The United States is abandoning the idea that countries need to build every layer of artificial intelligence technology at home, instead pushing allies to specialize and integrate into a shared Western tech ecosystem. This represents a fundamental shift in how Washington views technological competition with China and shapes relationships with partners like Italy, Germany, and the Netherlands.

What Is 'Innovation Sovereignty' and Why Does It Matter?

Jacob Helberg, the U.S. Under Secretary of State for Economic Affairs and architect of the Pax Silica initiative, recently challenged the United Nations' vision of "digital sovereignty," arguing that countries trying to replicate entire AI technology stacks domestically waste resources on yesterday's breakthroughs instead of investing in tomorrow's innovations. His alternative concept, "innovation sovereignty," reframes technological power not as the ability to reproduce existing technologies, but as the capacity to generate new ones through research, entrepreneurship, and industrial learning.

The distinction carries enormous practical weight. Rather than encouraging dozens of countries to build competing versions of the same AI infrastructure, the U.S. is now promoting a model where each ally contributes unique strengths to a shared ecosystem. One country might provide semiconductor manufacturing, another critical minerals, another energy infrastructure, and another frontier AI research.

How Is the U.S. Building This New Tech Alliance?

The centerpiece of this strategy is Pax Silica, a U.S.-led coalition that expanded to 24 participating countries during a summit in Washington this week. The initiative brings together trusted partners across the entire AI supply chain, from raw materials to advanced manufacturing to research institutions. Italy recently signed the Joint Statement on AI Opportunity at the summit, formally aligning itself with this vision of coordinated technological development.

This approach marks a deliberate departure from the assumption that every critical supply chain segment must be recreated inside the United States. Instead, Washington is emphasizing what Helberg calls "trusted interdependence," where allies specialize according to their comparative advantages while maintaining control over strategically sensitive technologies.

Steps Nations Can Take to Participate in Trusted Tech Ecosystems

• Identify Comparative Strengths: Countries should assess where they have unique advantages in the AI supply chain, whether in semiconductor manufacturing, critical mineral extraction, renewable energy infrastructure, or frontier research capabilities, rather than attempting to replicate every technology domestically.
• Build Bilateral Trusted Technology Dialogues: Nations like Italy are establishing formal coordination mechanisms with the United States and other allies to align industrial policy, research priorities, and supply chain resilience around shared technological objectives.
• Integrate Into Multilateral Frameworks: Participating in initiatives like Pax Silica allows countries to shape the emerging Western technology architecture while securing access to cutting-edge AI models and infrastructure that would be prohibitively expensive to develop independently.

What Triggered This Shift in U.S. AI Strategy?

The timing of this policy pivot is not coincidental. Just days before the Pax Silica summit, the Trump administration invoked export controls to effectively ban access to Anthropic's "Mythos class" AI models, including the consumer version Claude Fable 5, to foreign nationals both inside and outside the United States. The proximate cause remains disputed: some observers point to Anthropic's sharing of powerful model versions with a South Korean telecommunications company without White House approval, while others cite the company's attempt to restrict Pentagon use of its Claude product for autonomous weapons and mass domestic surveillance.

Regardless of the specific trigger, the ban revealed a critical vulnerability in the current regulatory framework. As Michael Froman, President of the Council on Foreign Relations, noted, the United States has demonstrated its capacity to deny foreigners access to frontier AI models at will, much like how China controls rare earth magnets or Iran controls the Strait of Hormuz. This realization is now driving countries to reconsider their technological dependencies.

"The United States has sought to embed as many allies and partners as possible in the U.S. tech stack. Here, there will be a trade off between that objective and the risk of sharing some of our most powerful technologies. Other countries may well opt for AI sovereignty over serfdom, even if the alternatives are sub-optimal," explained Michael Froman.

Michael Froman, President, Council on Foreign Relations

How Does This Reshape Europe's Technology Strategy?

The new U.S. approach creates a strategic dilemma for European governments, many of which continue to frame technology policy around achieving "digital sovereignty" and reducing external dependencies. Helberg's counter-argument suggests that the greater vulnerability lies in spending scarce resources replicating technologies others already produce instead of investing in next-generation innovation that creates lasting competitive advantage.

Italy's participation in this week's AI Opportunity declaration signals that Washington's vision is already moving from theoretical debate to practical implementation. The U.S.-Italy dialogue on Trusted Technologies, coordinated by Roberto Baldoni and the Krach Institute of Tech Diplomacy, is no longer simply a bilateral initiative but increasingly appears aligned with the broader architecture emerging through Pax Silica.

The fundamental question reshaping industrial policy across the West is no longer how to achieve complete technological self-sufficiency, but rather how each trusted partner can create the greatest value within a shared ecosystem while preserving control over strategically sensitive technologies and reducing dependence on systemic competitors like China.

What Are the Risks of This New Model?

The current domestic regulatory paradigm for AI remains inadequate to manage these new dynamics. The United States still lacks comprehensive federal laws regulating AI, forcing the government to rely on ad hoc legal tools like export controls and supply chain risk designations. This creates uncertainty for private industry, which now faces the volatility of policymakers with few constraints on state intervention and no certainty regarding the nature or timing of such interventions.

The Mythos class model ban also exposed a deeper governance challenge: it is now settled that government, not companies, will determine whether a model or particular use case is too dangerous. The Pentagon previously criticized Anthropic for attempting to restrict military use of its products for "all lawful purposes," including autonomous weapons. Now the government has decided that another Anthropic product is too dangerous for public release. This inconsistency leaves private industry vulnerable to rapid policy reversals.

For countries outside the U.S. sphere, the choice is increasingly stark: build with American models and risk losing access at any time, or develop domestic and open-source alternatives largely from China that lag behind the cutting edge but pose additional geopolitical risks. This dynamic could accelerate fragmentation of the global AI landscape, slow innovation growth, and ultimately undermine the U.S. goal of defining the global AI architecture.