Panel, Conference Presentation
Sovereign AI: Piece by Piece | NetApp, Radiant, Hydra Host, Coatue & More | RAISE Summit 2026
RAISE SummitGeorge Kurian, Mahdi Yahya, Aaron Ginn, Max Cook, Benjamin Germin, Thomas Koehrer, Thomas Kova, Maddy, Ben
- Sovereign AI is projected to evolve into a fundamental transformation driven by governments taking proactive policies to reduce foreign dependencies, shifting focus from model layers to tangible infrastructure including chips, data centers, and strict data residency, a trend described as durable due to the GPU technology shift being inherently "anti-cloud."
- Control plane sovereignty is identified as the most critical component of the stack, with nations expected to navigate a continuous spectrum of ownership and regulation rather than achieving a binary end state, while the "AI industrial cloud" project in Germany specifically aims to increase national compute capacity by 50%.
- Infrastructure build-outs face significant bottlenecks related to power availability, particularly gas turbines, and the complexity of global chip supply chains, leading to predictions that the US will maintain full-stack dominance for at least five years while China develops an alternative ecosystem with a coalition of nations and open-source models.
- Market dynamics are expected to resemble the airline industry with two core global pillars and localized winners, driven by "inelastic" demand where governments prioritize strategic value over cost, resulting in deflationary token prices as deployment scales and increased spending to avoid the "cost of not having it."
- Regional strategies vary significantly, with Latin America and South America leveraging comparative costs, France, the UK, and Canada utilizing power and policy advantages, and Asian nations driven heavily by export controls, though Europe is predicted to retain heavy dependency on foreign hyperscalers and chips within the next five years.
- The regulatory landscape will treat AI as critical infrastructure and a utility similar to gas or electricity, with governments acting as a third, exogenous market player enforcing strict controls on the "fastest ever" developing technology, where current restrictive postures are anticipated to be permanent and self-regulation will emerge alongside government mandates for powerful models.
- Economic constraints and export controls are predicted to foster innovation and create lower-cost models, potentially accelerating the innovation cycle through necessary duplication of investments, while the emergence of "AI native" expressions is contingent upon the prior establishment of foundational ecosystem elements like chips, data, and capital.