The telecommunications industry is facing a stark economic reality. Monthly global mobile network data traffic has surpassed 200 exabytes, yet telecommunications revenue is projected to grow at a compound annual growth rate (CAGR) of just 1.3% from 2025 to 2030. This widening gap between data volume and financial return is prompting operators to reassess how they create and capture value. The old playbook of selling connectivity by the gigabyte is under increasing pressure. The new imperative is to move beyond transporting data and begin turning network intelligence into governed, commercially viable services.
This is the core premise of the AI Opco model. AI Opco refers to an operating model in which artificial intelligence is treated as a governed, production-grade capability across the telecom business, not simply as a software feature or customer-service chatbot.
The network can evolve from a passive pipe into an intelligent, programmable fabric that supports inference at the edge, assists with autonomous decisions, and turns selected network signals into potentially monetizable data products.
From Connectivity Provider to Decision-Making Engine
One of the biggest shifts telecom companies are facing is in identity. A traditional telco asks, “How reliably can I deliver a packet?”
An AI Opco asks, “How quickly can I support a decision?” For many enterprise use cases, the network’s value lies not only in carrying traffic but also in providing the context, controls, and intelligence needed to support business decisions.
This value comes not only from carrying internet traffic but also from providing the data and decision services required by AI agents, autonomous vehicles, and industrial control systems.
This demands the coordinated evolution of cloud, edge, and core network architectures into an interoperable operating environment that supports AI at scale while respecting regional data residency laws.
Such an architectural shift enables a move from reactive operations to a more predictive and increasingly automated stance. Modern networks are already using real-time analytics for anomaly detection and automated provisioning. A self-optimizing network may identify a potential equipment failure and reroute traffic before service quality is materially affected. But this internal efficiency is just the starting point. The real prize is to expose selected capabilities externally in forms that solve measurable enterprise problems. An operator that can autonomously manage its own network may be able to offer related capabilities to an enterprise partner managing a global logistics fleet or a chain of automated factories. These offerings would need clearly defined APIs, security controls, service levels, and commercial models, as internal automation does not automatically translate into a market-ready product.
It’s important to avoid layering AI over fragmented, siloed data. The operators making real progress have focused on improving data consistency, ownership, quality, and access across domains. A ‘single source of truth’ is often better understood as a set of governed, interoperable data sources with clearly defined semantics and lineage. By standardizing data patterns, operators can reduce deployment timelines. This speed matters. In a market where the value of data can decay rapidly, the ability to move fast while maintaining data quality and controls is a competitive advantage.
Building the Governed Data Plane
The technical heart of an AI Opco is a robust data control plane. This layer should connect radio access network telemetry with business support systems and identity signals, ensuring data remains governed and trustworthy from ingestion to inference. Using an open table format like Apache Iceberg can support high-performance data lake architectures and more flexible data management, but the technology alone does not guarantee lower storage costs or effective governance. More importantly, this infrastructure must enforce strict lineage and access controls. Every data product delivered to a partner or internal AI model must meet compliance and privacy standards. Without this trust foundation, enterprise adoption and long-term monetization become difficult to sustain.
This transition also demands a new financial model. The move away from simple gigabyte-based pricing toward more value- and usage-aligned commercial models is essential. Billing could be tied to API calls, inference events, verified transactions, service tiers, or dataset subscriptions, depending on the product and buyer. This aligns the operator’s revenue more closely with the value it provides. If a developer uses a location verification API to prevent a fraudulent transaction, the operator could charge per request, through a subscription, or under a value-based commercial arrangement, subject to the API’s accuracy, coverage, and contractual terms.
Scaling Through Ecosystems and APIs
No operator can build the future alone. Achieving platform scale requires a decisive move away from bespoke, one-off integrations and toward standardized API frameworks. The CAMARA project and GSMA Open Gateway initiative provide a common direction for exposing selected network capabilities through standardized APIs, although adoption and availability vary by operator and market. These APIs can allow third-party applications to request, verify, or configure selected network capabilities. By exposing “network truths” through a simple, well-documented interface, telcos can become more directly integrated into the application development lifecycle. They are no longer just the carrier of the final product; they can provide infrastructure capabilities that form part of the product’s value proposition.
