Episode SummaryIn this episode of Carrier 2.0, host Steve Saunders argues that the real competitive advantage in the AI era isn’t hype, scale, or autonomy, it’s execution.
As hyperscalers dominate headlines, carriers face a more grounded challenge: how to modernise live networks without breaking trust. Drawing on conversations with operators and technology leaders, the episode explores why there is no “greenfield reset” in telecom, only the messy reality of legacy infrastructure, exponential endpoint growth, and software complexity layered over physical networks.
From deterministic APIs to probabilistic LLMs, from cloud-native ambitions to the reality that only a fraction of network functions are truly cloud-native, the discussion exposes the operational gap between vision and delivery. The episode makes a clear case: AI changes the network, but uptime, throughput, integration, and disciplined prioritisation still define success.
Ultimately, the end state for carriers isn’t full autonomy, it’s trust. And trust requires humans in the loop, clean data foundations, horizontal integration, and infrastructure delivered at software speed.
Key Talking PointsThe Execution Gap (00:00)
Why reinventing networks with software fails when execution collapses under complexity.
The Greenfield Myth (01:44)
Why carriers don’t get to start over, and must modernise live networks with legacy systems intact.
Changing the Propeller Mid-Flight (02:21)
The operational reality of transitioning to software-defined architectures while maintaining service continuity.
From Silos to Horizontal Platforms (02:32)
Why 21st-century network ecosystems demand integration over organisational segregation.
Trust Over Hype (03:42)
Why enterprises ultimately buy competence, reliability, and realistic delivery, not AI marketing.
AI Isn’t New, But the Interface Is (05:00)
From machine learning to LLMs: what has actually changed, and what hasn’t.
Uptime Is the Real AI Constraint (05:24)
Why stability, fibre throughput, and predictable performance matter more than hallucinations.
Data First, AI Second (06:38)
Why clean databases and narrow, high-impact use cases are the true starting point of AI transformation.
Deterministic vs Probabilistic Systems (07:43)
The clash between contract-based APIs and semantic AI agents, and why telecom must reconcile both.
Cloud Native Reality Check (09:17)
Why most networks remain far from fully cloud-native, despite years of readiness claims.
Autonomy vs Trust (09:38)
Why complete network autonomy is not the goal, and why humans must remain in the loop.
Infrastructure at Software Speed (10:17)
Why the winners will deliver network infrastructure with the speed, flexibility, and user experience of cloud software.
The Carrier QuestionIf there is no forklift upgrade and no greenfield reset, what does disciplined execution actually look like in a live, AI-enabled network?
For this episode, the answer lies in modernising without breaking trust, building horizontal platforms instead of silos, prioritising data clarity, using AI selectively, and delivering infrastructure at software speed while keeping humans firmly in control.
LinksJoin Steve Saunders’ mailing list for bonus insights