AI Moves to the Network Edge with Telco Push
A major infrastructure shift is underway as AI gets embedded directly into telecom networks. The GSMA launched 'Open Telco AI' to accelerate development, while Deutsche Telekom is already offering device-agnostic AI services like live translation and call summaries directly from its network. This trend treats AI less like an app and more like a core, high-availability utility.
The GSMA's 'Open Telco AI' initiative aims to tackle the underperformance of general-purpose AI on telecom-specific tasks. Founding supporters AT&T and AMD are contributing open-source models and GPU compute capacity, respectively, to improve AI's ability to interpret network data and automate operations. The initiative is supported by over two dozen other organizations, including major operators like Orange, SK Telecom, and Vodafone. This industry-wide push is creating a significant market opportunity, with the global AI in telecommunication market valued at $3.6 billion in 2025 and projected to reach $45.1 billion by 2034. The broader Edge AI market is forecast to grow from approximately $17.88 billion in 2023 to over $118 billion by 2033, with North America currently holding the largest market share. This growth is fueled by the expansion of IoT and the demand for low-latency, real-time data processing. For platform engineering leaders, this network-level AI integration demands a strategic shift. Instead of treating AI as a siloed capability, platforms must provide structured, self-service AI agents and pre-configured coding assistants that embed organizational knowledge. This prevents "shadow AI" and ensures that as developers leverage AI, they adhere to established governance, security, and cost-management patterns from the start. The move to the edge directly impacts API strategy, as processing data closer to its origin enhances performance and reliability. For developers, this means leveraging edge computing to reduce API response times, which can directly influence business outcomes like e-commerce conversion rates. This requires a focus on API design patterns that support serverless edge architectures and local data processing to improve regional compliance and maintain functionality during central cloud outages. From a leadership perspective, this trend blurs the lines between traditional IT operations and network management. Adopting an AIOps (AI for IT Operations) model becomes critical for managing the complexity of software-defined networks. AIOps leverages machine learning to automate routine tasks, predict network issues, and analyze vast amounts of data, ultimately reducing operational expenditures and improving service quality. Major European operators are already advancing their network automation capabilities. Telefónica has 12 "Level 4" autonomous network use cases—where the network acts autonomously based on human intent—and aims to reach this level across its main operations by 2030. Deutsche Telekom is also expanding its collaboration with partners like Nokia to accelerate the development of AI-native and Open RAN solutions. This shift towards intelligent, programmable networks is enabled by standardized Network APIs, with organizations like ETSI, TM Forum, and CAMARA developing frameworks to expose network capabilities to developers. These APIs allow applications to dynamically control connectivity, a key factor for monetizing 5G and edge computing investments in sectors like manufacturing, where real-time data exchange is critical for automation. For investors, the telecommunications sector offers steady cash flows and dividend yields, though it requires significant capital expenditure for infrastructure like 5G and fiber. The integration of AI is seen as a key driver for future growth and efficiency. The convergence of telecom and AI has also spurred partnerships, such as Bharti Airtel's collaboration with Google to use AI for enhancing Rich Communication Services (RCS) messaging security.