Privacy-first collaboration is non-negotiable in the current regulatory environment. Technologies like data clean rooms allow operators to partner with banks, retailers, and governments while limiting the sharing of raw personally identifiable information, provided the environment, data policies, and permitted analyses are appropriately designed. A telco can use aggregated or otherwise privacy-preserving insights to help a retailer analyze foot traffic patterns for a new store location, and the retailer may not need access to individual subscriber records. This capability unlocks cross-industry monetization while supporting data sovereignty. Analysis can take place where the data resides or within a controlled collaboration environment, reducing certain data-sharing risks and supporting compliance efforts.
Naturally, the distribution model is also evolving. Marketplace models allow operators to package data products and native applications for easy discovery by a global developer audience, lowering the barrier to entry for third-party innovation and supporting ecosystem development around the core infrastructure.
The AI Factory and the Telecom Knowledge Plane
To generate intelligence consistently, an operator must establish an internal “AI Factory”: an operational framework that handles feature engineering, model management, and monitoring within a secure governance boundary. It enables repeatable intelligence production and controlled deployment, ensuring automated decisions can be evaluated against current network telemetry. Moreover, it moves the organization from one-off experiments to a state of more systematic optimization across the network footprint. The factory mindset treats AI development not as a series of projects, but as a core, repeatable business process.
This internal capability can be complemented by what can be called a telecom knowledge plane: a governed layer of network metadata, ontologies, policies, and operational context that helps AI systems interpret the environment in which they operate. This plane enables AI systems to understand the network’s physical and logical nuances. An AI agent managing an Internet of Things (IoT) fleet, for instance, could query this layer for information about suitable edge locations for an inference workload based on current network congestion, latency, capacity, and power availability. The quality of that recommendation would depend on data freshness, coverage, model accuracy, and the operator’s ability to enforce the resulting policies. This level of contextual awareness can help make the AI Opco more valuable to the next generation of autonomous systems.
The final goal is to turn these internal analytics into external value. Insights derived from the knowledge plane (such as device risk indicators, IoT fleet health metrics, and predictive maintenance alerts) can be packaged as products when they address a defined customer need and meet applicable privacy, accuracy, and liability requirements.
Navigating the Hard Tradeoffs
The path to becoming an AI Opco is not without friction. There’s a significant upfront investment required, covering not just technology but also the fundamental reskilling of the workforce. Engineers who once optimized for network uptime must learn to think about data pipelines and model governance. Product, sales, legal, security, and operations teams must also be able to translate technical capabilities into reliable enterprise offerings. This cultural transformation is often the most difficult part of the transition.
There is also the tension between speed and governance. The market rewards operators who can move fast, but the regulatory environment demands meticulous control over data. Cutting corners on governance to accelerate a product launch can create significant long-term operational, legal, and reputational risks.
Furthermore, the promised returns from AI-driven services are not guaranteed. While the potential for higher-value data products is real, capturing that value requires sustained execution, including reliable data, differentiated distribution, measurable customer outcomes, and repeatable operations. Building a developer ecosystem can take years of consistent investment.
In Closing
The telecommunications industry has reached an inflection point. The operators that will thrive are those that recognize that their core asset is not only the physical network itself, but also the intelligence and programmable capabilities that flow through it. The AI Opco model provides a strategic framework for this transformation, prioritizing data governance, programmable interfaces, and ecosystem collaboration.
This is not a simple technology upgrade, but rather a fundamental redefinition of what a telecommunications company is and does. The leaders in this space are those who will adopt consumption-based and other value-aligned economics, build robust data control planes, and foster developer-centric environments, while tying these capabilities to specific enterprise outcomes. The future may belong not simply to the largest networks, but to the operators that can turn network intelligence into trusted, useful, and commercially sustainable services.
