{"id":56576,"date":"2026-10-26T23:53:24","date_gmt":"2026-10-26T23:53:24","guid":{"rendered":"https:\/\/www.intelligentcio.com\/apac\/?p=56576"},"modified":"2026-10-26T23:53:25","modified_gmt":"2026-10-26T23:53:25","slug":"can-australias-infrastructure-keep-pace-with-ai-data-centre-growth","status":"publish","type":"post","link":"https:\/\/www.intelligentcio.com\/apac\/2026\/10\/26\/can-australias-infrastructure-keep-pace-with-ai-data-centre-growth\/","title":{"rendered":"Can Australia’s infrastructure keep pace with AI data centre growth?"},"content":{"rendered":"\n

As Australia moves towards national standards for large data centres, Nandakumaran Muthuvel, Director of Power Systems at Cundall, explains why power, cooling, water and site selection must be planned together from day one if infrastructure is to keep up with AI-driven demand.<\/em><\/p>\n\n\n\n

Australia’s plans for national AI standards are sparking a broader discussion about the infrastructure needed to enable the country’s next generation of data centres.<\/p>\n\n\n\n

The Federal Government first set out its direction in July, proposing a national regulatory framework for large data centre facilities around power supply, grid connection costs, demand management, water efficiency and site selection. National Cabinet subsequently agreed to progress a nationally consistent framework, with Commonwealth legislation intended for early 2027.<\/p>\n\n\n\n

For developers, the important point is the extent to which these requirements link digital growth with the physical systems which support it; securing land and obtaining a connection to the grid will not be sufficient, since the projects will also need a solid plan regarding where their electricity is obtained, their behaviour when the network is constrained and the protection of local water resources.<\/p>\n\n\n\n

That creates a more demanding development test, but potentially a more useful one. Australia has an opportunity to create clearer expectations for investors without allowing the disjointed approach that can emerge when energy, planning and water are considered separately. The detail will still matter, particularly because state grids and resources vary considerably.<\/p>\n\n\n\n

Power strategy starts before the connection<\/strong><\/p>\n\n\n\n

Large AI facilities involve a highly concentrated electrical demand. Because high-density GPU clusters operate continuously and can alter demand quickly, they put pressure on generation facilities, substations and transmission systems that may not have been designed to handle this level of demand.<\/p>\n\n\n\n

The load’s electrical characteristics also matter. Power-electronic equipment can include harmonics, reactive power requirements and system-strength considerations. These factors need to be assessed early in the connection process. These are not issues to address after a site has been selected; they may affect whether the proposed connection is practical in the first place.<\/p>\n\n\n\n

Timing adds another layer of risk, since a data centre can usually be designed and built faster than the grid infrastructure needed to supply it; when reinforcement is required, the electricity connection can quickly become the critical path to delivery.<\/p>\n\n\n\n

That changes the questions developers need to ask during due diligence. Available megawatts are only part of the answer. They also need to know when that capacity will be available, what network upgrades are required and how much additional generation or firming the project may need to support.<\/p>\n\n\n\n

Although renewable procurement helps, it does not eliminate local constraints. Renewable Energy Certificates and virtual Power Purchase Agreements can support investment in clean generation and emissions reporting. However, they do not ensure that electricity is available at the correct location and time.<\/p>\n\n\n\n

A more robust power strategy should connect procurement with physical delivery; in different cases this may include dedicated renewable generation, battery storage, grid supply and, where appropriate, on-site generation. The aim is not to make each campus completely electrically self-sufficient, but to align supply more closely with the facility’s operating profile and lessen strain on overloaded networks.<\/p>\n\n\n\n

Flexibility must reflect the workload<\/strong><\/p>\n\n\n\n

The Government\u2019s proposed approach may require large data centres to provide greater demand flexibility, raising an important technical question: how flexible can an AI facility realistically be?<\/p>\n\n\n\n

Storage can help answer part of it. Battery energy storage systems can absorb renewable generation, reduce peaks and support a facility over periods ranging from minutes to hours. UPS systems and supercapacitors operate on much shorter timescales and can respond to power-quality disturbances closer to the computing equipment.<\/p>\n\n\n\n

They work best when the systems are coordinated. The energy management system can decide when batteries should charge or discharge, how UPS assets should respond and whether non-essential electricity demand can be reduced. Viewing each technology as a separate layer tends to create duplication and leaves useful flexibility unused.<\/p>\n\n\n\n

