There is a particular kind of anxiety that has crept into Australian technology conversations over the past two years, a nagging sense that the country is being left behind while the United States and China pour tens of billions of dollars into building ever larger artificial intelligence systems. Every week brings news of another giant model, another data centre the size of a small suburb, another founder promising machines that will soon outthink us. Against that backdrop, the natural instinct is to ask how Australia keeps up.
Veteran technology investor Daniel Petre thinks that is the wrong question. In an opinion piece for Forbes Australia, the AirTree Ventures co-founder argues that the country will not win the global AI arms race, and, more provocatively, that it does not need to. His case is that Australia has spent too long measuring itself against the wrong scoreboard, fretting about foundation models and compute clusters it was never going to build, when its real leverage sits somewhere else entirely.
The news
Petre’s argument, laid out in his essay for Forbes Australia, starts from a blunt acknowledgement. AI is the biggest technological shift in modern history, it is already better than humans at a growing list of tasks, and it touches almost every corner of the economy. None of that is in dispute. What he rejects is the idea that Australia should respond by trying to match the hyperscalers on their own terms.
The maths simply does not work. Training a frontier model now costs sums that dwarf the entire Australian venture capital pool, and the companies doing it, from the American labs to their Chinese rivals, are backed by capital and state interest on a scale no Australian effort could realistically rival. Attempting to compete head on, in Petre’s telling, would mean burning scarce national resources on a contest that is already lost before it starts.
Instead, he wants the country to play to its structural advantages. Australia holds some of the world’s richest deposits of the critical minerals that go into the hardware underpinning the AI boom, including lithium and rare earths. It has abundant sunshine and wind, and therefore the potential to generate the cheap, clean electricity that the enormous new data centres are desperate for. Those are not consolation prizes, in his framing, they are the foundations of the whole industry, and Australia happens to sit on them.
The second half of his pitch is about money, specifically about making sure the global technology giants that profit from Australian users actually pay a fair share of tax here. If the platforms are going to hoover up value from the Australian economy as AI reshapes it, the argument runs, then the revenue they generate locally should be taxed properly and reinvested in the things that genuinely lift national capability, from skills to infrastructure.
Two ways to read it
Petre’s position lands in the middle of a live and sometimes tetchy debate about ambition. On one side sit those who worry that “play to your strengths” is a polite way of accepting permanent junior status, a digital echo of the old lament that Australia digs things up and ships them off while the clever, high-margin work happens somewhere else. Critics of the mining-and-energy framing argue that selling minerals and megawatts into someone else’s AI supply chain leaves the country exposed to exactly the same commodity cycles it has always ridden, and captures little of the value created further up the stack.
On the other side are those who see clear-eyed realism where others see defeatism. This camp points out that Australia already has genuine software success stories, from Atlassian to Canva, built without ever trying to out-spend Silicon Valley, and that the smarter move is to keep backing applied AI, the businesses that take existing models and turn them into useful products for healthcare, mining, agriculture and finance. On that reading, Petre is not lowering the country’s sights, he is pointing them at targets it can actually hit.
Both readings can be true at once, which is part of why the essay resonates. The question of whether Australia should chase sovereign capability or specialise has been simmering through recent policy fights, including the row over sovereign large language models and government procurement, and Petre’s intervention gives the specialisation camp a sharpened set of talking points.
Why it matters for Australia
The stakes here are not abstract. Governments at every level are being asked to make expensive bets, on data centre precincts in the regions, on gas and renewables to power them, on skills programs and research funding, and the framing that wins the argument will shape where billions of public and private dollars flow over the next decade. If the country accepts Petre’s premise, the priority becomes securing the mineral and energy supply chains, negotiating harder with the platforms on tax, and pouring resources into applied AI rather than foundation models.
There are already signs of both instincts pulling at once. Regional towns from Wagga Wagga to the Northern Territory are being pitched as future data centre hubs, drawing on cheap land and, in some cases, gas, while universities and enterprises scramble to lift the workforce capability that any AI-enabled economy will need. Meanwhile the tax question Petre raises is hardly settled, with successive governments having struggled to make multinational technology companies pay in proportion to the value they extract from Australian consumers. His essay effectively argues that the AI era makes solving that problem more urgent, not less, because the sums at play are only going to grow.
There is also a national security dimension that sits just beneath the surface. Critical minerals and clean energy are not only economic assets, they are strategic ones, and Australia’s position as a reliable supplier to allied nations gives it leverage that a middling model-training effort never would. Playing to those strengths, on this view, buys the country a seat at the table in a way that vanity projects would not.
What’s next
None of this is settled, and Petre’s piece is a provocation rather than a policy. The test will come in how governments and investors respond, whether the next tranche of funding chases headline-grabbing sovereign compute or the less glamorous work of securing supply chains, cleaning up the tax base and commercialising applied AI. Petre and AirTree have skin in that second game, backing the kind of software companies that would benefit if the country doubled down on application rather than infrastructure, and readers are entitled to weigh that interest against his argument.
What is harder to dismiss is the underlying discipline of his case. In a debate that too often runs on fear of missing out, the suggestion that Australia should stop trying to be something it is not, and instead get very good at what it already is, has a certain stubborn logic. Whether the country’s leaders have the confidence to act on it is another matter entirely.
Sources: Forbes Australia.
