Energy and sovereignty will decide Australia's AI future, not algorithms

Australia is investing heavily in digital infrastructure, yet risks becoming structurally dependent on global technology systems it does not control, writes Paul Budde.

Deloitte’s latest TMT Predictions 2026 Australia report confirms what has been quietly emerging over the past few years: the AI revolution is no longer about software innovation — it is about infrastructure, energy, and control.

For Australia, this represents a critical strategic moment. The country is investing heavily in digital infrastructure, yet risks becoming structurally dependent on global technology systems it does not control.

AI is now an energy problem

One of the most important insights from the report is the scale of energy demand driven by AI. Data centre capacity in Australia is expected to more than double, potentially reaching 6.5 GW, with tens of billions of dollars in private investment.

At the same time, AI servers alone could increase power consumption by 1.7 times, with data centres consuming more than 2.5% of the nation’s total energy supply in the near term.

This fundamentally shifts the AI debate. What was once framed as a question of innovation and productivity is now a question of energy systems, grid capacity, and national planning.

This directly reinforces the argument made in my recent article, "Power to the people: Rewiring Australia’s energy future." There, I argued that digital infrastructure – particularly AI and data centres – has become a primary driver of energy demand, yet policy frameworks still treat energy and digital development as separate domains.

Power to the people: Rewiring Australia’s energy future

Australia’s energy future won’t be built by corporations alone, it’ll be co-created by communities, technology and trust.

Deloitte’s analysis makes clear that this separation is no longer viable. AI is not just another industry layered onto the economy; it is reshaping the underlying energy architecture. Without integrated planning, Australia risks power supply bottlenecks, infrastructure rollout delays, and rising costs that will ultimately undermine competitiveness.

The illusion of frictionless AI

At the same time, the report highlights the rapid adoption of 'agentic AI' — embedded systems automating tasks across enterprise software. By 2026, more than 40% of applications are expected to include such capabilities.

Yet Deloitte introduces a critical warning: up to half of organisations may experience 'AI agent sprawl,' leading to duplication, data silos, and governance risks.

This challenges the prevailing narrative of seamless productivity gains. As I have argued in earlier work on AI and societal impact, technology does not automatically translate into efficiency. Without governance, coordination, and institutional maturity, complexity increases rather than decreases.

In other words, the real challenge is not deploying AI — it is managing it.

The sovereignty gap

Perhaps the most strategically significant theme in the report is the growing importance of technology sovereignty.

Australia faces rising semiconductor costs, constrained global supply chains, and continued dependence on offshore manufacturing for critical digital components. Compute costs alone are expected to rise by around 15% due to chip constraints.

Frankenstein and AI: Innovation without accountability

The risk of AI is not innovation, but abandonment.

This aligns closely with my earlier Independent Australia article on digital sovereignty, where I highlighted the risks of relying on foreign-owned cloud platforms and digital infrastructure. The recent case involving restrictions on international digital services only underscores how quickly geopolitical tensions can translate into technological vulnerability.

Deloitte’s findings deepen that concern. While Australia is investing heavily in data centres – the visible layer of the digital economy – it remains dependent on global supply chains for the invisible but critical layers: semiconductors, cloud platforms, and AI models.

This creates a structural imbalance. Australia may host the infrastructure, but it does not control the system.

A new geopolitical layer

The report also points to expanding critical minerals agreements and strategic partnerships linked to semiconductor supply chains.

This highlights a broader shift: digital infrastructure is becoming part of geopolitical strategy. Control over chips, data, and energy is now as important as traditional resource security.

In this context, Australia’s position is paradoxical. The country is rich in the raw materials needed for the digital economy – lithium, rare earths, and energy resources – yet lacks the downstream capabilities to translate these into technological sovereignty.

This mirrors patterns seen in earlier resource cycles, in which value is created offshore while domestic economies capture only part of the benefits.

Australia’s AI paradox: Mass adoption, minimal strategy

As generative AI becomes an everyday tool for Australians, new OECD data exposes a dangerous gap between mass uptake and the absence of any serious national strategy to govern it.

The real policy challenge

Taken together, the Deloitte report reveals a convergence of issues that are too often treated in isolation:

AI is driving exponential growth in energy demand

Infrastructure investment is accelerating without full system integration

Sovereignty risks are increasing as dependence on global platforms deepens

This convergence requires a fundamentally different policy approach.

Australia needs an integrated strategy that links energy policy, digital infrastructure, and national sovereignty. Without this, the country risks building a highly advanced digital economy that remains externally dependent and strategically vulnerable.

Conclusion

The AI debate has moved beyond algorithms and applications. It is now about who controls the infrastructure, who supplies the energy, and who owns the underlying systems.

Australia still has the opportunity to position itself as a leader in this new landscape. But doing so will require moving beyond fragmented policy settings and short-term market responses.

The real question is no longer whether Australia participates in the AI economy. It is whether it does so on its own terms — or someone else’s.

Paul Budde is an IA columnist and managing director of independent telecommunications research and consultancy, Paul Budde Consulting. You can follow Paul on Twitter @PaulBudde.