Leading the Next Era of Intelligent Connectivity 

Cisco’s AI-Native Wireless Network for Agentic and Physical AI

The mobile industry is entering one of the most profound transitions in its history. Networks are no longer simply channels for voice and data; they are becoming intelligent fabrics capable of sensing, reasoning, and acting in real-time. As generative and agentic AI applications accelerate, communication service providers sit at the epicenter of this evolution, shaping how intelligent machines, autonomous systems, and immersive experiences will connect and perform. 

Customer Challenges in an AI-Driven Wireless Era 

This opportunity comes with formidable hurdles. Inferencing workloads are growing at an unprecedented rate as models expand in size, frequency, and modality, ingesting video, audio, and sensor data continuously. That surge drives massive demand for compute, storage, and transport across the network. Uplink traffic, historically lighter than downlink, is soaring as intelligent endpoints push rich data upstream for real-time processing. Latency has become mission-critical: a few extra milliseconds can degrade performance for applications such as public safety, Industrial IoT or robotics amongst others. 

Telecom operators must scale capacity at the edge, orchestrate seamlessly from the core through the Radio Access Network, and secure critical infrastructure end-to-end. They need a network that is not only fast but also observable, adaptable, and trustworthy. And they must do all this at speed because the market will not wait. 

Cisco’s Two-Prong Approach: AI for Wireless and Wireless for AI 

Cisco addresses these demands with a powerful two-pronged approach with AI for Wireless and Wireless for AI. AI for Wireless embeds artificial intelligence directly into the network – from the radio layer to the core, Cisco’s architecture is built to sense, learn, and optimize. Edge inference brings AI processing closer to endpoints, enabling instant decision-making and ultra-low latency. Multimodal sensing, embedded with integrated sensing and communications (ISAC), computer vision, and environmental IoT inputs, creates networks that understand their surroundings and adapt in real time. This secure, resilient design is open and modular, ready to integrate new technologies and support the most demanding AI services. 

At the same time, Wireless for AI transforms the mobile network into a distributed AI engine. Here, the network itself becomes the supply chain hub for AI services, dynamically allocating compute, storage, network and spectrum resources to sustain uplink-heavy, inference-driven traffic. Cisco’s distributed User Plane Function (dUPF), integrated with NVIDIA accelerated computing platform, brings mobile data processing to the network edge, enabling ultra-low latency, secure, and resilient performance so agentic AI applications can sense, decide, and act in real-time. By keeping traffic local, dUPF optimizes bandwidth and safeguards sensitive data, delivering faster decisions, lower costs, and trusted outcomes over mission-critical 5G Advanced and 6G mobile infrastructure. Cisco’s Agile Services Networking provides the high-performance network fabric connecting core and edge, ensuring that inference can occur where data originates, driving efficiency and enabling next-gen applications from intelligent machines to immersive applications.  

These capabilities bring more than throughput and latency; they enable a new class of intelligent machines. Think of these endpoints—robots, industrial equipment, vehicles, sensors—as physical AI devices which learn, reason, and act, requiring real-time data and decision-making. Cisco’s AI-native wireless architecture becomes the foundation for physical AI, where digital intelligence is embodied in machines that perceive and interact with the physical world. Agentic and physical AI aren’t optional enhancements; they are the destination. Every architectural choice—from secure, low-latency transport to edge inference—is a means to that end. 

Accelerating Innovation: The AI-WIN Project 

Cisco is amplifying this vision through the AI-WIN project, a groundbreaking collaboration with NVIDIA, Booz Allen, MITRE, the O-RAN Development Company, and T-Mobile, announced earlier this year. In just six months, the team has built the first fully AI-native mobile network stack, from radio access and core mobility to orchestration and security, with three live pre-6G applications. This collaboration brings accelerated computing, open standards, and deep industry expertise to create a scalable, future-ready platform, spearheading American technology leadership in AI-Native 6G—a evolutionary step that builds on and is fully aligned with Cisco’s AI for Wireless and Mobility Services Platform strategy. While much attention focuses on 6G, the inflection point starts now and accelerates through 5G Advanced. The most exciting part of this development is the speed at which the collaboration was completed, providing additional evidence on the power of a software-defined architecture as the key to rapid innovation. While agile innovation is a hallmark of the software, cloud and AI era, it must also become the norm for the telecom industry as AI use cases explode.  

“Our partnership with Cisco puts U.S. leadership at the forefront of AI-native telecom,” said Ronnie Vasishta, senior vice president of telecom at NVIDIA. “Working together in the AI WIN project, we are combining Cisco’s mobile, networking, and security innovations with NVIDIA’s AI Aerial platform to deliver a high performance, software-defined and AI-native solution that can scale to meet operator’s needs for next-generation networks that go beyond connectivity.” 

The Business Impact 

For telecom operators, the rewards are substantial. AI-enabled services unlock new revenue streams, from real-time enterprise analytics and autonomous systems to advanced public safety and environmental sensing. McKinsey estimates that AI-driven telecom services that AI-driven telecom services could unlock over $150 billion in annual savings by 2030. This underscores what is at stake, and why service providers need to act now. Bringing inference closer to the edge lowers cost and latency dramatically, enabling experiences that were once impossible. In the era where bits are turning into tokens as traditional voice, video, and data applications become intertwined with AI, inherently, secure, multi-tenant and multi-modal infrastructure reduces overhead and risk. This will allow communication service providers to host diverse and distributed AI services efficiently, with cost-effective tokenomics. And integrated hardware and software shorten research to deployment cycles, turning innovation into faster revenue generating machines. 

Leading the Intelligent Connectivity to lay the foundation for Intelligent Machines  

Telecom providers are uniquely positioned to lead this transformation. They not only own the infrastructure, relationships, and regulatory expertise to deliver AI-powered services at scale but also have the inherent sovereign advantage in every nation. Cisco’s AI for Wireless and Wireless for AI strategies, combined with industry-leading partnerships like AI-WIN, provide the blueprint for this next era of intelligent connectivity. 

Come by the Cisco Booth #109 for live demos of AI for Wireless applications and mobility innovation for the AI era. Also, join us and industry leaders as we discuss the future of Telecom and AI in depth: 

• Tuesday, October 28 at 3pm EDT: Cisco Breakout Session on: Accelerate With Confidence: Security-First AI From Cisco and NVIDIA, and  
• Wednesday, October 29 at 10am EDT: AI-Native Wireless Networks Panel Discussion on AI-Native Wireless Networks: Shaping U.S. Leadership in 6G and Beyond  

Register now at NVIDIA GTC DC to hear the latest mobility innovations and see how agentic and physical AI plus distributed AI supply chains are reshaping telecom networks for intelligent connectivity.