Here are our predictions for 2026 on how Data Centre environments will evolve and what you will need to do globally and in Australia. Artificial intelligence (AI) workloads and real-time digital services are reshaping the physics, economics, and regulation of data centres. By 2026, firm power access, liquid cooling, higher voltage/DC distribution, and granular sustainability reporting will define competitive advantage. In Australia, these trends meet a unique mix of grid constraints, strong privacy/cyber regulation, and rapid 5G/edge expansion. This article distils our predictions for 2026 and the practical steps enterprises should take to stay ahead.
No.1. Power Availability, The Critical Challenge for Data Centres in 2026
OK let’s get the Elephant in the Room out there straight away. As AI adoption accelerates, Data Centres face an unprecedented surge in electricity demand. The International Energy Agency projects global Data Centre power consumption to more than double by 2030, making grid access the single biggest constraint for new capacity. Power availability is now the gating factor for site selection, driving operators toward hybrid solutions such as onsite generation, microgrids, and firm power purchase agreements (PPAs) to reduce “time to power.”
In Australia, the Australian Energy Market Operator (AEMO) has begun forecasting Data Centre demand separately in its National Electricity Market outlook, anticipating significant growth through 2034–35. Reliability will depend on timely delivery of generation, storage, and transmission infrastructure.
What enterprises must do – Secure firm PPAs early, explore behind-the-meter battery storage, and consider hybrid onsite generation. Align expansion plans with AEMO’s ESOO reliability projections to mitigate grid constraints and ensure sustainable growth.
No.2. Liquid Cooling, The New Standard for AI-Driven Data Centres
The rise of AI workloads is transforming Data Centre design. Average rack densities have doubled in recent years, with AI clusters now reaching 60–120 kW per rack, and next-generation systems pushing even higher. Traditional air cooling, which typically caps out at around 20 kW per rack, can no longer keep pace. This shift makes direct-to-chip liquid cooling essential for AI training and inference environments.
Industry surveys, including Uptime’s 2025 report, confirm that adoption is accelerating, driven by both performance and sustainability goals. Liquid cooling not only handles extreme heat loads but also enables warmer water temperatures, reducing chiller energy consumption and improving overall efficiency.
Action for enterprises, Design for hybrid cooling (air and liquid), integrate coolant distribution units (CDUs), manifolds, and quick-disconnects, and specify facility loops that support higher temperature ranges. Future-proofing your infrastructure today ensures readiness for tomorrow’s AI-driven demands.
No.3. High-Density AI Racks Drive a Shift to High-Voltage Direct Current Power
The next wave of AI infrastructure is rewriting the rules of Data Centre design. Rack-scale AI systems, such as NVIDIA’s GB200 and GB300 NVL72, are engineered for liquid cooling and extreme density, pushing 120–140 kW per rack. This surge in power demand is forcing a rethink of traditional AC distribution. Industry leaders are now moving toward 800 V DC architectures, which significantly reduce copper requirements, cut conversion losses, and improve overall efficiency.
For operators, this isn’t a distant trend, it’s an imminent design imperative. High-voltage DC enables streamlined builds and better scalability for AI-driven workloads, but it also introduces new safety and compliance considerations.
Action for enterprises, Evaluate 800 V DC busways and protection schemes in upcoming projects, ensure DC-aware monitoring and commissioning, and engage early with electrical authorities to meet standards. Future-proofing power distribution today will be critical for tomorrow’s AI-ready data centres.
No.4. Water and Energy Efficiency
Cooling remains one of the largest energy consumers in Data Centres, accounting for 30–40% of facility energy in traditional builds. As AI workloads drive higher densities, operators are turning to liquid cooling, which enables warmer water temperatures and can eliminate mechanical chillers in many climates. This shift dramatically improves sustainability metrics, including Water Usage Effectiveness (WUE). NVIDIA reports up to 25× energy efficiency and 300× water savings for liquid-cooled rack-scale systems compared to conventional air-cooled designs.
With regulators and investors demanding transparency, water and energy performance are now board-level priorities.
Action for enterprises, In water-stressed regions, prioritise dry coolers and closed-loop systems to minimise potable water use. Implement metering and publicly report WUE alongside PUE to meet compliance and ESG expectations. Designing for efficiency today ensures resilience and sustainability for tomorrow’s AI-driven Data Centres.
