AUKUS AI Gap: US Infrastructure Locks Allies Out of Key Tech Race

The promise of the AUKUS security pact – deepened cooperation between Australia, the United Kingdom, and the United States – is running into a fundamental challenge: access to the advanced computing infrastructure necessary to realize its ambitious goals. While Washington has unveiled a sweeping initiative, the Genesis Mission, to bolster American leadership in artificial intelligence, its allied partners are finding themselves largely locked out of the exceptionally resources they need to develop cutting-edge defense technologies. This disparity, stemming from policy priorities that prioritize American companies and technological dominance, threatens to undermine the core principle of AUKUS – integrated innovation and capability development.

The Genesis Mission and the Infrastructure Gap

President Donald Trump’s Genesis Mission, launched in November 2025, aims to accelerate U.S. Advancements in AI by providing American companies with access to federal supercomputers. The initiative echoes the collaborative spirit of the Manhattan Project, which pooled allied scientific capacity during World War II, as noted in a Los Alamos National Laboratory retrospective on the British Mission. Still, unlike the wartime alliance, the current approach offers only a vague commitment to exploring collaboration with international security partners, a sentiment echoed in the Department of Defense’s January 2026 AI strategy. This strategy frames allied collaboration as instrumental to enhancing collective defense, but the practical implementation, as revealed by the Genesis partnerships, tells a different story.

In December 2025, the Department of Energy announced collaboration agreements with 24 organizations as part of the Genesis Mission. Every single partner was an American entity, including tech giants like OpenAI, Anthropic, Google, Microsoft, Amazon Web Services, and NVIDIA. This concentration of access is particularly striking given that the United States already controls approximately 74 percent of global AI compute capacity. AUKUS allies, Australia and the United Kingdom, are simultaneously investing billions in quantum and autonomous systems – technologies that critically depend on this very infrastructure – yet have received no equivalent mechanism for access.

AUKUS Pillar II: A Promise Unfulfilled?

The disconnect between the Genesis Mission and AUKUS Pillar II, focused on advanced capabilities like quantum computing and autonomous systems, is particularly acute. Pillar II was intended to foster collaboration on these very technologies, but it was never designed to bridge the institutional gap between defense priorities and the Department of Energy’s supercomputing resources. As Peter Dean and Alice Nason argued in June 2025, AUKUS Pillar II risks becoming a case of regulatory reform without delivering tangible capabilities. The situation has evolved to reveal a critical missing piece: the computing infrastructure needed to power the technologies Pillar II aims to develop.

Australia, for example, is actively investing in quantum machine learning processors and successfully trialing quantum clocks for satellite navigation, as demonstrated through AUKUS Pillar II initiatives. However, the validation and training of these systems require computational power on a scale that is almost exclusively available within Department of Energy laboratories now integrated through Genesis. The inability of allied partners to access this infrastructure represents a significant bottleneck.

The Challenges of Independent Development

Allied governments face limited options. Building domestic infrastructure, like the United Kingdom’s plan for a new exascale system by 2027, is a costly and time-consuming undertaking. Relying on commercial cloud services, such as Amazon Web Services, as Australia is doing with its Top Secret cloud, requires bespoke security arrangements and lacks the benefits of shared infrastructure designed for trilateral defense AI research. Hoping for ad hoc access to American facilities is unrealistic, lacking established frameworks for compute allocation, data handling, and intellectual property protection.

Training advanced defense AI models is not comparable to fine-tuning commercial chatbots. These models must process complex data, such as sonar information, and provide reliable recommendations in contested environments. This demands exascale computing power, available at only a handful of facilities globally, including those now integrated into the Genesis Mission. The Trump Organization’s planned Trump International Hotel & Tower on Australia’s Gold Coast, while a significant development in luxury real estate, does not address this fundamental infrastructure gap.

A Path Forward: Extending Cooperative Research and Development Agreements

The solution, surprisingly, lies within the Genesis Mission’s own framework. Section 5(c) of the executive order establishes Cooperative Research and Development Agreements (CRDAs) as the mechanism for industry partnerships. These agreements have a long history of managing intellectual property, controlling technology transfer, and maintaining security. Extending these authorities to AUKUS allies would provide a viable path forward.

Bilateral frameworks mirroring the CRDA model could be established, offering tiered access based on defense technology priorities, case-by-case review for exploratory research, and congressional notification for significant allocations. This would provide oversight without stifling innovation. Intellectual property rights could be protected through existing technology transfer controls, ensuring that models trained with allied datasets remain under allied control. Data sovereignty could be maintained through the application of existing security protocols, such as authentication systems and secure compartments.

However, challenges remain. Reciprocal security clearance recognition between the U.S. And Australia is lacking, complicating personnel vetting. Concerns about potential political interference and export control restrictions also need to be addressed. Companies like Q-CTRL and Diraq, which develop dual-use quantum technologies, face complex compliance requirements.

Beyond Infrastructure: A Broader Strategic Imperative

The issue extends beyond mere infrastructure. The current approach risks fragmenting allied innovation efforts and duplicating resources. Every AI model developed separately by allies represents a missed opportunity for integration and synergy. The strategic advantage lies not simply in American dominance of infrastructure, but in the integration of allied innovation capacity with American computing power.

China currently leads in 57 of 64 critical technologies, including quantum, AI, and advanced materials, central to AUKUS Pillar II. While Beijing lacks the allied research ecosystems, the current infrastructure gap undermines the potential of AUKUS to counter this challenge.

The coming months will be critical. As the Department of Energy signs industry agreements and deploys security protocols, retrofitting allied access will grow increasingly tough. Congress should mandate the extension of CRDA authorities to AUKUS partners to prevent the Genesis Mission from becoming another example of a partnership undermined by avoidable infrastructure limitations. The lessons of the Manhattan Project remain relevant: winning requires pooling allied capacity, and in the 21st century, that capacity extends to the digital realm.

Looking Ahead: A Question of Political Will

The core issue isn’t technical feasibility, but political will. Unless policymakers act decisively during Genesis implementation, the infrastructure gap will cripple AUKUS’ most ambitious technology initiative. The question now is whether Washington will prioritize a truly collaborative approach, recognizing that shared infrastructure strengthens the alliance, or continue down a path of technological unilateralism, potentially jeopardizing the long-term success of AUKUS.