Australia Made the Right Call on Globalisation. Will It Make the Right Call on AI?

In the early 1980s, Bob Hawke and Paul Keating stared down one of the hardest calls any Australian government has faced. The tariff walls that had kept local industry comfortable for decades were also keeping it soft. Globalisation was coming whether Australia wanted it or not. They chose to tear the walls down, let the cold air in, and back Australian businesses to compete. The short-term pain was real. The thirty-year prosperity that followed was realer.

Peter Hartcher, political and international editor at The Age, made a pointed argument in his column this week: Australia simply cannot afford to miss this AI wave. He’s right. And the parallel he’s reaching for goes deeper than a metaphor. Same party, same fork in the road. And the stakes are, if anything, bigger.

The same fork, forty years on

Anthony Albanese stood at the UN General Assembly on 25 September and said plainly: “We can’t ignore AI or prevent it.” He then signed a guardrails agreement with 19 nations, notable in itself, except that the US and China didn’t sign. The two countries doing most of the building sat this one out. The agreement is symbolic at best.

That’s not a criticism of the intent. But it illustrates the gap between where Australia’s political energy is going (governance, guardrails, responsible frameworks) and where the actual economic action is happening.

The scale of what’s coming

Google, Amazon, Microsoft and Meta have spent a combined US$1.1 trillion on AI infrastructure over the past three years. Another US$745 billion is planned for this financial year alone. Goldman Sachs puts global AI investment in 2026 at roughly US$1 trillion. The Brookings Institution projects US$10 trillion in AI outlays over the next six years. These aren’t normal numbers.

The Wall Street Journal called it “the biggest economic bet in US history, dwarfing the railroads, the highway system and the plumbing for the internet.” That’s not hype. That’s infrastructure. And like every infrastructure build that came before it, from railways to the internet, the speculative frenzy will eventually cool and the physical reality will remain, still running, still compounding its advantages for wherever it was built.

Australia is not building it.

We’ve already fumbled one wave

Here’s the cautionary tale sitting right in front of us: renewables.

For years, Australian politicians from both major parties called this country a “renewable energy superpower.” The geography supports it. The resources support it. Then last year, large-scale solar and wind investment halved to $4.4 billion. We had the rhetoric. We didn’t hold the investment.

AI is tracking toward the same pattern. Anthropic signed a lease for inference capacity at Queensland’s $30 billion Western Downs Digital Park, which sounds impressive until you read the fine print. They’re running requests there from 2027. They’re not training models there. Model training, where the economic weight actually sits, stays in the US. Australia gets the warehouse. America keeps the factory.

What this costs small businesses right now

Andrew Charlton, the Assistant Minister for Science, put a number on it at a recent ANU Crawford School lecture: Australians are spending $5 to $8 billion a year on AI tokens. Most of that flows directly to US companies. Charlton called it what it is: a new import bill. (We’ve written about this dynamic in detail: Is Your AI Subscription Funding Silicon Valley? Australia’s 80-Cent Problem.)

One in five Australians is already paying for AI. One in three Australian businesses. So the adoption is happening. The question is who captures the value from it.

Right now, the answer is mostly: not us.

Small business owners paying for AI subscriptions are contributing to that import bill every month. That’s not an argument against using the tools. It’s an argument for Australia getting serious about where those tools come from, where they’re built, and where the economic return lands.

For individual small businesses, the calculation is simpler and more urgent: you’re already competing against businesses that use these tools. The gap between those who do and those who don’t will widen faster than most people expect. Waiting to see how it plays out is a decision with consequences.

What Australia has to get right

The Hawke-Keating government didn’t just open the economy to global markets. They made a bet that Australia’s competitive strengths would prove greater than the disruption costs. They were right, eventually. The transition was brutal for some industries and they knew it going in.

The bet on AI requires something similar. Hosting data centres isn’t enough. As the renewables story shows, you can build the physical infrastructure and still miss the economic capture if the high-value work happens elsewhere. The only way Australia gets real economic weight from AI is by persuading frontier labs to train models here, not just run them here. That requires a different level of commitment than signing guardrails agreements with 19 countries, two of whom matter most and didn’t show up.

Albanese’s Labor has the same choice Hawke and Keating had. Make the bold call and absorb the short-term costs of getting it right, or manage the optics while the wave builds offshore.

Globalisation is now going into reverse. Trump’s tariffs are unwinding the very order Australia benefited from for thirty years. AI isn’t reversible in the same way. The infrastructure, once built, compounds. The country that gets the training compute gets the models, gets the products built on top of them, gets the economic weight. Australia has a window. It’s not infinite.

For small businesses: don’t wait for Canberra to sort this out before you act. The adoption curve is moving now. Get on the right side of it while the cost of entry is still low.

More hot takes: Hot Takes →

If this interests you

• Australia Signed a Global AI Regulation Deal. Here’s What It Actually Means for Your Business.
• Your Staff Have Already Moved On. Have You?
• CommBank Data: Australian Small Business AI Investment Doubled in a Year. What Are They Actually Buying?

More hot takes: Hot Takes →

Sources

• Peter Hartcher, The Age: Why Australia simply cannot afford to miss this AI wave (26 September 2026)
• Wall Street Journal: The AI build-out is becoming the biggest economic bet in US history
• Goldman Sachs: Global AI investment forecast to exceed US$1 trillion in 2026
• Brookings Institution: Financing the AI build-out, US$10 trillion projected over six years
• Andrew Charlton, Assistant Minister for Science: ANU Crawford School lecture on AI import costs
• Clean Energy Council: Clean Energy Australia Report 2026
• BloombergNEF: Energy Transition Investment Trends

📖 More deep dives: News Deep Dives

📖 More deep dives: News Deep Dives →