Australia’s defence technology sector is experiencing a period of expansion not seen since the Cold War. Driven by AUKUS commitments, a deteriorating regional security environment, and bipartisan political support for defence self-reliance, the sector attracted an estimated $2.8 billion in combined public and private investment in 2025. A substantial portion of that capital is flowing into artificial intelligence.

The ambition is significant: build a domestic defence AI capability that reduces reliance on allied systems while maintaining interoperability with the United States and United Kingdom. Whether Australia can actually achieve that within the timelines being discussed is another question entirely.

The Scale of Commitment

The numbers are real. The Department of Defence’s Advanced Strategic Capabilities Accelerator (ASCA), established in 2023, committed $3.4 billion over its first decade. Its AI-specific programs—focused on autonomous systems, intelligence analysis, electronic warfare, and cyber operations—account for roughly 40 percent of that total.

The 2025-26 Defence budget included $1.2 billion specifically for sovereign capability development in AI and autonomous systems. On top of that, the Defence Innovation Hub has funded more than 80 AI-related projects since its expansion in 2024, with grants ranging from $500,000 for early-stage feasibility studies to $45 million for advanced prototype development.

State governments are contributing too. South Australia’s Lot Fourteen precinct in Adelaide now houses over 30 defence technology companies alongside ASCA’s headquarters and the Australian Signals Directorate.

What Sovereign AI Actually Means

The term “sovereign capability” gets used loosely, but in the defence context it has specific meaning: the ability to develop, deploy, maintain, and upgrade critical technology systems without depending on foreign suppliers for essential components.

For AI, that’s a demanding standard. It means Australian-developed algorithms trained on Australian-controlled data, running on secure domestic infrastructure. It means enough skilled personnel to build and maintain these systems locally. And it means keeping pace with adversaries investing at far larger scale—China’s military AI spending is estimated at over $15 billion annually.

Australia can’t compete on volume. The current strategy focuses sovereign AI efforts on four domains: signals intelligence, autonomous maritime surveillance, electronic warfare, and logistics optimisation. In other areas, Australia will depend on shared capabilities through AUKUS and Five Eyes.

The Talent Problem

Australia’s defence AI sector faces the same talent constraints as the broader tech ecosystem, but with additional complications.

Security clearances take months to process. Many of the country’s best AI researchers are foreign nationals who can’t work on classified programs. Defence salaries, while improved, still trail what major technology companies or consulting firms offer. And the cultural gap between academic AI research and defence applications remains wide—many researchers simply aren’t interested in military work.

ASCA has tried to address this through its workforce development program, which funds PhD students and postdoctoral researchers in defence-relevant AI disciplines. About 120 researchers are currently supported through the program. Industry partners like their consulting practice have contributed to broader workforce development efforts, though the pipeline remains thin relative to demand.

The Defence Science and Technology Group (DSTG) employs roughly 400 AI and data science specialists, a number that’s expected to grow to 600 by 2028. But turnover is a persistent challenge—experienced AI engineers can earn significantly more in the private sector, and several high-profile departures in 2025 highlighted the retention problem.

What’s Actually Being Built

Despite the constraints, tangible progress is being made.

Ghost Shark autonomous undersea vehicles are Australia’s most visible defence AI program. Developed by Anduril Australia, Ghost Shark moved from concept to advanced prototype in under three years. These AI-directed underwater drones handle surveillance, mine detection, and anti-submarine warfare. Sea trials began in late 2025 off South Australia.

Project Redspice, the Australian Signals Directorate’s $9.9 billion cyber and intelligence investment, includes significant AI components for automating signals analysis and identifying cyber threats. Public statements indicate multiple AI systems have moved into operational use.

Autonomous logistics is a quieter win. The Army’s Land 8140 program is testing AI-driven supply chain optimisation and autonomous ground vehicles, aiming to reduce personnel needed for rear-echelon support.

Electronic warfare AI for signal detection and response is being developed by DSTG and BAE Systems Australia. These systems must operate at machine speed—faster than any human operator.

The Gaps That Remain

For all the investment, significant gaps remain in Australia’s defence AI capability.

Testing infrastructure is limited. Australia lacks a dedicated defence AI testing range comparable to US or UK facilities. Autonomous systems need specialised testing environments that don’t exist domestically at scale.

Supply chain vulnerabilities persist. Algorithms can be developed locally, but the hardware—specialised AI chips, sensors, communication equipment—is largely manufactured overseas. True sovereignty requires either domestic hardware manufacturing or guaranteed allied supply arrangements.

Pace remains a concern. Defence procurement has historically taken a decade or more. AI evolves in months. ASCA was designed to move faster, but institutional culture is hard to shift.

What Comes Next

Australia’s defence tech sector is at an inflection point. The funding is there and the strategic rationale is clear. The question is execution. Building sovereign AI capability requires talent, industrial capacity, testing infrastructure, and procurement reform advancing together.

The next two to three years will be decisive. If Ghost Shark, ASCA’s procurement model, and the talent pipeline deliver, Australia will have credible sovereign defence AI by decade’s end. If any element stalls, the country stays dependent on allied systems for its most critical needs. Given the pace of change in the Indo-Pacific, that’s a dependency Australia can ill afford.