Executive Insights

AI in Telecom: Why It’s the Next Battleground for Network Survival and IoT Domination

14.07.2025 As a founding CTO and co-founder of emnify, I’ve been closely watching how artificial intelligence (AI) evolves from a buzzword into a practical tool transforming the telecommunications landscape. The impact of AI on telco is threefold: It drives network utilization through new data-hungry use cases. It integrates directly into mobile networks to optimize both the RAN and the core. It empowers enterprises to manage and optimize their IoT deployments in smarter, more automated ways. Let’s explore these in detail. Will AI use cases drive network utilization on mobile networks? Absolutely. We’re only at the early stages of what this means for mobile operators and infrastructure providers. Generative AI applications, from real-time voice synthesis to multimodal assistants combining video and text, are pushing unprecedented volumes of data across mobile networks. Even seemingly simple consumer services like AI-enhanced translation or live transcription involve constant back-and-forth communication with cloud models. Consider: Edge-based AI cameras on construction sites sending continuous video streams for anomaly detection or compliance, often over 4G/5G. Retail robots using AI vision to track inventory, uploading rich datasets in near real-time. Autonomous drones relying on AI for navigation, uploading telemetry and video back to control centers over mobile networks. These examples don’t just add a handful of megabytes; they multiply demand on both uplink and downlink. As AI moves into more latency-sensitive, real-time applications, we can expect tighter integration with mobile networks. How will AI be used inside mobile networks? AI is not only driving demand on networks; it’s also making those networks smarter and more efficient. On the RAN side: AI/ML is increasingly embedded in radio access network software to handle: Predictive traffic management, where machine learning forecasts demand spikes by hour, day, or event, adjusting resources like spectrum or power accordingly. Beamforming and massive MIMO optimization, where AI algorithms determine the optimal beam patterns in near real-time, improving both coverage and capacity. On the core side: AI helps operators: Automate slice management, critical in 5G SA (Standalone), to allocate resources to enterprise customers on-demand. Detect anomalies and prevent fraud by continuously learning from traffic patterns. Predict and pre-empt failures, reducing downtime and enhancing SLAs. Telcos can now move towards intent-based networking, where you describe the desired outcome, and AI dynamically configures the network. It’s no longer about static profiles; it’s about continuous learning and adjustment. How will enterprises use AI to optimize their IoT connectivity? This is an area close to my heart. Many enterprises today struggle with managing global fleets of connected devices, from balancing roaming profiles to ensuring consistent performance across diverse networks. AI is changing this by: Learning optimal connectivity patterns: Instead of static roaming lists or manual APN switches, AI can analyze signal quality, cost metrics, and usage to proactively select the best network. Predicting SIM churn or device issues, allowing preemptive maintenance. Automating policy enforcement, e.g., instantly throttling or isolating suspicious devices. As an example, I was recently granted a European patent (EP4401436) on using AI to optimize eUICC (embedded SIM) settings. Our approach analyzes connectivity data across millions of sessions and automatically updates eSIM profiles or network preferences on the fly over the air, maximizing uptime and minimizing cost without manual intervention. This is exactly the kind of intelligence enterprises will expect as they deploy thousands (or millions) of IoT devices globally. Closing thoughts AI is no longer just a layer on top of telco; it’s woven directly into the fabric of mobile networks. It’s driving more traffic through innovative applications, optimizing the very networks that carry this traffic, and putting powerful tools in the hands of enterprise customers to streamline their IoT connectivity. As the ecosystem matures, we’ll see networks that are not only faster and more reliable but also fundamentally more adaptive, learning, predicting, and self-optimizing in real-time. I’d love to hear your thoughts. How are you seeing AI shape your connectivity strategies, whether as a telco, enterprise or tech innovator? Add your comments to my LinkedIn post here.