Building Sovereign AI: Lessons from Deploying Secure RAG Systems using Confidential Computing
-
Isha Salania
Microsoft
Cloud & AI Solution Engineer
Digital sovereignty is a hot topic not only for governments but also for organisations that operate across jurisdictional and national boundaries - and therefore all major financial businesses. Restricting the ability of foreign-owned or foreign-controlled organisations OR GOVERNMENTS to access data owned by, about or important to national citizens or interests and reducing dependence on foreign-owned or -controlled infrastructure is complex, but becoming a firm requirement from regulators and lawmakers.
Most approaches attempt to move computing resources - often to less-appropriate or favourable locations - or to rely on contractual agreements.
This session introduces an alternative: using Confidential Computing - a set of widely-available, hardware-based technologies - to isolate applications and their data from the underlying infrastructure, operators and owners, implementing a truly isolated zero-trust friendly architecture. We will discuss the basics of the technology and then examine how well it fits the needs of digital sovereignty, explaining its applicability for use cases from finance, AI, Web3 and the broader Enterprise.
World Congress 2026 Europe
Isha Salania
Microsoft
Cloud & AI Solution Engineer
Franz Kiraly
German Center for Open Source AI
Director
Ross Kukulinski
Tailscale
VP of Product
Jan Peer Stöcklmair
Sentry
Senior Software Engineer