Executive Summary

This report addresses a key strategic conflict for telecommunications operators (“telcos”): the necessity of partnering with hyperscalers versus the profound long-term risk of ceding control. We look now at the situation in the network edge, just as our earlier report looked at core networks (see “Telco-Cloud Strategies – Balancing Risks and Opportunities in Core Network Outsourcing”). 

Partnerships are essential to gain operational agility, to access world-class Generative AI (GenAI) platforms, and to transition from a capital-intensive to an operational expenditure (“OpEx”) telco business model. However, these partnerships create an existential threat of customer capture, value-chain disintermediation, and relegation to a “dumb pipe” connectivity provider. 

This risk is generated by potential lock-in to hyperscaler ecosystems. We present a framework to evaluate these partnerships. The GenAI edge stack is resolved into four “control planes” (Infrastructure, Platform, Application, and Data). This provides a framework for auditing hyperscaler lock-in mechanisms across each layer.

Key Findings:

1. The New Lock-In: The most significant lock-in mechanism is no longer just proprietary APIs. It is the hyperscalers’ proprietary management plane ( e.g. Azure Arc, Google Anthos) that extends hyperscaler control over the telco’s own on-premise and edge assets, creating deep, structural dependency.
2. The “Reseller Trap”: Simple reseller agreements ( e.g. reselling Microsoft Copilot) represent a critical form of customer disintermediation. In this model, the telco uses its own trusted sales force to transfer its enterprise customer relationships directly to its biggest competitor.
3. The “Sovereign AI” Advantage: The telco’s greatest point of leverage is its status as a trusted, national infrastructure provider. This unique position enables the telco to build “Sovereign AI Factories,” a high-value service that US-based hyperscalers, subject to foreign laws like the CLOUD Act, are legally and operationally compromised in delivering.

Telcos cannot replicate what hyperscalers deliver, and need partnerships. The optimal strategy is not isolation but “agile autonomy.” The only long-term defence against lock-in with one hyperscaler is a credible, demonstrated ability to use another provider. This involves the telco building and owning a federated, multi-cloud orchestration platform based on open-source standards ( e.g. Kubernetes, StarlingX, Sylva). This platform allows the telco to treat all providers – its own private cloud, AWS, Azure, and Google – as interchangeable component suppliers, thereby retaining control, avoiding dependency, and preserving the all-important customer relationship. 

This means different things for Tier 1 and Tier 2 & 3 telcos. For Tier 1, it means leaning into sovereign AI infrastructure. For Tier 2 & 3 that lack the capital and technical resources for the sovereign strategy, it means forming alliances with other telcos and rigorously sticking to open alliance standards.