Australia’s National AI Plan is a welcome start on skills and safety, but it plays too safe. While the US, Europe and the Gulf pour sovereign capital into chips, compute and energy, Canberra is still talking about catalysing investment rather than committing.
A significant supply chain attack has struck the US financial sector, with fintech vendor Marquis Software Solutions confirming a ransomware incident that exposed the sensitive data of hundreds of thousands of customers from dozens of American banks and credit unions.
South Korean e-commerce giant Coupang has confirmed a massive data breach exposing the personal information of 33.7 million customers. The incident, which began in June 2025, is one of the largest in the nation's history and is linked to a former employee's active credentials.
Australia’s National AI Plan is a welcome start on skills and safety, but it plays too safe. While the US, Europe and the Gulf pour sovereign capital into chips, compute and energy, Canberra is still talking about catalysing investment rather than committing.
The Albanese government’s National AI Plan is, on its face, a solid and timely step. It sets out a people-focused roadmap to build an “AI-enabled economy”, tied to the Future Made in Australia agenda, and it does several important things well.
It recognises that AI is not just a tech story, but a workforce, education and social story. It commits to skills development through TAFEs and universities, emphasises collaboration with workers and unions, and flags the need for models trained on Australian data so local culture and context are not lost. The creation of an AI Safety Institute, backed by 29.9 million dollars, sends a useful signal that Australia wants guardrails, not a free-for-all.
Business, tech and digital-economy groups have broadly welcomed the plan. Estimates of up to 200,000 new AI-related roles speak to the genuine upside if Australia can lift adoption across industry, services and the public sector. For companies such as AirTrunk, NEXTDC, CDC and emerging players like Firmus Technologies, the plan provides a policy backdrop that acknowledges data centres and digital infrastructure as serious economic assets, not just “sheds full of servers”.
On safety and inclusion, the language is strong. The government stresses that “technology should serve Australians, not the other way around”. It talks about spreading the benefits, protecting vulnerable communities, and ensuring workers have a voice as AI reshapes jobs and workplaces. That is good policy, and it aligns with Australian values.
Yet read closely and a different picture emerges. This is a document that manages AI, rather than one that sets out to make Australia a genuine AI power. Its rhetoric is about “catalysing” private investment that “could scale up to more than 100 billion dollars”, but it is vague on binding commitments.
What does “catalysed” really mean? How much of that 100 billion dollars is contracted, financed and shovel-ready? Where are the sovereign guarantees, the long-term offtake agreements, or the federal balance sheet being put to work in the way we see in the United States, Europe or the Gulf?
Equally striking is who is not fronting this agenda. The plan has been rolled out by the Minister for Industry and Science and the Assistant Minister for Science, Technology and the Digital Economy. Important portfolios, but this is not being presented as a whole-of-government economic transformation on par with defence, energy or trade.
Where is the Treasurer? Where is the Prime Minister standing beside industry and infrastructure leaders and making it clear that AI is a national competitiveness priority, not just a digital policy file? The absence of those voices tells markets that this is still a secondary agenda.
The plan also leans heavily on existing laws, with relatively modest new mandates. Safety functions are largely advisory. Regulators are asked to interpret AI through current privacy, competition and consumer frameworks, rather than being given a bespoke regime for high-risk systems. Civil-society groups and the Greens have already labelled it “light touch” and “toothless” on systemic risks.
In fairness, that light touch is one reason the plan is seen as pro-business and pro-infrastructure. It lowers political risk for investors looking to deploy capital in data centres and AI services. But it stops short of answering the core strategic question: is Australia content to be a very good customer of global AI, or does it intend to become a serious producer and owner of the technology stack?
At present, the language is about benefiting from AI, not building it. That is a subtle but important distinction.
To understand where Australia really sits, it helps to look at how other jurisdictions are moving.
In the Gulf, Saudi Arabia and the United Arab Emirates are explicit about wanting to be the third global AI pillar alongside the United States and China. Saudi’s HUMAIN initiative is backed by a sovereign fund approaching 1 trillion US dollars, with plans for about 6 gigawatts of data centre capacity by 2034 and deep partnerships with Nvidia, AMD, AWS, Qualcomm and Cisco. The UAE is stacking sovereign capital behind MGX as a national AI investment vehicle.
In Europe, the InvestAI program is mobilising roughly 200 billion euros in public and private funds, including 20 billion euros for “AI Gigafactories” that combine compute, data and talent in 4 or 5 sovereign facilities dedicated to training very large models.
In the United States, industrial policy has swung behind AI with unusual clarity. Federal financing, accelerated permitting, and an American AI Exports Program are designed to promote the full US stack: chips, models, software and standards. Tight export controls over advanced semiconductors are being used to tilt the playing field in favour of American firms and allies.
Set against that, Australia’s 29.9 million dollars for an AI Safety Institute and loosely defined 100 billion dollar “catalysed” private investment look modest. The real gap, however, is not just in scale but in posture. Others are using sovereign capital, industrial policy and export controls to build national capability. Australia is hoping to attract capital and manage risks.
On the ground, capital is already moving faster than Canberra. AirTrunk has built a portfolio of hyperscale data centres across the Asia-Pacific and is now talking about a 1 gigawatt campus in the region, far beyond what was imagined when its first 100 megawatt site was pitched less than a decade ago. Its founder, Robin Khuda, has repeatedly framed AI data centres as enabling infrastructure for productivity, the clean energy transition and digital sovereignty, and has called for streamlined regulation, faster approvals and clear rules on land and energy to seize the moment.
