Artificial intelligence may look like a competition between software companies, but its foundations are physical: semiconductor factories, mineral-processing plants, power grids and data centers.

Washington’s Pax Silica initiative reflects that reality. Rather than focusing only on better models or faster processors, the United States is trying to connect the industrial capabilities needed to build and operate AI at scale.

The initiative’s original eight invited partners were Australia, Japan, South Korea, the Netherlands, the United Kingdom, Israel, Singapore and the United Arab Emirates. Announced ahead of the December 2025 summit, that list should not be confused with the countries that immediately signed the declaration. The initial signatories were the United States and six partners; the UAE formally joined in January 2026. A second summit followed in June, broadening cooperation.

Not Just a Club of Mining Countries

The choice of partners reveals Washington’s strategy. Most are not major mineral exporters. They matter because they occupy different positions in the technology supply chain.

Australia brings mineral resources and opportunities to expand processing. Japan supplies essential semiconductor materials and manufacturing equipment. South Korea contributes memory technology, including the high-bandwidth memory used alongside AI accelerators. The Netherlands supplies advanced lithography equipment through ASML.

Britain’s strengths include chip design, intellectual property and compound semiconductors. Israel contributes chip-design and data-center technologies. Singapore combines semiconductor manufacturing and packaging with international logistics. The UAE offers investment capital, energy and infrastructure.

Together, these capabilities suggest an effort to link the entire production system—not merely secure additional shipments of ore.

What Does the Cooperation Actually Cover?

The declaration promotes cooperation across mineral refining, energy, semiconductors, manufacturing, communications infrastructure, logistics, models and applications. It also calls for coordination on investment security and the protection of sensitive technologies and infrastructure.

That makes Pax Silica an economic-security framework, not a single mining agreement or a guaranteed purchasing program. Its significance will depend on whether governments and companies turn shared principles into commercially viable projects.

A separate U.S.–Australia minerals framework illustrates the types of tools available: coordinated investment, financing, supply arrangements and strategic reserves. Those bilateral commitments should not, however, be treated as obligations automatically shared by every Pax Silica participant.

The Bottleneck Is Often Processing, Not Geology

The phrase “resource race” can obscure an important distinction. Finding minerals is not the same as producing the high-purity materials that manufacturers need.

Refining requires specialized equipment, technical expertise, environmental management and reliable customers. China’s position in several mineral supply chains rests on those industrial capabilities, not simply on the location of deposits.

The International Energy Agency’s 2026 assessment found that supply remained highly concentrated, particularly in refining. Much of the recent increase in processed supply came from already dominant producers. Diversifying mining therefore does not automatically diversify the finished-material market.

Nor should every critical mineral be described as an ingredient in an AI processor. The initiative covers a much wider system, including electricity supply, communications and supporting infrastructure.

What It Means for American and European Communities

For Western households, this competition reaches beyond technology rankings. It raises questions about industrial jobs, taxpayer support, environmental safeguards and access to electricity.

New mines and refineries may create employment, but they also require permits, infrastructure and community acceptance. New data centers need substantial power supplies. Governments must decide how to support strategic projects without underwriting indefinitely uncompetitive businesses.

Companies face another trade-off: a more diversified supply chain may be more resilient, but it will not necessarily be cheaper. Investors need confidence that new facilities will have customers even when commodity prices fall or political priorities change.

The broader lesson is that the AI race has not stopped being a contest of ideas. It has become a contest of ideas supported by industrial capacity.

Pax Silica can help organize investment and cooperation. It cannot instantly recreate established processing and manufacturing ecosystems. Its ultimate test will be whether its partners can convert resources into dependable materials, chips and computing capacity—at a cost businesses and taxpayers are willing to bear.

Editorial note

This essay is intended for general information and analysis. It is not investment advice. Facts and interpretations may be revised as new information becomes available.