No.5. Sustainability Reporting Goes Global
From 2024, the EU’s revised Energy Efficiency Directive (EED) and Delegated Regulation mandate annual reporting for data centres over 500 kW. Operators must disclose key metrics such as PUE, WUE, renewable energy share, and waste-heat reuse to a central database. In 2025, the Corporate Sustainability Reporting Directive (CSRD) raises the bar further, requiring audited ESG disclosures and detailed Scope 3 emissions reporting. These frameworks set a global benchmark that multinationals, and their supply chains, cannot ignore.
Australia, Boards and investors increasingly expect credible, audit-ready ESG metrics, even without local mandates.
Action for enterprises, Implement metered, verifiable KPI reporting, map Scope 3 emissions, and prepare CSRD-aligned disclosures if operating in or selling into the EU. Proactive compliance not only mitigates regulatory risk but also strengthens sustainability credentials in a competitive market.
No.6. Edge Computing Surges to Power Real-Time AI and 5G
Enterprises are moving compute closer to users to achieve sub-10 ms latency for critical applications in manufacturing, healthcare, finance, gaming, and telecommunications. This shift is driving double-digit growth in edge computing investments, with colocation providers building micro and edge data halls near population centres and carrier hubs.
Australia, With 5G coverage expanding and strict data sovereignty requirements, demand for localised edge nodes in NSW, Victoria, and Queensland is set to rise sharply. These nodes will enable real-time AI inference and analytics, supporting next-generation digital services.
Action for enterprises, Plan compact, energy-efficient edge footprints equipped with liquid cooling, resilient power systems, and automated lifecycle management. Integration with core and hyperscale regions is essential to ensure seamless performance and scalability.
No.7. “Clean Firm” Power
Data Centres face mounting pressure to secure reliable and sustainable energy. Operators are increasingly blending renewables with storage and exploring innovative solutions such as carbon capture and storage (CCS) – enabled gas and nuclear PPAs (but NOT in Australia YET!). A notable example is Google’s Illinois CCS project, designed to capture ~90% of CO₂ emissions and store them in saline aquifers, signalling a bold shift toward “clean firm” power strategies.
Analysts predict Data Centres will account for a significant share of global electricity demand growth, influencing market design and grid planning worldwide.
Australia, Enterprises should leverage the National Electricity Market’s growing pipeline of utility-scale renewables and storage, while deploying behind-the-meter batteries to firm up supply and participate in FCAS and ancillary markets.
No.8. Talent Shortages Accelerate Automation and Software-Defined Power
The Data Centre industry faces a dual challenge, rising power constraints and a growing management talent gap. As AI-driven workloads demand complex infrastructure, operators are turning to automation and advanced tools to maintain resilience and compliance. Gartner’s 2025 outlook highlights Data Centre Infrastructure Management (DCIM) and microgrids as strategic enablers for AI-ready operations, while regulatory frameworks require granular reporting and predictive capabilities.
Manual processes can no longer keep pace with high-density environments and sustainability mandates.
Action for enterprises, Invest in DCIM platforms that model thermal and high-density scenarios, integrate real-time energy market signals, and automate capacity planning. Ensure these systems generate audit-ready ESG outputs, including PUE, WUE, renewable share, and heat reuse, to meet compliance and investor expectations. Embracing automation today will bridge the talent gap and future-proof operations for tomorrow’s AI-driven Data Centres.
We will review these predictions this time next year and see how accurate we were.
This BARM DC thought leadership piece explains by 2026, Data Centres will be transformed by AI workloads, driving the adoption of firm power access, liquid cooling, high-voltage/DC distribution, and granular sustainability reporting, with Australia facing unique challenges from grid constraints and regulatory demands. Enterprises must secure reliable power, embrace advanced cooling and automation, and prepare for global ESG compliance to stay competitive and resilient in the evolving digital landscape. (www.barmdc.com)
At BARM DC, we specialise in designing, optimising, and migrating Data Centre and IT environments that deliver maximum efficiency and resilience. From energy-conscious fit-outs to advanced cooling strategies and performance tuning, our team ensures your infrastructure is ready for the future, reducing costs, improving sustainability, and supporting business growth. Whether you’re planning a new build, upgrading existing systems, or you need to review your current environment, we provide end-to-end expertise to help you achieve your goals with confidence.