Intelligence Is No Longer Just a Digital Tool
The conversation around artificial intelligence is no longer limited to tools, automation, or smarter digital services. At the panel discussion “From AI to AGI: Should Nations Build Intelligence as Critical Infrastructure?”, the debate moved into a much bigger space: whether intelligence itself should now be treated as part of a nation’s core infrastructure.
The session was moderated by Nate Busa, Director – AI, Automation at NEOM, with insights from panelists Manus Ward, Director of Operations for the National Transformation Institute at KAUST, Adel Miran, Director General of Administration & Facilities Management at Dubai Holding, Dr. Ibraheem Sheerah, Chief Transformation Officer at Saudi Arabian Airlines Holding, and Andrew Chen, Vice President of Platform at MAGNA AI. The discussion brought together questions that governments can no longer afford to keep in the background.
The Big Question: What Should Governments Own?
For years, governments have looked at AI as something to adopt. A tool for improving services, automating operations, supporting decision-making, and increasing productivity. That view still matters, but it may no longer be enough.
As AI systems become more powerful, nations are starting to ask which part of the intelligence layer they should actually own. Is it compute? Could it be data? Or does the real power sit inside the models? Or is orchestration the real strategic layer, the ability to connect systems, govern access, and control how intelligence is deployed across public services, industries, and national operations?
This question is not only technical. It is political, economic, and strategic. A nation that depends completely on external AI systems may be able to move quickly at first, but that dependency can become a weakness during global outages, supply chain pressure, cloud disruptions, or geopolitical uncertainty.
Compute Is Becoming a National Asset
Compute was one of the central issues in the discussion. Without enough compute, advanced AI systems cannot be trained, deployed, or scaled. And as the world moves from AI toward AGI-level ambition, compute becomes more than a backend resource. It becomes a national asset.
Countries that want to build sovereign intelligence capabilities will need reliable access to high-performance infrastructure, not just software partnerships or strategy documents. If intelligence becomes deeply embedded into government, transport, aviation, healthcare, energy, and security systems, then losing access to that intelligence layer is no longer just an IT problem. It becomes a national resilience problem.
Data Centers Are Now Part of AI Strategy
That brings data centers into the spotlight. Hyperscale data centers are no longer invisible technical facilities sitting behind digital services. They are becoming part of the foundation of national AI readiness.
AGI-level systems will demand massive energy supply, advanced cooling, resilient cloud architecture, and long-term infrastructure planning. The challenge is not glamorous, but it is decisive. A country can announce an AI strategy, launch innovation programs, and attract talent, but without power, cooling, and compute capacity, the strategy stays fragile.
Energy and Cooling Could Shape AGI Readiness
Energy and cooling may become some of the most important constraints in the race toward advanced intelligence. The panel raised the question of how hyperscale data centers can meet AGI-level cooling and energy demands, and it is a question every AI-driven economy will eventually face.
The next generation of intelligence systems will not run on ambition alone. They will require physical infrastructure that can operate continuously, efficiently, and securely at enormous scale. In that sense, AI infrastructure planning is no longer separate from energy planning. They are starting to overlap.
Public-Private Cloud Partnerships Matter
The discussion also touched on the role of public-private cloud partnerships in building national intelligence resilience. Very few governments can build every layer of the AI stack alone. Private cloud providers and AI companies bring scale, technical expertise, and operational maturity.
But relying fully on private infrastructure also creates risks. The real challenge is building partnerships where governments can benefit from private-sector speed while still protecting national continuity, sovereignty, and control during disruption.
Resilience Cannot Be an Afterthought
If a country’s public services or critical industries depend on AI systems, then those systems must remain available during outages, cyber incidents, regional disruptions, or global platform failures. Public-private partnerships may help, but they need strong governance, redundancy, local capacity, and clear rules on who controls what when things go wrong.
This is where intelligence infrastructure becomes a serious policy issue. It cannot be treated like a normal technology contract. If intelligence becomes infrastructure, then downtime becomes more than a service interruption. It becomes a continuity risk.
Decentralized Compute Enters the Conversation
Another major idea from the panel was decentralized compute. The question was whether decentralized compute grids could help nations move toward autonomous, sovereign AGI. It is a bold concept, but not unrealistic.
Instead of depending only on centralized hyperscale data centers, nations could build distributed compute networks across government institutions, private partners, universities, cloud facilities, and regional infrastructure. This could reduce single points of failure and make national AI systems more flexible.
Of course, decentralized compute comes with its own complications. Security, coordination, trust, governance, standards, and performance all become harder when infrastructure is distributed. But the idea is gaining relevance because the future of national intelligence may not depend on one massive system. It may depend on a connected grid of compute, data, models, and orchestration layers working together across different environments.
AI Readiness Is Not the Same as AGI Readiness
The strongest message from the panel was that AI readiness and AGI readiness are not the same thing. AI readiness can mean adopting tools, launching pilots, training staff, and improving workflows.
AGI readiness asks something much larger. Can a nation run advanced intelligence systems at scale? It must also protect its data, guarantee compute access, and handle global outages. Beyond that, the country needs enough power and cooling for the required infrastructure. Finally, it must manage the relationship between public control and private capability.
These questions are no longer distant future topics. They are becoming present-day planning issues for governments, technology leaders, and infrastructure builders. The panel made it clear that nations hoping to lead in the next phase of artificial intelligence cannot treat intelligence as just another digital service. They may need to treat it as infrastructure, and once that happens, the AI race becomes a very different kind of race.
