Mobile World Congress (MWC) Barcelona 2026 made one thing abundantly clear: artificial intelligence (AI) monetisation is accelerating, but most of the value is still flowing away from telecom operators. Over the next 12 months, the critical question for operators is less about how to deploy AI, and more about how to capture the economics. This marks an important inflection point where the industry must pivot from technological hype to pragmatic revenue generation.
Omdia has consistently argued that AI creates economic value only when it is embedded into operational workflows with clear ownership, measurable outcomes, and budget accountability. MWC 2026 reinforced this view. The conference's discussions moved away from potential and possibilities toward pragmatic matters such as latency, failure tolerance, governance, and data sovereignty. This reflects a broader enterprise reality: AI that sits outside core workflows remains discretionary spend. Conversely, AI that directly improves uptime, throughput, yield, or safety competes much more effectively for budgets owned by CIOs, chief operating officers, and line-of-business leaders.
The implication is straightforward: revenue follows deployment, and deployment follows outcomes. The most credible AI narratives at MWC were those demonstrating operational impact rather than technological novelty. For example, in manufacturing, AI-powered quality inspection systems reduce defect rates by up to 30%, directly improving yield. In logistics, AI-driven route optimisation cuts fuel costs and delivery times. These are the kinds of outcomes enterprises are willing to pay for.
The Vertical Battleground: The Operational AI Stack
What became increasingly clear in Barcelona is that enterprise AI is no longer being deployed as a standalone capability. Rather, it combines digital twins, edge AI, and dedicated connectivity to form an operational AI stack. It is one system, not three separate projects. From Singapore to Shanghai, and Beijing to Busan, ports and factories are already running on these AI stacks. These are not proofs of concept; they are real-life operations in commercial use, where digital twins are increasingly used to simulate and optimise physical processes in near-real time.
Digital twins allow operators to model entire production lines or port operations, testing changes virtually before applying them physically. This reduces downtime and accelerates process improvements. Edge AI handles time-critical inference for quality inspection, autonomous vehicles, and robotic coordination. Without edge processing, latency from cloud-based AI would make real-time decisions impossible. Meanwhile, private or campus networks provide deterministic latency, security, and availability that public networks cannot guarantee. These elements are interdependent, and enterprises do not purchase them piecemeal.
This interdependency explains why vertical AI is proving monetisable. Enterprises are not buying AI per se. They are buying higher asset utilisation, faster throughput, and safer factories. Crucially, this AI stack requires telecom capabilities – latency, resilience, and regulated data paths – which gives operators pricing power when bundled as managed outcomes. For instance, a managed service that combines private 5G, edge compute, and an AI quality inspection application can be sold as a per-machine subscription, aligning costs directly with the value delivered.
The Road to Revenue
For telecom operators, this creates a narrow but credible path to AI revenue. Operators tend to struggle when competing with hyperscalers on foundation models or generic platforms. Instead, value capture lies at the intersection of AI, connectivity, and industry operations. Orange CEO Christel Heydemann articulated this well at MWC. She argued that operators must evolve from connectivity providers to architects of trust, turning resilience, cyber security, and data protection into competitive advantages. At the same time, she acknowledged the economic imbalance facing operators, noting that despite operating critical digital infrastructure, telcos are not reaping rewards proportionate to their role.
Trust and performance are not abstract concepts when placed in the context of an operational workflow within enterprise verticals. When AI stacks control physical processes, operators can price predictability, accountability, and compliance. Multi-year managed services around private 5G, edge compute, and secure data pipelines can then be bundled and contracted out to enterprises. This approach moves telcos away from commoditised bandwidth sales toward high-value outcome-based contracts. For example, a port operator might pay for guaranteed container throughput improvements rather than a flat network fee.
APAC Shows What Execution Looks Like
Asian operators at the event provided a blueprint for how to monetise AI. SK Telecom CEO Jung Jaihun stated that operators' proprietary infrastructure and operational expertise are key to building AI infrastructure, adding that telcos must move beyond data delivery to play a leading role in shaping AI services for enterprises. Singtel made a similar argument, but with a strong emphasis on outcomes. Announcing its expanded 5G Advanced strategy at MWC, Singtel Singapore CEO Ng Tian Chong said: 'We see 5G Advanced not as a network upgrade, but as a foundational platform for Singapore's AI-powered future. By embedding intelligence, programmability and service differentiation into our network, we enable enterprises to innovate with confidence and compete globally.'
This framing from SK Telecom and Singtel is significant. They are not selling AI as a standalone product or an overlay; rather, AI is inseparable from infrastructure and industry services. This integrated stack combines private networks, edge compute, and service-level agreements (SLAs). In sectors such as manufacturing and logistics, this bundling approach aligns closely with how enterprises buy technology. For instance, a factory might purchase a 'smart production line' package that includes sensors, edge servers, AI analytics, and guaranteed network performance, all under a single SLA. Such packages reduce procurement complexity for enterprises and create recurring revenue streams for telcos.
Why Execution Will Decide AI Winners
Many operators understand where AI value is forming, but few are organisationally prepared to capitalise on it. Vertical AI requires a different path to market, new ecosystem partnerships, and a willingness to move away from selling bandwidth to selling outcomes. This means telcos must invest in industry vertical expertise, hire solution architects who understand manufacturing or logistics, and build go-to-market teams that can speak the language of chief operating officers rather than only CTOs. It also requires rethinking internal incentive structures: sales teams traditionally rewarded for selling gigabytes need new metrics tied to outcome delivery and customer retention.
Partnerships are equally critical. Telcos cannot build all components of the operational AI stack alone. They need partnerships with cloud providers for digital twin platforms, with AI software vendors for inference models, and with system integrators for deployment and support. Successful operators will act as orchestrators, bringing together these partners while owning the connectivity and trust layer. The battleground for operators is clear: it is not in the large language model (LLM) or raw compute power, but in the enterprise vertical. Operators that align AI with operational outcomes, trust, and performance will find real revenue. Those that do not will continue to enable AI growth, but without getting their fair share of the rewards.
Source: ComputerWeekly.com News