The AI-Native Telco Forum in Düsseldorf this week was a good event — and an encouraging sign that telcos are starting to get ready for the next huge shift in technology: AI.
Day 1 focused on AI inside the network — smarter RANs, predictive maintenance, and energy optimisation.
Day 2, however, moved into much more exciting territory: AI at the edge and the emerging opportunity for telcos to deliver AI as a Service.
As Andrew Collinson and Dean Bubley noted in their Unthinkable Lab breakfast session, there’s a big difference between AI for the network (AI Enablers) and the network for AI (AI Services)!
🧭 From Enablers to Providers
Telcos have always been the enablers of digital transformation.
When cloud computing arrived, they provided the connectivity while hyperscalers built the platforms — and captured the profits.
Fifteen years later, history risks repeating itself.
Most operators still focus on internal efficiency, but the real opportunity lies in offering AI and Model-as-a-Service to enterprises and governments: AI that runs locally, securely, and under sovereign control.
🔄 Connecting the Dots: From Network AI to AI Services
The Dell Technologies session provided the bridge between those two worlds.
Selenga Akiner from Dell explained how telcos are already sitting on a goldmine of AI-ready infrastructure — much of it built on Intel technology that already powers their data centres, edge sites, and network clouds.
Dell’s work still sits close to the network — using AI to automate RAN, accelerate service creation, and monetise operational data — yet it points directly toward what comes next.
By exposing that intelligence through APIs and orchestration layers, telcos can transform AI that runs the network into AI that runs on the network.
Those same capabilities evolve naturally into Models-as-a-Service, where enterprises consume telco-hosted AI resources through marketplaces and open interfaces.
It’s a straight line from internal optimisation to external value.
And with partners like Intel, Dell, and others providing the compute, orchestration, and integration layers, this isn’t a distant vision — it’s technically achievable today.
🌍 AI Moves to the Edge
As Wind River reminded the audience, AI is escaping the datacentre and moving into the physical world — cars, factories, hospitals, energy grids, drones.
To make that work, intelligence has to live at the edge, where latency, privacy, and reliability matter most.
Telcos already have what’s needed: distributed mini-DCs, metro POPs, 5G edge sites, and high-capacity fibre connecting them all.
They also have local trust, regulatory oversight, and service-level discipline.
In other words, telcos are the natural hosts of edge AI.
☕ Proof in Action: COLT Edge AI
A strong example came from Chetan Narang of Colt Technologies, describing how a global coffee-shop chain uses AI for real-time video analysis to monitor queues and optimise staffing.
Each store’s cameras generate gigabytes of data every hour — far too much to ship to the cloud, and impossible under privacy constraints.
By running lightweight inference models on Colt’s edge nodes, the retailer processes footage locally and sends only anonymised insights upstream.
That’s Sovereign Edge AI in practice: data stays local, latency drops, and business value appears instantly.
🇦🇪 A Glimpse from the Middle East
Meanwhile, e& enterprise has partnered with Intel to build a Sovereign AI Platform in the UAE — a telco-grade environment that lets enterprises deploy AI models on national infrastructure while keeping data compliant and private.
Powered by Intel’s confidential-computing and AI stack, it delivers secure model hosting and low-latency performance across the country.
It’s a bold example of what European and Middle-Eastern telcos can achieve when they combine connectivity, compute, and compliance into one sovereign fabric.
☁️ From Hyperscale Cloud to Sovereign AI
In the cloud era, scale was everything.
Hyperscalers built massive, centralised datacentres to achieve economies of scale — enabling elastic, pay-as-you-use models and relentless innovation.
That worked brilliantly for generic workloads, but AI changes the equation.
AI depends on data proximity, real-time responsiveness, and sovereign control — three factors that don’t scale when everything is centralised.
Latency becomes critical, compliance rules tighten, and the cost of moving data across borders is soaring.
The pendulum is swinging back.
The future of AI will be distributed, local, and trusted — powered by smaller, smarter nodes close to where data is created and consumed.
