Ahead of DTW Ignite 2026, TM Forum published new research showing CSPs are moving ahead on sovereign AI, with 77% identifying a commercial opportunity. The findings are backed in a new leadership discussion paper, Making Sovereign AI Real, which makes clear that the next phase of value will come not from infrastructure alone, but from the ability to govern and control AI in motion.
While 72% of CSPs are already investing in or exploring sovereign AI, only 25% are very confident in demand and 67% say the term is not yet universally understood. Yet 98% are adapting or exploring changes to network architecture, and 96% are aligning more closely with national or regional policy – pointing to a future where AI services are trusted, auditable, compliant and under clear control. Making Sovereign AI Real is a first, practical step toward industry clarity and faster progress from uncertainty to practical solutions.
“Sovereign AI is still being defined – and often over-simplified – while market demand and regulatory policy are moving faster than the industry’s thinking. This is not about owning GPUs or building data centres. It is about controlling how AI behaves in motion, in real time. With geopolitical context heightening urgency, the race is on for CSPs to turn network advantage into AI control,” said Nik Willetts, CEO, TM Forum.
The research shows that, according to CSPs, customers are placing less emphasis on where AI runs and far greater importance on how it behaves across complex, connected environments. This includes priorities such as control over data flows across borders, the ability to audit and evidence AI decisions, and assurance that systems are compliant and regulator ready. Yet only 43% of CSPs are already investing in the governance, audit and assurance capabilities required to make sovereign AI operational.
Where sovereignty becomes real – and where value is created.
In the discussion paper, TM Forum CEO, Nik Willetts, and Chairman, Steffen Roehn, flag this enforcement and assurance layer as the next strategic battleground for the industry. Through the paper for Members and industry, they reinforce that that this is not a challenge any single player can solve alone. Real progress will depend on industry collaboration using common standards, trusted frameworks and interoperable approaches that allow CSPs to turn network-level governance into a scalable market opportunity.
“What this whitepaper makes clear is that the opportunity for operators is not just to build sovereign AI infrastructure, but to create the governance and assurance layer around it. That is where long-term value, trust and differentiation will be won,” continued Nik.
The whitepaper sets out a shift in priorities from infrastructure ownership, localisation strategies and capacity build-out to real-time policy enforcement, audit and assurance services, and monetisation of trust and compliance. Operators that move early can position themselves not at the Edge of the AI value chain, but at its control point – using the network to deliver the governance, assurance and trust that sovereign AI will require.
Trustworthy AI, live at DTW Ignite
Sovereignty and control will be explored across DTW Ignite through the Trustworthy AI and Data Summit and related sessions on AI governance, assurance, accountability and control – including discussions on agentic AI, data products, model-as-a-service, agent-to-agent security and the practical frameworks needed to operationalize trustworthy AI at scale.
