How to build a sovereign European AI compute infrastructure
summary
Argues for building local, sovereign compute by pairing GPU vendors with European data centers and clouds (examples: NVIDIA partnerships, Northern Data, Tiger Cloud) to offer H100 access and on-premise inferencing for GDPR-sensitive apps; runs an AI accelerator that lends hardware free for proofs of concept; cites concrete workloads like a radio-frequency checkout deployed in company canteens, multilingual translation covering 35–40 languages, and a forecasted power need rising from 17 GW to 35 GW in 48 months.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:32Why sovereign AI compute matters
- 8:46Local ecosystems and AI competitiveness
- 15:54Overcoming skepticism in AI adoption
- 23:43Investing in AI infrastructure
- 27:59AI energy and sustainability
- 30:22Closing remarks
related talks
32mSovereign AI
49mScaling Data in a Sovereign AI Platform
28mPowering Europe’s Future: Investment in AI and Quantum Infrastructure
17mCloud as the Catalyst: Bridging AI and Quantum for a Trustworthy European Future
28mKeynote - Cloud Policies for an AI Continent
13mPolitical Keynote European Commission
01mCan Europe protect digital sovereignty as AI advances exponentially?
52mHow Sovereign Solution Providers are Securing Europe’s Digital Borders?