Australia’s AI story is not limited to how local companies are using the technology. The country is also positioning itself as an emerging regional market for the physical infrastructure that supports AI.

With abundant land, significant renewable energy potential, reliable digital infrastructure and large organisations operating in data-intensive sectors such as banking, healthcare and government, Australia is attracting significant investment in data centres.

As highlighted in Roland Berger’s Australia’s AI journey report, the domestic data centre market has expanded rapidly, growing from approximately 700 MW in 2017 to around 2,800 MW in FY25. This represents a 20 per cent compound annual growth rate (CAGR).

The global consulting firm said Australian data centre demand is increasingly being driven by AI workloads, with colocation AI accounting for most capacity take-up in 2024 and gaining share from traditional cloud.

AI has also increased power requirements, with modern racks demanding densities of 30 to 100 kW, compared with legacy baselines of 5 to 15 kW.

Australia’s data centre sector is facing a surge in demand as hyperscalers and AI-focused cloud providers accelerate investment in large-scale infrastructure. AWS has announced a AU$20 billion investment programme across Sydney and Melbourne through 2029, while AI providers are pursuing data centre capacity blocks of 10 MW to 50 MW or more, Roland Berger said.

Operators are also expanding campus-scale developments, including AirTrunk’s planned SYD3 facility, which is designed to scale up to 320 MW.

However, growth is being constrained by infrastructure bottlenecks, Roland Berger points out.

“Grid capacity remains the biggest challenge, with around 8 GW of connection requests competing for limited transmission capacity and new 50 MW facilities potentially facing multi-year connection timelines.”

Supply chain pressures are adding further delays, with generator lead times of about 42 weeks and switchgear delivery times of around 30 weeks.

Roland Berger notes that high construction costs in Sydney and Melbourne, combined with AI-readiness gaps in legacy facilities, are increasingly pushing developers towards secondary markets such as Brisbane and Perth.

While Australia’s digital AI story is shaped by its services economy, the firm said the relative absence of heavy manufacturing shapes its physical AI story.

“Physical AI is already relevant to Australia, particularly in sectors where the technology fits existing operating conditions,” the report said.

Mining is one of Australia’s leading sectors for physical AI adoption. The report highlights Rio Tinto’s Pilbara iron ore operations as a prominent example, with autonomous systems supporting mining and transport operations.

“Remote locations, high labour costs and demanding safety requirements have created strong incentives for autonomous mining vehicles, and Australia has developed large-scale autonomous mining infrastructure, for instance, in its remote Pilbara region.”

Agriculture is following a similar trend, with Roland Berger highlighting that Australia’s large-scale farming operations and ongoing labour shortages are driving increased adoption of robotics across the sector.

“These sectors demonstrate a clear pattern: physical AI has found some of its strongest applications in Australia where the technology solves a well-defined operational problem, particularly in environments where distance, safety, labour costs or difficult operating conditions make autonomy valuable.”

Yet manufacturing and processing remain a significant investment base, with $155 billion of foreign direct investment invested in the sector in 2025, creating a meaningful foundation for AI-enabled automation.

Australia’s manufacturing sector has a lower level of robotics adoption than many global peers, with around 134 industrial robots per 10,000 employees compared with the global average of 162. Roland Berger notes that the rate remains well below that of leading regional manufacturing economies.

Early physical AI deployment therefore tends to concentrate in sectors where companies can efficiently collect operational data and where automation can address clear operational challenges. However, Australia has one of the lowest shares of physical AI deployment in Roland Berger’s Asia-Pacific sample, suggesting it risks falling behind its regional peers.

Closing the AI value gap
To assess the current state of enterprise AI adoption, Roland Berger surveyed 211 C-suite and technology executives across seven major Asia-Pacific economies in June 2026, including 30 Australian business leaders.

Seventy per cent of surveyed Australian executives genuinely believe that AI is transformative, yet 37 per cent of companies are still investing in it without demonstrating meaningful value.

One third of Australian companies plan to increase AI spending next year, despite Australia having the strongest levels of AI governance among developed APAC countries.

This comes as Australia takes a more active role in AI governance, with the federal government establishing the Office of AI within the Department of the Prime Minister and Cabinet in July 2026 to coordinate national AI standards.

“The proposed framework targets areas including AI training, large-scale data centres, energy and water use, grid impacts, and copyright protections, with legislation expected in early 2027. The government has described the initiative as the first AI governance framework of its kind to be legislated globally,” the report said.

According to Roland Berger, the move signals a broader shift towards increased regulatory oversight as AI becomes more integrated into Australia’s economy, with businesses likely to face greater government involvement in the sector.

Roland Berger’s managing partner at Southeast Asia & Australia, Damien Dujacquier, said the firm’s latest report shows that Australia’s challenge is no longer simply getting AI into production, but proving its value.

“Despite strong governance and substantial investment, many companies still measure AI activity rather than financial impact. Australian executives must introduce continuous ROI measurement to ensure that AI initiatives deliver measurable business outcomes.”

“Australia has the foundations to capture greater value, but the challenge is no longer whether Australian companies will adopt AI, it is whether they can turn that adoption into sustained business value,” the report said.

The difference between companies that convert AI investment into value and those that do not, according to Roland Berger, lies less in access to technology, capital or executive support than in execution discipline.

“Many still measure AI activity rather than AI value and remain cautious about giving agentic AI real decision-making authority. The analysis places Australia at the centre, using selected Asia-Pacific markets only to clarify its position,” the report said.

“Australia is advanced in AI governance, adoption and deployment. Yet these strengths have not consistently translated into business value, risking an adoption-value gap. The opportunity now is to learn from Australian companies already converting AI investment into measurable business outcomes.”

The structure of Australia’s economy shapes its AI opportunity, Roland Berger said.

“Over recent decades, the country has developed a services-led economy with deep capabilities in financial services, banking, healthcare, education, professional services, telecommunications, infrastructure and natural resources. That economic structure matters enormously for where AI is likely to create value.”