, an Intelligent Fabric and end‑to‑end device–network–cloud architectures — as operators and regulators grapple with spectrum, security and vendor dominance.

Abstract illustration depicting complex digital neural networks and data flow.

Key Takeaways

  • 1Vendors pitched an "AI‑Centric Network" model at MWC2026 to transform telecoms from data pipes into distributed AI platforms.
  • 2Huawei showcased Atlas 950 and TaiShan 950 SuperPoD super‑nodes overseas for the first time, emphasising UnifiedBus interconnects and large‑scale on‑premise training capability.
  • 3U6GHz is emerging as a pivotal band for 5G‑A and a transitional step toward 6G; commercial device and CPE support is expected in 2026–27.
  • 4Qualcomm and Nokia Bell Labs demonstrated wireless AI interop and large‑MIMO work as part of an end‑to‑end 6G engineering push aimed at pre‑commercial terminals by 2028.
  • 5Industry aims to complete 5G deployments, embrace AI challenges and secure networks and data — the three ‘mountains’ identified by GSMA leadership.

Editor's
Desk

Strategic Analysis

MWC2026 revealed that the telecom industry’s strategic axis has shifted from radio throughput to compute and intelligence. Whoever controls the orchestration layer and the distributed compute fabric will not only supply networks but also gatekeep data and services that run on them. That raises geopolitical and competitive stakes: nations will press for local control over compute and data flows, operators will demand open interfaces to avoid lock‑in, and vendors will balance openness with the commercial logic of integrated stacks. Regulators must therefore prioritise spectrum harmonisation and interoperability standards while updating rules on data governance and security to reflect networks that make autonomous decisions. The effective winners will be those who can sell operators a believable path to both technical performance and regulatory compliance — and that contest will shape how AI is embedded into everyday connectivity.

China Daily Brief Editorial
Strategic Insight
China Daily Brief

Barcelona’s Mobile World Congress opened under a familiar paradox: mobile networks remain both indispensable and unfinished. As vendors and operators paint the coming era as one in which networks will host and enable artificial intelligence at scale, the immediate agenda is simultaneously mundane and strategic — finish the commercial work of 5G, retool networks for AI-driven services, and harden infrastructure against growing security and data‑sovereignty demands.

On day one of MWC2026, industry heavyweights sketched complementary blueprints for that future. Huawei placed “AI‑Centric Network” at the centre of its stand, arguing that cloud, AI and mobile connectivity must knit together to make an “intelligent internet of agents”. The company rolled out new tools for operators — an A2A‑T intent interface jointly developed with TM Forum for cross‑vendor orchestration, a RAN Agent that links wireless digital twins and communication models, and a suite of U6GHz products it says will unlock the next phase of 5G‑Advanced (5G‑A) while smoothing the path to 6G.

Ericsson framed a similar vision in less proprietary language, promoting an “Intelligent Fabric” that treats the network as a distributed cognitive platform. Its pitch emphasises edge inference and a proliferation of machine agents on phones, wearables, vehicles and sensors that will interact continuously and learn cooperatively. The message was the same: networks must evolve from pipes into emergent infrastructure that senses, predicts and acts in near real time.

Crucially, MWC became a forum for compute as well as radio innovation. Huawei’s first overseas display of Atlas 950 and TaiShan 950 SuperPoD “super‑nodes” signalled how network vendors plan to collapse networking and data‑centre stacks. Built around a proprietary UnifiedBus interconnect and capable of logically presenting thousands of accelerator cards as a single machine, these systems are pitched at pre‑training large models and running latency‑sensitive AI workloads close to the radio.

Qualcomm, meanwhile, emphasised a more distributed end‑to‑end approach. It is investing across terminals, radio access research and cloud integration, demonstrating large‑scale MIMO tests and wireless AI interop with Nokia Bell Labs to validate hybrid on‑device and cloud workflows. Qualcomm’s narrative anticipates a heterogeneous device ecosystem — smartphones, AI PCs, cars, robots and XR headsets — that will carry increasing AI capability and help define the early contours of 6G pre‑commercial devices around 2028.

The industry’s near‑term roadmap is anchored to specific technical and regulatory milestones. The U6GHz band, identified as a global mobile candidate at WRC‑23, features heavily in vendor roadmaps for 5G‑A because it offers a sweet spot of bandwidth and coverage. Vendors are racing to ensure chipsets and customer premises equipment are ready in 2026–27, giving operators practical levers to deliver the low latency and capacity that AI services will demand.

The ambitions on stage at MWC are wide ranging, but three structural tensions run through them. First, there is a commercial tussle over who supplies the integrated stacks — radio, software and massive on‑premise compute — that will underpin AI‑driven services. Second, regulators and operators must reconcile global spectrum harmonisation and national security pressures while allocating U6GHz and other bands. Third, the concentration of AI training capability at the network edge raises questions about interoperability, vendor lock‑in and how privacy and data governance will be enforced when intelligence is both distributed and proprietary.

For operators the opportunity is clear: AI can lower operating costs, raise network quality and create new monetisable services from differentiated network slices. For cloud and silicon vendors the prize is control of compute and data flows. For regulators and civil‑society actors the challenge is to shape rules that preserve competition and protect users as networks become active decision‑makers in social and economic life. MWC2026 made one message plain — the next decade of mobile is not just about speed; it is about intelligence, control of compute, and the battle over who sets the rules for a newly agentic internet.

Share Article

Related Articles

📰
No related articles found