Australia brings more than a quarry to Pax Silica

The inaugural Pax Silica Summit in December 2025 was more than the launch of a new economic initiative. The US-hosted event was the declaration of a new organising principle for great power competition – clear on purpose yet vague on implementation. That vagueness is an opportunity for Australia. One with a closing window.

Pax Silica brings together partners with a shared objective of a trusted AI supply chain capable of resisting coercion and deploying transformative technologies at scale. So far, it includes the United States, Australia, Finland, Greece, India, Israel, Japan, Norway, Qatar, South Korea, Singapore, Sweden, the Philippines, the United Arab Emirates and the United Kingdom, with Taiwan a non-signatory participant. Yet the outcome of the talks in December provided no clear roadmap or governance framework to achieve its objective. In that void, the division of capabilities and resulting economic opportunities will be contested. Countries that engage early and deliberately, clearly articulating what they bring, will have disproportionate influence over how those arrangements take shape.

Trade Minister Don Farrell captured the prevailing view of Australia’s role when he described it as holding “vast resources of the critical minerals that the world is going to need to decarbonise, to build data centres, to process AI”. The emphasis on “the world”, not Australia, was notable.

Australia’s critical minerals agenda is too limiting if pursued in isolation. It must also extend beyond the default focus on minerals in the ground. Without a deliberate effort to expand Australia’s role into technology initiatives, the country’s node within Pax Silica risks being defined by geology rather than the full range of AI infrastructure it can provide.

The figures show the scale of Australia’s potential. Between 2023 and 2025, announced investment in Australian data centre capacity exceeded $100 billion. In 2024, Australia ranked second globally after the United States as a data centre investment destination, driven by renewable energy, robust privacy protections, stable governance, and high-quality Indo-Pacific connectivity. All the industry needs now is certainty and support.

The economic case for AI infrastructure is sometimes challenged on the grounds that Australia can increase productivity from AI adoption regardless of where servers sit. While true, that argument fails to appreciate the larger strategic and economic picture. A significant subset of workloads related to government, defence, health, and financial services either legally cannot sit offshore or carry unacceptable risk if they do. By 2030, this could result in up to 40% of AI workloads being driven to sovereign environments, representing a market projected to reach US$600 billion.

Beyond productivity, failing to develop local infrastructure means forgoing Australia's opportunity to serve as a regional hub for trusted compute – an export opportunity valued at $11 billion with substantial spillover benefits for the broader economy.

With Industry and Innovation Minister Timothy Ayres affirming that domestic data centre investment “strengthens our security, supports our startups and researchers and ensures Australian data benefits Australians”, the government is aware of the stakes. In November 2025, Minister Ayres announced the establishment of an AI Safety Institute to support the adoption of emerging capabilities, while mitigating their risks. In December 2025, the National AI Plan consolidated over $460 million in pre-existing research and skills funding and launched an AI Accelerator program. In March 2026, the Expectations for Data Centres and AI Infrastructure Developers established a nationally consistent framework. These are meaningful steps, but frameworks are not the same as financial commitment and diplomatic engagement. The critical minerals agenda progressed because the government backed it with loan guarantees, equity support, and bilateral agreements. Now is the time to establish similar architecture for AI infrastructure.

The announcement of the Pax Silica Fund in March 2026, committing US$250 million to seed a co-investment pool whose founding members manage more than US$1 trillion, highlights the opportunity and urgency of this moment. Notably, the fund explicitly includes energy infrastructure as a response to supply chain disruptions caused by conflict in the Middle East. As conflict risk continues to reprice stable operating environments, Australia’s geography, institutional stability, and Five Eyes integration will increasingly move from background features of its Pax Silica membership to active competitive advantages. However, these advantages cannot be fulfilled without the government making a compelling case and demonstrating financial commitments.

Pax Silica’s implementation agenda is still being written, the fund being capitalised, and the project pipeline assembled. Those who engage with a clear offer and the resources to underwrite it will shape what gets built, where, and by whom.

The next step for Australia is to match the energy already directed at critical minerals toward AI infrastructure, with the government delivering a dedicated investment vehicle, bilateral compute agreements within the Pax Silica framework, and sustained diplomatic engagement to make the case. The minerals matter. The data centres matter more than the current framing suggests. That argument needs to be made, and funded, before it is lost.