“Location is not sovereignty. Value capture is.”
Washington and Beijing are not ending their technological competition. They are learning to manage it while continuing to scale it. AI, advanced technology, critical minerals and industrial capacity remain embedded in the strategic relationship.
For Europe, that is not reassurance. It is a deadline.
A more predictable U.S.–China relationship could allow both systems to invest faster, scale further and compound their technological advantages. The strategic question is therefore no longer how many Gigafactories Europe can build. It is whether Europe can convert capital, power and industrial demand into sufficient AI capacity to remain an autonomous economic pole.
The Systems at Play
how the race has reorganized
-
The United States combines frontier models, GPUs, hyperscalers, software and capital markets. China combines manufacturing, power, critical minerals, domestic AI, strategic capital and deployment scale. These are no longer competing technology industries — they are economic systems organized around intelligence.
Europe cannot answer them with isolated subsidies, disconnected national initiatives or a collection of data centers. It needs its own integrated architecture: capital × power × compute × industrial AI. signal: The Gigafactory programme must become an integrated architecture, not a set of projects.
-
Today's £2.45 billion VIRTUS financing, backed by a 13-bank syndicate including BNP Paribas, Crédit Agricole CIB and Société Générale, proves European institutions can finance AI-ready infrastructure at scale. The constraint is not insufficient savings. It is capital allocation.
Europe must avoid the structural loop of European savings funding foreign AI infrastructure and platforms, only to buy back the output as European AI consumption feeding foreign cash flow. The alternative flywheel keeps the return at home: European savings → European compute → industrial AI → productivity → cash flow → reinvestment. signal: The constraint is allocation, not savings — build the capital-formation flywheel.
-
Reduced geopolitical volatility improves investment visibility in both the United States and China. Lower tail risk raises AI capex, scale, ecosystem depth — and entry barriers. More compute attracts more models, which attract more customers, which generate more cash flow, which finances more compute.
Once the scale gap becomes sufficiently large, catching up becomes exponentially harder. Europe is competing against compounding, not simply against today's capacity gap. signal: Do not mistake geopolitical stabilization for a pause in technological competition.
From Leverageable MW to Strategic MW
the framework evolves
The Physical Stack Is the Strategy
where Europe already leads
The Hidden Pattern
fragmentation is the enemy
Capital without compute
Is merely savings.
Power without compute
Is merely electricity.
Compute without applications
Is merely capacity.
Applications without scale
Remain experiments.
The Velocity Insight
the missing variable
who should do what
- Investors: a more stable U.S.–China relationship could accelerate, rather than moderate, AI investment — watch Europe's relative scale and capital-formation velocity, not simply absolute CAPEX.
- CEOs & Boards: AI sovereignty increasingly means reliable control over compute, power, critical inputs and industrial applications.
- Banks: Europe has financing capacity — the question is whether European savings finance European intelligence assets.
- EU policymakers: stop measuring success principally through Gigafactory MW; measure European economic value captured per MW.
- European governments: energy, AI, semiconductors, infrastructure and capital markets are no longer separate agendas — they are one industrial strategy.
- The new metric: move beyond Leverageable MW toward Strategic MW — financeable AI capacity that also generates European technology, IP, applications and cash flow.