JEAN DONG AND CHRISTOPH NEDOPIL |
Positioning Australia as a global hub for secure, trusted, low-emissions artificial intelligence data
At a recent business lunch, one of us posed a question to the audience: is a renewable energy target of 80% by 2035 ambitious for Australia? Most in the audience believed yes. When asked whether this target strengthens Australia’s competitiveness, the reaction was more mixed. We see it differently: the target should be higher to drive Australia’s competitiveness to utilise a transformative opportunity offered as the world enters an AI-driven era. Australia is uniquely positioned to utilise its energy wealth to become a superpower in smart exports based on reliable, secure, and low-carbon data processing and infrastructure.
Imagine a future where data processed and stored in Australia carries the label “Processed in Australia”—signalling to global partners that these services are built on strong legal protections, low-emissions infrastructure, and governance anchored in democratic values. Just as “Made in Australia” has long stood for quality in goods, this new designation could represent confidence in AI systems that are secure, accountable, and rights-respecting. With the right policy foundations, Australia’s future export identity can be anchored in its legacy of resource reliability, green based on its energy, and trusted based on democratic governance of digital infrastructure.
As artificial intelligence becomes a foundational force in global economic and technological competition, the infrastructure required to support it—particularly energy-intensive data centers—has become a strategic priority. Countries are now racing not only to develop AI capabilities but to host the physical infrastructure that powers them. Blackstone estimate that the US will see over $1 trillion invested in data centers over the next five years, with an additional $1 trillion invested internationally.
Data centres are highly energy-intensive, requiring large and stable electricity inputs to power computation and cooling systems. Estimates vary, but in aggregate, global data centre electricity use is projected to more than double by 2030, reaching approximately 945 terawatt-hours (TWh)—comparable to Japan’s total annual electricity use today. Where this demand is met with fossil fuels, it could generate 500–800 MtCO₂ per year[i]—comparable to the annual emissions of global aviation or one-quarter of the steel industry—and consume up to 2% of the world’s remaining carbon budget to limit warming to 1.5°C.[ii] In parallel, the concentration of sensitive data and computational capacity in data centres makes them increasingly attractive targets for malicious cyber activity, physical sabotage, or geopolitical coercion.
Australia is uniquely well-positioned to mitigate both risks while contributing strategically to the global AI economy.
First, Australia possesses world-class renewable energy advantages, including some of the planet’s highest solar irradiance levels, abundant wind resources, and vast available land for clean energy projects. As global AI services increasingly seek low-carbon compute, Australia can deliver green power at scale. Australia’s data centers are already achieving industry-leading energy efficiency. Thus, rather than adding 50GW of solar and wind by 2035 to meet the 85% percent target, let’s utilise falling cost and add 10 GW per year (as a comparison: China added 360 GW in 2024 alone). Investments in long-duration energy storage—including batteries, green hydrogen, and thermal solutions—can further accelerate a renewable energy momentum to enhance grid reliability and enable continuous AI operations.
Second, Australia’s proximity to ASEAN, combined with new high-speed subsea cables like the Darwin-Jakarta-Singapore link and Google’s planned “Australia Connect,” positions it as a low-latency, high-capacity computing hub for the Indo-Pacific. With Sydney, Perth, and Darwin now within millisecond reach of Singapore and Jakarta, Australia offers “edge-adjacent” infrastructure—close enough to serve Asia’s AI inference demand in real time, yet with more space, energy, and stability than capacity-constrained hubs like Singapore or Hong Kong.
Thirdly, Australia is a trusted democracy with robust cyber and legal protections. Australia offers a high-trust, stable regulatory environment for AI infrastructure, with strong privacy laws, cybersecurity protections, and clear frameworks for critical data assets. Its commitment to digital sovereignty and ethical AI—backed by the Critical Technologies Policy—makes it an attractive destination for companies seeking secure, responsible AI inference at scale.
It’s not just cities and tech hubs that can power Australia’s AI ambitions—old mines have a role to play too. With existing grid connections, transport access, and naturally cooler environments, former mine sites are well suited for conversion into secure data centres, experimental facilities, or AI research hubs. Their geographic isolation provides a layer of protection against physical and cyber threats, while nearby solar generation and green hydrogen or battery storage can offer reliable, low-emissions energy. Repurposing these sites opens up new business models for mine rehabilitation, supports regional economic diversification, and as demonstration platforms for frontier technologies, such as kinetic or thermal energy storage. This not only can turn legacy industrial sites into assets for Australia’s digital and clean technology future but also contributes to innovation in clean infrastructure made in Australia.
The case is clear: sovereign, green data infrastructure can position Australia at the forefront of global decarbonisation and digitalisation. But this opportunity demands more than vision—it requires coordinated action. Australia must deliver clear rules on digital infrastructure, data security, and emissions, invest in a skilled AI and data workforce, and show partners it can lead in responsible, inclusive digital development. Crucially, natural gas should not be part of Australia’s long-term clean infrastructure strategy. Its high cost, limited flexibility, and misalignment with net-zero goals make it an unsustainable option. Only green energy—solar, wind, batteries, and green hydrogen—can power the infrastructure needed to compete globally, drive innovation, and anchor a national strategy linking energy, trade, AI governance, and investment. A national strategy that links energy policy, digital trade, AI governance, and foreign investment could unlock this next generation of export leadership.
Jean Dong is a Visiting Fellow at the Harvard Kennedy School and Professor Christoph Nedopil is the Director, Griffith Asia Institute.
[i] Fossil-fueled power generation emits a high amount of carbon dioxide per unit of electricity. Coal-fired power plants emit roughly 820 g of CO₂ per kWh generated, while natural gas power plants emit around 450 g CO₂ per kWh according to natureoffice.com. Using these typical emission factors:
- If all 945 TWh were coal-fired: CO₂ emissions would be on the order of 775 million metric tons (MtCO₂).
- If all 945 TWh were gas-fired: CO₂ emissions would be roughly 425 MtCO₂.
[ii] Based on the IPCC estimate of ~250 GtCO₂ remaining carbon budget for a 50% chance of limiting warming to 1.5°C (as of 2023). If data centers emit 0.5–0.8 GtCO₂ annually and sustain this over a decade, they would cumulatively consume 5–8 GtCO₂—equivalent to 2–3% of the remaining budget.