Firmus Technologies’ Project Southgate, backed by Nvidia and working with CDC Data Centres, is planning around 1.6 gigawatts of AI capacity across Tasmania, Melbourne and future mainland hubs, with an initial 150 megawatts slated by mid 2026. The project aims to use tens of thousands of GB300 GPUs for sovereign “AI factory” workloads and to integrate renewables at scale.
Craig Scroggie
NEXTDC has struck a landmark deal with OpenAI to co-develop a hyperscale AI campus and GPU supercluster at its S7 site in Sydney, with market talk of a roughly 650 megawatt facility at Eastern Creek and a long-dated anchor commitment from OpenAI. CDC has secured New South Wales approval for a 3.1 billion dollar, 504 megawatt campus at Marsden Park, described as the largest data centre campus in the Southern Hemisphere.
Taken together, these moves show that Australia is becoming a serious destination for AI-grade infrastructure. That is good news, and it reflects genuine strengths: political stability, strong institutions, advanced capital markets and a deep pipeline of engineering and construction capability.
Yet these advances are occurring in spite of, not because of, a tight federal AI-energy strategy. Grid constraints are already limiting new connections in key locations. The Australian Energy Market Operator has warned that data centres may account for a double-digit share of grid demand within 15 years, while renewable generation and transmission build-out are lagging stated climate ambitions.
Saudi Arabia’s HUMAIN chief Tareq Amin has been blunt in pointing out that the kingdom’s existing energy infrastructure saved him 18 months of work because he did not need to build new substations to support AI campuses. In Australia, by contrast, energy is the bottleneck. There is a risk that data centre proponents exhaust available capacity in pockets of the grid, while other industries and households face higher prices and reliability concerns.
The National AI Plan nods to renewables and acknowledges data centre growth, but it does not treat energy as a central pillar. There is no integrated blueprint that couples AI investment with guaranteed, bankable access to low-carbon power at scale. There is no explicit mechanism to ensure that AI infrastructure expands in lockstep with transmission upgrades, storage and firming capacity.
This is not just a technical oversight. It goes to competitiveness. AI investors now have options. They can deploy capital in Saudi Arabia, the UAE, the United States, Europe or Southeast Asia, where governments are willing to move quickly on land, permitting, power and finance. Australia’s advantage as a safe, rules-based destination will not compensate forever if energy is expensive or unreliable and strategic direction is unclear.
Australia has strong AI research, innovative startups and a fast-growing ecosystem of integrators and applied AI companies. The plan does talk about sovereign models trained on Australian data and an AI Accelerator funding round under the Cooperative Research Centres program. But there is no explicit national AI champion, no commitment to domestic chip manufacturing at any scale, and no export control or procurement strategy aimed at giving domestic firms a structural advantage.
In effect, Australia is betting that it can be an attractive host for global players such as OpenAI, Nvidia and AWS, while its own companies compete in the middle and upper layers of the stack. That may deliver jobs and tax revenue. It will not deliver sovereign control over critical AI infrastructure, nor will it create the kind of national champions that rival economies are now cultivating.
Australia has never been afraid of competition. As a country we celebrate punching above our weight on the sporting field and in business. We understand what it takes to win: clear goals, disciplined investment and high-performance cultures that are built, not wished into existence.
The opportunity in AI is no different.
The National AI Plan is a useful starting point. It sets a tone of inclusion, safety and collaboration that fits the Australian temperament and provides enough policy certainty to keep global investors interested. But it is only a first lap.
To convert “good enough” into genuine leadership, Australia will need to move from aspirational language to binding commitments. That means treating AI and digital infrastructure as core economic assets, not peripheral policy interests. It means putting the Treasurer and Prime Minister at the centre of the conversation, embedding AI into energy planning and industrial strategy, and being honest that sovereign capability will cost real money and require real trade-offs.
The country is standing at a digital crossroads. With OpenAI, Nvidia, AirTrunk, NEXTDC, Firmus and CDC already using Australia as a proving ground for hyperscale campuses, the window to turn a light-touch plan into a full-blooded national strategy is open, but not indefinitely.
Australia can choose to be a cautious but capable adopter of other nations’ AI, or it can decide to compete for gold medals in the global AI race. The first path is safe and familiar. The second is harder, but it is the one that matches who we say we are as a nation.
Get the stories that matter to you.
Subscribe to Cyber News Centre and update your preferences to follow our Daily 4min Cyber Update, Innovative AI Startups, The AI Diplomat series, or the main Cyber News Centre newsletter — featuring in-depth analysis on major cyber incidents, tech breakthroughs, global policy, and AI developments.
Oracle has become a test of how much debt markets will tolerate in the AI buildout. Rising credit costs and growing use of swaps show investors backing the AI story while seeking protection in case promised workloads and cash flows fail to arrive on time.
NVIDIA’s staggering 57 billion quarter didn’t just calm the market, it rewrote the AI story. What looked like a bubble unwind now looks like a misread. With sovereign AI deals, a Saudi megacentre and a global capex surge, NVIDIA has reset the race and reminded investors this era is only beginning.
NVIDIA’s blockbuster quarter has reset the AI narrative, turning fears of a bursting tech bubble into renewed conviction in a structural shift. With record data-centre sales and sold-out Blackwell GPUs, NVIDIA now looks less like a chip stock and more like core AI infrastructure in the AI build-out
Australia stands at a digital crossroads as leaders race to turn AI ambition into real infrastructure. With global rivals surging ahead, the question is whether policy, power and capital can finally align to secure true AI sovereignty.
Where cybersecurity meets innovation, the CNC team delivers AI and tech breakthroughs for our digital future. We analyze incidents, data, and insights to keep you informed, secure, and ahead. Sign up for free!
