Telcos pivot from AI operational efficiency to infrastructure monetization strategies

This shift indicates that the 'telco as a pipe' model is being challenged by the 'network as an AI control point.' By utilizing existing distributed facilities for GPU-as-a-Service and sovereign AI, operators can capture value from the $2.5 trillion AI spending forecast. For the streaming ecosystem, this move toward localized, intelligent compute promises lower latency for real-time video processing and more dynamic traffic management as agentic AI increases network symmetry. Watch for the emergence of multi-agentic framework benchmarks as TM Forum moves closer to defining Level 5 'fully autonomous' networking standards.

The transition to AI-centric networking is accelerating as operators report tangible pilot results. Per a March 2026 report from AT&T, the company launched its 'Connected AI for Manufacturing' platform at MWC 2026, which combines 5G, Nvidia GPUs, and Microsoft Azure to deliver machine diagnostics at the edge. Early results from these pilots showed a 70% reduction in waste and identified potential hardware failures up to four hours before they occurred. These successes are shifting executive focus toward end-to-end automation, with 60% of operators now prioritizing AI capabilities in their infrastructure procurement cycles, according to Omdia data from February 2026.

Financial markets are beginning to price in this infrastructure pivot. Cisco's stock reached record highs in June 2026 following CEO Chuck Robbins' keynote at Cisco Live, where he emphasized that 'the network is more important than the node' in an era of agentic AI. Per Dell'Oro Group, global data center capex is expected to surpass $1 trillion in 2026, driven by a 78% spending increase from major cloud providers. This massive investment creates a halo effect for telcos, as enterprises seek hybrid architectures that combine massive hyperscale training with localized, sovereign inference to comply with data residency laws.

Globally, the gap between AI experimentation and industrial-scale execution is narrowing. Per Swisscom's full-year results in early 2026, the operator has successfully launched a sovereign AI platform designed specifically for the Italian and Swiss public sectors. Simultaneously, China Mobile has developed over 50 industrial large models under its 'Jiutian' brand, which use over 20 trillion data tokens to automate cross-domain network functions. Industry analysts at IDC noted in May 2026 that the 2026-2028 window will be defined by this 'execution discipline,' as operators move from internal opex reduction to launching SLA-backed, AI-orchestrated services for enterprise clients.