Compute itself may also offer some room to manoeuvre, but the amount depends heavily on the workload. Certain non-critical tasks could be moved away from periods of network constraint. Battery charging could be delayed. Other processes cannot be shifted so easily.<\/p>\n\n\n\n

AI training programmes may be tied to customer commitments or operational deadlines. A grid-interactive data centre therefore needs to be based on a realistic assessment of which loads can move and for how long.<\/p>\n\n\n\n

It is essential to implement broad policy expectations carefully. Any requirement for demand flexibility will therefore need to reflect how individual facilities operate and which loads can realistically be shifted.<\/p>\n\n\n\n

Water, cooling and location as considerations<\/strong><\/p>\n\n\n\n

Australia’s proposed standards also place greater attention on water, an issue that becomes more important as AI changes cooling design.<\/p>\n\n\n\n

Dense GPU environments produce concentrated heat loads. Liquid cooling is increasingly used to remove that heat efficiently, but the wider cooling system still determines the facility’s overall energy and water performance.<\/p>\n\n\n\n

Since all the methods involve making concessions, evaporative cooling can reduce electricity consumption but may increase on-site water use, while dry cooling can substantially reduce water consumption at the development itself but may require more electricity, particularly during hot weather. Where this additional electricity is supplied by water-intensive thermal power generation, the reduction in on-site water use may be partly offset by greater water consumption elsewhere in the power system. Direct-to-chip liquid cooling can support high-density AI infrastructure efficiently, but its overall energy and water performance depends on how the closed-loop liquid system is integrated with the facility\u2019s wider cooling and heat-rejection systems.<\/p>\n\n\n\n

Power Usage Effectiveness is therefore an inadequate measure on its own. A facility might achieve a good PUE while still using large amounts of water in a place where that resource is already limited. On the other hand, a design that uses a little more electricity could deliver a better local environmental outcome if it greatly reduces water consumption.<\/p>\n\n\n\n

The right decision begins with the site. Developers need to understand water availability, recycled-water infrastructure, treatment requirements and climate before fixing the cooling strategy. Waste heat recovery can add value where a viable nearby user exists, but its usefulness also depends on local conditions.<\/p>\n\n\n\n

The same logic applies across the wider development. Grid capacity, renewable supply, water, land and fibre cannot be evaluated as separate checkboxes and reconciled after the site is acquired. They interact and improving one part of the system can create pressure elsewhere.<\/p>\n\n\n\n

Australia’s emerging framework recognises that relationship. If implemented well, it could encourage developers to treat large AI campuses as active participants in local infrastructure rather than passive consumers of electricity and water.<\/p>\n\n\n\n

That does not mean asking every data centre to become self-sufficient. It means setting clearer expectations around additional power supply, network contribution, realistic demand flexibility and responsible resource use.<\/p>\n\n\n\n

The projects most likely to succeed should deal with those dependencies from the start; power, cooling, water and computing facilities must be planned as a unit and the site should be chosen in light of what the surrounding infrastructure can support throughout the facility’s entire life.<\/p>\n","protected":false},"excerpt":{"rendered":"

As Australia moves towards national standards for large data centres, Nandakumaran Muthuvel, Director of Power Systems at Cundall, explains why power, cooling, water and site selection must be planned together from day one if infrastructure is to keep up with AI-driven demand. Australia’s plans for national AI standards are sparking a broader discussion about the […]<\/p>\n","protected":false},"author":21,"featured_media":56577,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1463,12018,14,7975,22,12026,1576,44],"tags":[1108,76,2229,11344,72,987,558,12241,1705],"class_list":["post-56576","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-apac","category-australia","category-data-centres","category-expert-opinion","category-industry-verticals-government","category-government","category-analysis-insights","category-top-stories","tag-ai","tag-australia","tag-cooling","tag-cundall","tag-facilities","tag-infrastructure","tag-power","tag-site-selection","tag-water"],"acf":[],"publishpress_future_workflow_manual_trigger":{"enabledWorkflows":[]},"_links":{"self":[{"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/posts\/56576","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/users\/21"}],"replies":[{"embeddable":true,"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/comments?post=56576"}],"version-history":[{"count":1,"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/posts\/56576\/revisions"}],"predecessor-version":[{"id":56578,"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/posts\/56576\/revisions\/56578"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/media\/56577"}],"wp:attachment":[{"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/media?parent=56576"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/categories?post=56576"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.intelligentcio.com\/apac\/wp-json\/wp\/v2\/tags?post=56576"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}