That’s why telcos have a genuine second chance:
their networks and edge sites are already everywhere.
Combine that reach with AI-ready infrastructure and they can deliver what hyperscalers can’t — real-time, sovereign intelligence.
🔒 Why Sovereign Edge AI Matters
The demand for trusted, compliant AI is exploding.
Enterprises and governments want to know where data lives and how it’s used.
Sovereign AI provides:
- Proximity & performance – real-time responsiveness through local inference.
- Privacy & compliance – data never leaves national borders.
- Resilience – services continue even if global links fail.
- New ecosystems – APIs and marketplaces for models, data, and use cases.
This isn’t just a commercial opportunity; it’s a societal one.
AI will soon underpin healthcare, education, mobility, and public safety.
Keeping that intelligence sovereign safeguards both citizens and competitiveness.
🧱 The Building Blocks Are Here
The technology stack is ready:
- Compute & Confidential AI – Intel Xeon & Gaudi with TDX/SGX for encrypted inference.
- Edge Platforms – Wind River for deterministic, secure industrial performance.
- Orchestration Software – Red Hat OpenShift, VMware, and others for multi-site AI.
- Connectivity & APIs – MEC (Multi-Access Edge Computing) and GSMA Open Gateway to expose AI capabilities to developers.
Add partners like BT, Intel, Dell, and others, and the blueprint for a sovereign, distributed AI fabric is complete.
🔄 AI for Telecom vs. Telecom for AI
The final panel captured the real challenge: AI for telecom versus telecom for AI.
Most telcos today are focused on the first — using AI to make networks more efficient.
That’s necessary, but limited.
The second — telecom for AI — is where new revenue streams and business models will emerge: providing the connectivity, compute, and orchestration layer that every AI application depends on.
It’s also the harder challenge.
Telcos don’t build chips or powerplants, and they don’t yet have a seat at the table with the big AI players.
Those companies simply assume the network will “just be there.”
But that assumption collapses under the weight of video-driven, real-time AI, which demands massive capacity and local edge processing.
That’s why telcos must move now — not only to claim their role in the AI value chain, but to help define what responsible, sovereign AI infrastructure looks like.
This isn’t just about technology or market share; it’s about ensuring that the most transformative innovation in decades is built on foundations of trust, inclusion, and accountability.
Here, Europe has a unique responsibility, and has already started work on this as part of the IPCEI-CIS project.
Regulation, privacy, and data sovereignty aren’t constraints — they’re competitive assets that can define a global model for trustworthy AI.
If European and Middle-Eastern telcos step up, they can lead the world in delivering Sovereign Edge AI that balances innovation with integrity.
AI is too important to leave to others.
📊 What the Audience Thinks
The day closed with a live poll asking how telcos can best leverage AI to improve efficiency and develop profitable new services.
The top answer was clear — 55% of participants said “work with vendors to develop new AI capabilities.”
That’s good news. It shows that telcos increasingly recognise they can’t — and shouldn’t — do this alone.
The next era of telecom innovation will be built through partnerships — combining the reach, trust, and infrastructure of telcos with the compute and AI expertise of technology partners like Intel, Dell, and others.
It’s the strongest signal yet that the industry is ready to move — from pilots to platforms, and from internal AI efficiency to AI as a Service.
🚀 The Moment to Act
AI is the biggest disruption — and the biggest opportunity — telecom has ever faced.
Within a few years, millions of AI applications will require local inference, edge orchestration, and sovereign guarantees.
Telcos have reach, trust, and infrastructure that no one else can match.
If they act decisively — and partner smartly — they can build the AI networks of the future, delivering not just connectivity but intelligence itself.
When cloud arrived, telcos built the pipes.
When AI arrives, they can build the brain.
Let’s hope the next AI-Native Telco Forum isn’t twelve months away — this conversation needs to keep moving, fast.
Key takeaway:
The telcos who move first will define the next decade — turning their networks into sovereign AI platforms for a smarter, safer, and more connected world.