Since 2024, five themes have shaped how we think about AI.
How our thinking has moved
Five questions, followed year after year. Start with 2027 and travel back to where it began.
Power and infrastructure
Sovereignty in AI now means owning the chips, the energy and the rules, not just the models.
We follow who has it, and who is trying to get it.
2027Citizens of an expanding world
AI is widening what people can know, make and decide. In 2027 we ask who gets to shape that bigger world.
AI as global power
In 2027, the question is who holds power over AI, from middle powers to orbit.
Coming soon
More about Davos 20272026A human intelligence shift
Agents reached everyday work and nations raced to build their own AI. We asked what stays distinctly human.
AI as public infrastructure
We asked whether AI should be run like public infrastructure, and who it should serve.
Should countries build their own AI, or build it together?
- Sovereign AI: Securing Autonomy in the Age of IntelligenceRoundtable, not recorded
- The Delicate Act of Balancing AI Sovereignty With Global AI InteroperabilityRoundtable, not recorded
- The Next Wave of AI Models: Small, Specialized, SovereignRoundtable, not recorded
2025After the hype
AI had arrived. The question moved from promise to proof: where does it create real value?
AI as a national priority
We asked why owning AI had become a matter of national security, and what it takes to build it.
Why has owning your own AI become a matter of national security?
2024Inaugural edition
The first AI House in Davos, as AI went mainstream. We asked where the technology really stood.
AI as a policy question
We asked who should write the rules for AI, and whether Europe could build its own.
Can Europe build its own AI, or will it depend on others?
