A conference discussion of trusted AI stacks built under national priorities, useful for seeing how sovereignty is framed at policy and infrastructure scale.
Why we selected it
Sovereign AI is often discussed as if the model were the whole system. This session puts the stack in the title and the national operating context in the frame. That makes it useful alongside the site’s six-layer map: models, compute, data, agents, edge, and physical systems do not become controllable by changing one procurement line.
What to listen for
Separate broad goals—trust, resilience, autonomy—from the controls that could be tested. Who operates the infrastructure? Which jurisdiction applies? Can the workload move? Who controls updates and keys? What continues to work if an external dependency disappears?
This is a conference perspective. Treat examples and performance claims as reported by the speakers unless they point to a method or primary record.