An analysis of federal contracts reveals that while Canberra relies on domestic companies for artificial intelligence guidance, it predominantly selects international technology behemoths for the implementation of AI systems. This disparity suggests a preference for foreign providers in high-value AI solutions, despite the presence of skilled local innovators.
New data indicates that Australian startups are largely confined to lower-value consultancy and training agreements. In stark contrast, larger, often foreign-owned, corporations secure lucrative multimillion-dollar deals for AI platforms. This trend has been observed in core AI contracts, which encompasses direct investments in AI software, platforms, and specialized advisory services, excluding general IT staffing and hardware.
Since 2015, foreign entities have collectively amassed approximately $38 million from 47 core AI contracts. Australian firms, despite winning a significantly higher volume of contracts—108—have only secured $31 million in comparison. The data from AusTender further illustrates that core AI expenditures have surged from about $2.2 million in 2024 to nearly $40 million by early September 2026, highlighting the rapidly increasing investment in this sector.
A closer examination of the AusTender records demonstrates that Australian businesses dominate the advisory and labor-related AI contracts, securing $15.3 million in AI consulting fees, while foreign competitors obtained $6.2 million. Furthermore, local companies have largely cornered the market for e-discovery contracts, essential for digital information retrieval in legal and investigative contexts.
However, contracts for foundational software and large-scale enterprise infrastructure overwhelmingly favor global tech giants. A notable example is the Australian Taxation Office (ATO), which awarded a $24 million limited tender to Teradata Australia, a subsidiary of a US-based corporation. The ATO's substantial $30.8 million expenditure in this area significantly skewed the overall data, with nearly 80 percent of this funding flowing to overseas entities.
In the domain of biometric and speech software, foreign companies secured 36 contracts totaling $13.4 million, whereas Australian developers managed to win just one contract valued at $36,300. This stark contrast underscores the challenge local firms face in competing for complex, high-value AI system deployments.
Senator David Pocock, an independent from the ACT, has voiced concerns regarding Australia's increasing dependence on overseas firms for vital technological products and services. He emphasized the critical need for developing robust sovereign AI capabilities, encompassing skills, computing access, and models, to ensure national security.
Senator Pocock remarked that many Australian AI companies, particularly in Canberra, are developing sophisticated solutions. Yet, these local innovators often find it easier to secure contracts with international governments than with their own Australian federal departments. This situation points to a systemic issue within the current procurement process rather than a lack of capability among Australian businesses.
Michael Gately, CEO of Trellis Data, a local firm, highlighted the significant hurdle Australian founders face in transitioning from advisory roles to winning large-scale enterprise platform deals. He attributed this challenge to structural flaws in procurement frameworks rather than any deliberate bias against Australian technology.
Gately expressed confidence that public servants are not intentionally disadvantaging Australian businesses. Instead, he believes the issue lies with underlying structural weaknesses in the procurement system, which, encouragingly, are rectifiable. He noted that many senior public servants are keen to see Australian businesses succeed.
The AusTender data also revealed considerable variations in AI spending patterns across different government departments. Some departments prioritize local AI spending, while others exclusively opt for foreign technology. For example, the Department of Defence allocated nearly 95 percent of its core AI contract value to Australian firms, a proportion only surpassed by the Industry department, which directed all of its $2.79 million core AI spending onshore.
Conversely, operational agencies such as the Australian Federal Police ($4.48 million) and Services Australia ($2.68 million) channeled 100 percent of their core AI expenditures to international entities. This divergence indicates a fragmented approach to supporting domestic AI capabilities across the government.
To rectify this imbalance, Gately proposed implementing a guaranteed quota for local software vendors. He argued that a mandatory set-aside for software procurement, specifically for sovereignly controlled Australian technology companies, including AI, could significantly boost national productivity, a key government objective. Such a policy would provide government agencies with the necessary authorization and incentive to prioritize local capabilities, especially in sensitive sectors like health, welfare, education, and critical infrastructure, where trust and sovereignty are paramount.
Despite these structural obstacles, Gately maintains an optimistic outlook regarding Australia's potential to evolve from being primarily consumers of global tech to becoming leading producers of sovereign AI. He hopes to continue collaborating with the government to address this vital issue, believing that sovereign AI technology can enhance productivity, create jobs, and foster a better future for coming generations.
The reliance on international technology giants for substantial AI contracts, despite the robust local AI advisory sector, poses strategic questions about Australia's long-term technological independence. Addressing this structural challenge through targeted procurement policies could unlock significant economic and national security benefits, fostering a thriving domestic AI industry.