Mobile World Congress 2026 in Barcelona marked a pivotal moment for AI-native networks, transcending years of discussion to deliver tangible evidence of their emergence. A flurry of announcements from leading telecom vendors, chipmakers, and operators showcased field trial results, commercial product launches, open-source toolkits, and a multi-operator commitment to build 6G on AI-native foundations. For enterprise and IT decision-makers, this signifies a fundamental architectural shift in telecom infrastructure, poised to redefine how connectivity is delivered, managed, and monetized.

Nvidia played a central role, securing commitments from over a dozen global operators and technology companies, including BT Group, Deutsche Telekom, Ericsson, Nokia, SK Telecom, SoftBank, T-Mobile, Cisco, and Booz Allen. This coalition is dedicated to constructing 6G on open, secure, and AI-native software-defined platforms, ensuring future connectivity infrastructure is intelligent, resilient, and trustworthy. Nvidia's founder and CEO, Jensen Huang, emphasized the stakes, stating, “AI is redefining computing and driving the largest infrastructure buildout in human history–and telecommunications is next.” Nvidia is a founding member of the AI-RAN Alliance, which now boasts over 130 companies, and has joined the US FutureG Office-led OCUDU Initiative to accelerate open, software-defined, AI-native 6G architectures. The company further released a suite of open-source tools for network operators, including a 30-billion-parameter Nemotron Large Telco Model (LTM) developed with AdaptKey AI and fine-tuned on telecom datasets. Additionally, Nvidia co-published an open-source guide with Tech Mahindra for building AI agents that reason like NOC engineers and introduced new Nvidia Blueprints for RAN energy efficiency and network configuration. The energy blueprint integrates VIAVI’s TeraVM AI RAN Scenario Generator for policy simulation, while the network configuration blueprint is already being deployed by Cassava Technologies for autonomous networks in Africa and by NTT DATA with a tier one operator in Japan to manage traffic surges.

Nokia demonstrated significant progress in its strategic AI-RAN partnership with Nvidia, successfully completing functional tests of its anyRAN software on NVIDIA’s GPU-accelerated AI-RAN platform with T-Mobile US, Indosat Ooredoo Hutchison (IOH), and SoftBank Corp. These validations were conducted in live, over-the-air conditions rather than controlled lab environments. At T-Mobile’s AI-RAN Innovation Centre in Seattle, Nokia’s AirScale Massive MIMO radio concurrently ran AI and RAN workloads on a single Nvidia Grace Hopper 200 server alongside commercial 5G, handling tasks like video streaming, generative AI queries, and AI-powered video captioning. IOH achieved Southeast Asia’s first AI-RAN-powered Layer 3 5G call, with both AI and RAN workloads running simultaneously on shared GPU infrastructure, demonstrating the technology's potential to benefit every Indonesian. SoftBank’s demonstration further showcased how its AITRAS Orchestrator could identify spare compute capacity to run third-party AI workloads, hinting at future monetization strategies for RAN infrastructure beyond mere connectivity. Nokia’s expanding AI-RAN ecosystem now includes Dell Technologies, Quanta, Supermicro, and Red Hat OpenShift for orchestration, providing operators with a growing range of commercial off-the-shelf options.

Ericsson presented a distinctly different approach to AI-native networks at MWC 2026. While Nokia has invested heavily in Nvidia GPU acceleration, Ericsson unveiled ten new AI-ready radios built on its own purpose-built silicon, featuring neural network accelerators directly embedded into its Massive MIMO hardware, thus eliminating the need for NVIDIA GPUs. This portfolio includes AI-managed beamforming, AI-powered outdoor positioning, instant coverage prediction using AI models, and a latency-prioritized scheduler delivering up to seven times faster response times. Ericsson's argument hinges on total cost of ownership, contending that custom silicon offers superior TCO and power efficiency compared to external GPU hardware, alongside the benefit of supply chain independence. Per Narvinger, head of Ericsson’s mobile networks business, indicated this stance is firm. Ericsson also announced a broad collaboration with Intel, encompassing compute, cloud technologies, and AI-driven RAN and packet core use cases, aimed at accelerating ecosystem readiness for AI-native 6G. Ericsson President and CEO Börje Ekholm underscored that “6G is not merely an iteration of mobile technology. It is the infrastructure that will distribute AI across devices, the edge and the cloud.” Intel CEO Lip-Bu Tan framed the partnership as a pathway to open, power-efficient networks grounded in AI inference, with future Ericsson Silicon to be built on Intel’s most advanced process nodes.

Beyond vendor announcements, two operators, SK Telecom and SoftBank, detailed how deeply AI-RAN integrates into their broader infrastructure strategies. SK Telecom CEO Jung Jai-hun outlined a full-stack AI-native rebuild, from its network core to customer service systems, including plans to upgrade its sovereign AI foundation model to over one trillion parameters and to build a new AI data center in Korea in collaboration with OpenAI. The company is also expanding autonomous network operations using AI to automate wireless quality management, traffic control, and network equipment, with AI-RAN technology crucial for improving speed and reducing latency. SoftBank, in collaboration with Northeastern University’s INSI, Keysight Technologies, and zTouch Networks, demonstrated its Autonomous Agentic AI-RAN (AgentRAN) system. This system leverages SoftBank’s Large Telecom Model to translate natural-language operator goals into real-time 5G and 6G network configurations, representing a significant step towards intent-based, self-managing networks.

The maturation of AI-RAN from concept to commercial infrastructure is evident in the burgeoning hardware ecosystem. At MWC 2026, Quanta Cloud Technology introduced commercial off-the-shelf AI-RAN products supporting Nvidia ARC platforms and Nokia software. Supermicro extended support across the full Nvidia AI-RAN portfolio, including ARC-Pro and RTX 6000-based configurations. MSI unveiled its unified AI-vRAN platform with dynamic GPU allocation between 5G and AI workloads. Lanner Electronics launched its AstraEdge AI Server lineup–the ECA-6710 and ECA-5555–specifically designed to co-locate AI inference, RAN functions, and high-performance packet processing at cell sites. AMD positioned its EPYC 8005 edge platform and Open Telco AI initiative as an alternative compute path for operators moving from AI pilots to production.

For enterprise decision-makers, the implications extend beyond mere telecom infrastructure procurement. AI-RAN networks, evolving continuously through software rather than costly hardware refresh cycles, will offer connectivity infrastructure with a cloud-like pace of change and flexibility. The integration of GPU compute within the RAN opens avenues for enterprise AI workloads to run at the network edge, closer to data generation. Nvidia’s State of AI in Telecom report highlights that 77% of respondents anticipate significantly faster deployment timelines for AI-native wireless architecture compared to previous network generations. The architectural debate between Ericsson’s custom silicon path and Nokia-Nvidia’s GPU-accelerated approach is a critical one to monitor, reflecting a fundamental question about the optimal placement and cost of AI inference in network hardware. This debate will undoubtedly influence operator procurement decisions and vendor relationships for years to come. MWC 2026 unequivocally demonstrated that AI-native networks are no longer a research agenda. Field trials are live, hardware is shipping, and coalitions are forming. The pressing question for enterprises and operators alike is no longer if this transition will occur, but rather its speed and who will lead it.