AI and the Australian IT Leader: The Imperative for Sustainable Innovation
Listen to an Audio summary of this blog post HERE [16 Mins]
Artificial intelligence (AI) is rapidly emerging as a defining force of the decade, transforming everything from trading floors to hospital wards, and reshaping how Australian organisations compete and innovate. Yet, alongside this technological leap, climate change presents another urgent defining force, and AI’s accelerating climate footprint demands immediate attention.
For Australian IT leaders—Chief Technology Officers, Chief Information Officers, and data centre designers—sustainable AI adoption is no longer a niche concern; it is a critical strategic imperative. The challenge is to pursue innovation through sustainability, ensuring that our digital evolution supports, rather than undermines, our climate commitments.
AI & Climate - Australian Climate Leaders Coalition - B Team Australasia
Here is the essential case for Sustainable IT, focusing on the Why, What, and How for Australian leadership.
The Why: Confronting AI's Footprint and Seizing the Competitive Edge
As global AI investment is forecast to exceed US$600 billion by 2028, driven largely by generative and agentic AI, its appetite for computational power is scaling rapidly, bringing serious implications for environmental sustainability.
The Environmental Challenge
Energy and Emissions: AI is inherently energy-intensive. Global data centre energy demand could double to approximately 945 TWh by 2030, increasing associated emissions to 300 Mt by 2035. In Australia, data centres currently consume 2% of the National Electricity Market (NEM) total electricity, a figure expected to triple to 6% by 2030 and many times more in concentrated areas of Melbourne and Sydney.
Water Consumption: Data centres require vast amounts of water for cooling. By 2027, global AI is projected to consume between 4.2 and 6.6 billion cubic metres of fresh water annually, which is equivalent to roughly one-third to one-half of Australia’s total annual water consumption the driest inhabited continent.
Resource Use and Waste: The hardware for powerful new AI models relies on critical rare earth minerals, which are finite and environmentally costly to extract, leading to habitat destruction and pollution. Furthermore, the rapid requirement for next-generation chips accelerates hardware obsolescence, contributing to Australia's already high rate of e-waste generation—around 25 kilograms per person each year.
The Business Imperative
Organisations must balance the competitive advantages of AI with its potential environmental cost. Ignoring sustainability places organisations at risk of:
Competitive Disadvantage: Missing out on AI-enabled efficiencies and innovation opportunities.
Regulatory Non-Compliance: Australian organisations must anticipate new requirements. For instance, the Australian Government has mandated that, from mid-2025, data centres providing services to the federal government must achieve a 5-star rating from the National Australian Built Environment Rating System (NABERS), or an equivalent Power Usage Effectiveness (PUE) of 1.4 or less.
Reputational Damage: AI adoption without sustainable practices can lead to increased carbon emissions, resource depletion, and negative community impacts.
The What: Defining Sustainable AI
Sustainable IT is about taking a deliberate approach to evaluate the environmental impact of your technology choices, looking beyond simple performance metrics and price.
Sustainable AI, when designed and deployed responsibly, can also be a powerful tool to address climate change by accelerating decarbonisation and strengthening grid resilience. It means embedding ethical and environmental considerations into AI strategies from the start.
This holistic approach requires Australian organisations to examine how technology providers manage their data centres, source energy, and handle electronic waste. A strong track record in sustainability, transparent reporting, and ongoing efficiency innovation are all critical factors in procurement.
AI as an Enabler for Efficiency
When implemented responsibly, AI brings unique strengths to sustainability:
Grid Optimisation: AI can help forecast energy demand, allowing grid operators to integrate renewables more reliably, potentially unlocking up to 175 GW of transmission capacity globally.
Cooling Efficiency: AI-powered energy management can significantly reduce cooling energy in data centres by up to 40%.
Resource Optimisation: AI can analyse complex systems, predict environmental impacts, and optimise resource use across the value chain. For instance, AI algorithms can help identify low-carbon alternatives for materials or optimise delivery routes for lower emissions.
The How: Actionable Strategies for IT Leaders
To build a smart, efficient, and responsible AI infrastructure—the foundation of a durable competitive advantage—IT leaders should focus on three key areas: Infrastructure, Model Optimisation, and Governance.
1. Infrastructure and Data Centre Design
The convergence of AI and sustainability means bringing the engine rooms of data centres into the boardroom to spark new conversations about strategy and procurement.
Embrace Best Practice Design: Implement best practice data centre design, which includes efficient cooling methods, low-carbon energy sources, and sustainable materials.
Innovate Cooling and Water Management: Prioritise innovations in water efficiency. This means transitioning away from traditional water-intensive cooling toward solutions like air-assisted liquid cooling or direct-to-chip liquid cooling plates, which use closed-loop systems and minimise the need for fresh water.
Decarbonise Operations: Use renewable energy sources, such as solar or wind and ensure its 24/7 Carbon Free power traded and certified hourly.
Optimise Workload: Use dynamic workload balancing to shift computing tasks to times or locations where cleaner, cheaper renewable energy is available.
Adopt Circularity: Apply circular economy principles to hardware. This includes maximising server load, maintaining equipment to extend its lifespan to amortise embodied carbon, and participating in robust recycling programs for e-waste.
2. AI Model and Software Optimisation
Efficiency gains are needed not just in hardware but in the AI models themselves.
Implement Efficient Models: Implement energy-efficient AI models that minimise computational demands. Strategies include model pruning (eliminating redundant parameters) and quantisation (reducing calculation precision) to cut energy use without compromising performance.
Choose Smaller Models: Adopt Small Language Models (SLMs) which can achieve remarkable performance while being far less power-hungry than larger counterparts.
Measure Impact: Use tools like open-source carbon estimator tools (e.g., CodeCarbon) to assess the emissions intensity of models, guiding better decision-making during development and model selection.
3. Governance and Procurement (The C-Suite Call to Action)
The transformation requires leadership at the highest level. C-suite and technical leaders should focus on six key principles when choosing AI providers:
Evaluate Environmental Impact Holistically: Assess providers based on their full data centre operations, covering energy, water usage, and e-waste management.
Look for Verified Commitment: Prioritise providers that demonstrate strong, verified commitments to sustainability, such as alignment with NABERS, NGERS (National Greenhouse and Energy Reporting Scheme) reporting and ISO 14064 climate reporting, ISO 14067 product carbon footprints & ISO 14068 transitioning to net-zero.
Prioritise Innovation: Seek out partners who actively innovate in energy and water efficiency, developing sustainable AI models and retrofitting existing infrastructure.
Seek Collaborative, Transparent Partnerships: Favour providers open to sharing data and insights, adhering to ethical standards, and collaborating on sustainability projects.
Embrace Ecosystem Approaches: Leverage partnerships to source and deploy best practices, recognizing that no single organization can master this evolving field alone.
Embed Sustainability into Responsible AI Governance: Build sustainability into your Responsible AI strategies, including training users on efficient prompting and resource-aware usage to minimize unnecessary emissions.
The decisions made now regarding AI adoption will shape tomorrow’s legacy. By integrating sustainable practices at every layer of the tech stack—from governance and procurement to data centre design and model optimisation—Australian IT leaders can ensure that the rapid advancement of AI serves as a catalyst for a more sustainable, efficient, and resilient future.
-END-
References
AI for Climate - Use Case and Solution Mapping Framework [AU CLC May 2025]
AI and Scope 3: precision on the path to net-zero emissions [AU CLC May 2025]
AI Climate and Environment Strategies to Reduce Impact [AU CLC Sept 2025]
Australian Climate Leaders Coalition Website
NABERS (National Australian Built Environment Rating System) - Data Centres