For years, AI-native networks have been a visionary topic within the telecom industry, often associated with future 6G technology. However, Mobile World Congress (MWC) 2026 in Barcelona marked a pivotal moment by providing concrete evidence that these networks are rapidly transitioning from concept to commercial reality.
Major announcements from global telecom vendors, chip manufacturers, and network operators revealed not only the strategic vision for AI-driven Radio Access Networks (AI-RAN) but also detailed field trial outcomes, product launches, and collaborative commitments to build the next generation of connectivity on AI-native foundations.
Nvidia and Global Leaders Unite for AI-RAN and 6G Development
Nvidia emerged as a central figure at MWC 2026, securing commitments from over a dozen prominent operators and technology companies—including BT Group, Deutsche Telekom, Ericsson, Nokia, SK Telecom, SoftBank, T-Mobile, Cisco, and Booz Allen—to jointly develop 6G networks on open, secure, AI-native, software-defined platforms.
This initiative, supported by collaborations with governments in the US, UK, Europe, Japan, and Korea, aims to create intelligent, resilient, and trustworthy future connectivity infrastructure. Nvidia’s CEO Jensen Huang emphasized the transformative nature of AI on computing and telecommunications, framing it as the largest infrastructure expansion in human history.
As a founding member of the AI-RAN Alliance—which now includes over 130 companies—Nvidia also introduced an extensive suite of open-source tools for network operators. These include a 30-billion-parameter Nemotron Large Telco Model tailored to telecom datasets, an AI agent development guide co-created with Tech Mahindra, and new blueprints targeting RAN energy efficiency and network configuration.
Notably, the energy efficiency blueprint uses VIAVI’s TeraVM AI RAN Scenario Generator to simulate energy-saving policies before applying them to live networks. Early adopters like Cassava Technologies in Africa and NTT DATA in Japan are already deploying these innovations in real operational settings.
Nokia and Operators Demonstrate Live AI-RAN Deployments
Nokia showcased significant advancements by successfully testing its anyRAN software on Nvidia’s GPU-accelerated AI-RAN platform in live, over-the-air environments with partners such as T-Mobile US, Indosat Ooredoo Hutchison, and SoftBank.
For example, at T-Mobile’s AI-RAN Innovation Centre in Seattle, Nokia’s AirScale Massive MIMO radio operated concurrent AI and RAN workloads—including video streaming and generative AI tasks—on a single Nvidia Grace Hopper 200 server integrated with commercial 5G services.
Indosat Ooredoo Hutchison achieved Southeast Asia’s first AI-RAN-powered Layer 3 5G call during MWC, emphasizing the technology’s role in expanding digital inclusion across Indonesia. SoftBank demonstrated how unused compute capacity identified by its AITRAS Orchestrator can run third-party AI workloads, hinting at new revenue opportunities for operators beyond traditional connectivity services.
The AI-RAN ecosystem surrounding Nokia continues to grow, with partners like Dell Technologies, Quanta, Supermicro, and Red Hat OpenShift providing orchestration and hardware options. Nokia’s stock saw a notable increase following these announcements.
Ericsson’s Distinct Path with Custom Silicon AI-Ready Radios
In contrast to Nokia and Nvidia’s GPU-accelerated approach, Ericsson introduced ten new AI-ready radios featuring embedded neural network accelerators within their Massive MIMO hardware, eliminating the need for external GPUs.
This custom silicon strategy focuses on reducing total cost of ownership and enhancing power efficiency, while also ensuring supply chain independence. Ericsson’s portfolio includes AI-managed beamforming, outdoor positioning, instant coverage prediction, and latency-prioritized scheduling with significantly faster response times.
Additionally, Ericsson announced a broad collaboration with Intel to accelerate readiness for AI-native 6G networks, combining compute, cloud technologies, and AI-driven RAN and packet core solutions. This partnership aims to deliver open, power-efficient networks leveraging Intel’s advanced semiconductor process nodes.
Telecom Operators Embrace AI-RAN for Infrastructure Transformation
Operators SK Telecom and SoftBank articulated comprehensive strategies integrating AI-RAN within their infrastructures. SK Telecom plans a full-stack AI-native rebuild, increasing its sovereign AI foundation model to over one trillion parameters and establishing a new AI data center in collaboration with OpenAI.
The company is expanding autonomous network operations to automate wireless quality management and traffic control, with AI-RAN playing a critical role in enhancing speed and reducing latency.
SoftBank demonstrated its Autonomous Agentic AI-RAN system, which uses a Large Telecom Model to translate natural language operator goals into dynamic 5G and 6G network configurations—an important step toward self-managing networks driven by intent rather than manual input.
Emerging Hardware Ecosystem Supports AI-RAN Commercialization
The growth of AI-RAN is also evidenced by the expanding hardware ecosystem. Quanta Cloud Technology announced commercial off-the-shelf AI-RAN products compatible with Nvidia ARC platforms and Nokia software, while Supermicro extended support for Nvidia’s full AI-RAN portfolio.
MSI introduced a unified AI-vRAN platform enabling dynamic GPU allocation between 5G and AI workloads, and Lanner Electronics launched AstraEdge AI Servers designed to co-locate AI inference, RAN functions, and packet processing at cell sites. AMD positioned its EPYC 8005 edge platform and Open Telco AI initiative as alternative compute solutions for operators transitioning from pilots to production.
Broader Implications for Enterprises and the Future of Connectivity
The advancements showcased at MWC 2026 signal a fundamental architectural shift in telecom infrastructure, with AI-RAN networks evolving continuously via software updates rather than costly hardware replacements. This trend aligns connectivity infrastructure more closely with cloud computing in terms of flexibility and pace of innovation.
Embedding GPU compute within the RAN also opens possibilities for running enterprise AI workloads at the network edge, closer to data sources. According to Nvidia’s State of AI in Telecom report, a majority of industry respondents expect AI-native wireless architecture deployment to outpace previous network generations significantly.
The contrasting approaches between Ericsson’s custom silicon and Nokia-Nvidia’s GPU acceleration highlight ongoing industry debates about optimal AI inference placement and cost-effectiveness, influencing future operator and vendor decisions.
Ultimately, MWC 2026 made clear that AI-native networks are no longer distant ambitions tied solely to 6G research. They are live, actively deployed, and supported by robust ecosystems, with the primary questions now focused on the speed of adoption and leadership in this transformative wave.
Fonte: ver artigo original
Related Articles
Malaysia Captures 32% of Southeast Asia’s AI Funding Amid Major Infrastructure Expansion
Arcada Labs Launches Benchmark Testing Autonomous AI Agents on Social Media Platform X
Arm’s Vision for AI at the Edge Signals a Shift in Computing Paradigms
ScaleOps Launches AI Infrastructure Platform to Cut GPU Costs by Up to 70% for Enterprises Running Self-Hosted LLMs
