Programmable sovereignty: coordinated specialisation as a response to the N-2 problem in artificial intelligence

中文标题

可编程主权：协调专业化作为人工智能中N-2问题的应对之策

Natixis09/07/2026updated 10/17/2026

Multiple economies

Citable research brief

Core conclusion · Tlines extraction

Natixis introduces 'programmable sovereignty' and 'coordinated specialisation' as responses to the N-2 problem in AI, where only the US and China achieve full-stack sovereignty. The report argues that N-2 states should pool governance and hard-stack capacities, based on scorecards of seven economies, to make dependence reversible and credible.

Institution: Natixis
Published: 09/07/2026
Research topic: Coordinated specialisation · AI governance · Market leverage
Time horizon: medium-term

Key arguments

The N-2 problem: only the US and China can pursue full-stack AI sovereignty, leaving all other states dependent on one or both.
Programmable sovereignty: soft governance anchored by minimal hard-stack control, especially pooled compute, to make dependence reversible and governance credible.
Coordinated specialisation: pooling comparative advantages across complementary N-2 states under a shared governance framework, inspired by Airbus.
The EU, India, Japan, and South Korea possess governance capacity, while Taiwan and South Korea contribute fabrication and memory; no single N-2 state can achieve programmable sovereignty alone.
Regulatory actions by the US and China, such as rare earth controls and the Anthropic episode, demonstrate how dependence is leveraged.

Key numbers

Rare earth refining share (China): ~99% gallium, ~83% germanium
US AI compute share: ~75% of benchmarked GPU-cluster performance
China AI compute share: ~14%
US private AI investment 2024: $109.1 billion
China private AI investment 2024: $9.3 billion
EU market size for AI deployment: $20 trillion
Japan photoresist global supply share: ~90%
Top 5 equipment vendors revenue share: 56%-66%

Main risks

Coordinated industrial strategies have a mixed record; coalition may underperform due to state-capacity constraints, weak enforcement, and rent-seeking.
Coalition stability is uncertain; members with matured capabilities may defect and free-ride.
Political feasibility varies; Taiwan's security ties to the US may hinder formal membership.
The window for building alternatives is closing as supply chains and norms settle.

Corresponding source evidence

we argue that coordinated specialisation is the only feasible option available to the N-2 states. · medium-term
When it comes to attracting others to the partnership, the EU offers a $20 trillion market for AI deployment. · medium-term

Conditions / invalidation

If states adopt the proposed architecture and governance first.
If the EU offers market access as leverage.

Context: the conclusion is Tlines' automated extraction, not the institution's wording. Quoted evidence is publisher wording. Scope and date must travel with the conclusion.