Author: Digital Infrastructure Institute

Data centres, networks and AI systems don’t run themselves. The harder constraint is often not capital or even energy, but people — and building the skills to design, construct, operate and govern this infrastructure is a deliberate task, and a real regional opportunity.

Discussion of digital infrastructure tends to focus on investment and energy. Both matter. But ask people actually delivering these projects what keeps them up at night, and the answer is frequently the same: workforce. The skills to build and run this infrastructure are in demand everywhere at once, and they do not appear on their own. Treating the workforce as an afterthought is one of the surer ways to slow the whole agenda down.

The skills the sector actually needs

The requirement spans more roles than the “tech” label suggests. Construction and trades — electrical, mechanical, civil — to build facilities and the energy and network connections they depend on. Operations — the smaller, ongoing, high-skill teams that keep data centres and networks running reliably and securely. And a wider ecosystem of capability: energy and grid expertise, planning and assessment, cyber security, and the policy and governance skills needed to make good decisions about all of it. Sovereign capability, in particular, depends as much on people as on hardware.

An honest picture of the jobs

It is worth being straight about the shape of the employment. Construction phases create significant, time-limited demand for trades and contractors. Ongoing operational roles at a single facility are fewer, but they are skilled, well-paid and durable. The larger workforce story is the surrounding ecosystem and the capability it builds across a region — the apprenticeships, the training pathways, the expertise that, once developed, supports far more than any one project. Overstating direct headline jobs helps no one; understating the broader capability dividend misses the point.

A regional opportunity, if planned

Where these skills are developed is not predetermined. Left to chance, expertise is imported and flies in and out. Planned for, it can be grown locally — through partnerships with TAFEs and universities, apprenticeship and traineeship pipelines, and training tied to real projects. That turns a one-off construction boom into lasting local capability, and gives host regions a stake in the sector beyond the facility itself.

Tasmania as a worked example

This is squarely where an institute’s education role earns its place, and where Tasmania has an opening. Building digital infrastructure skills locally — connecting industry need with education and training providers — would let Tasmania supply its own talent rather than relying wholly on the mainland, and would anchor the benefits of growth in the community. Capability building of this kind is one of the three things the Digital Infrastructure Institute exists to do, alongside research and convening.

Workforce is infrastructure too

The lesson is simple: people are part of the infrastructure, not separate from it. A region can have the energy, the land and the investment and still be held back for want of skills. Planning the workforce as deliberately as we plan the facilities — early, locally and in partnership with educators — is what turns digital infrastructure from a set of buildings into a lasting source of capability and opportunity.

Part of the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure.

Large infrastructure projects rarely fail on engineering. They stall on trust. Community engagement done late, or as a formality, breeds the opposition it was meant to avoid. Done early and genuinely, it builds the social licence these projects can’t proceed without.

For all the focus on energy, water and economics, the factor that most often decides whether a piece of infrastructure gets built — and whether the community is glad it did — is trust. A project with strong local support moves faster and lands better than one fighting its neighbours, however sound its technical case. That trust has a name: social licence. And it cannot be bought or assumed; it has to be earned.

What goes wrong

Most engagement failures share a pattern. The community is informed after the important decisions are made, so consultation feels like notification rather than influence. Information flows one way, leaving residents to fill the gaps with worst-case assumptions about energy, water, noise, traffic or visual impact. Promised benefits are vague or arrive late. The result is predictable: people who feel done-to rather than worked-with, and a project that spends its capital fighting avoidable opposition.

What good looks like

Genuine engagement runs on a few principles. Start early — before the design is locked — so the community can actually shape outcomes rather than react to them. Be transparent about real impacts, in plain English, including the things people worry about. Listen and adapt: show where feedback changed the project, not just that it was collected. Share benefits visibly, so hosting the infrastructure delivers something local and lasting. And treat engagement as an ongoing relationship, not a one-off consultation that ends when approval is granted. Communities can tell the difference between being consulted and being managed.

Independent information matters

One quiet enabler of good engagement is trustworthy, independent information. When the only available explanations come from a project’s proponent, scepticism is natural. Accessible, even-handed material — explaining how data centres use energy and water, what the real impacts are, and what good practice looks like — gives communities a fair basis to form their own views. Providing exactly that kind of independent, plain-English information is part of why an institute like ours exists.

Tasmania as a worked example

Tasmania’s communities tend to be smaller, closely connected, and strongly attached to place — which makes engagement both more important and more powerful. Get it wrong and word travels fast; get it right and local support can become a genuine asset. There is an opportunity here to set a high standard from the outset: to make early, honest, benefit-sharing engagement the norm for digital infrastructure, rather than something retrofitted after the first dispute.

Earned, not assumed

Social licence is not a document or a milestone; it is the accumulated result of being trustworthy over time. Projects that treat engagement as a compliance step keep relearning that lesson the hard way. Those that treat it as central — early, honest and genuinely two-way — find that community trust is not an obstacle to navigate but the foundation everything else is built on.

Part of the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure.

On the driest inhabited continent, any large new water user invites scrutiny — and data centres can use water to keep cool. But water use varies enormously with design and location. With the right choices, it can be modest, transparent and uncontroversial.

After energy, water is the resource most often raised about data centres, and for good reason: in a dry country, communities are right to ask whether a new facility will compete with households, farms or the environment for a scarce resource. The honest answer is that it depends almost entirely on how the facility is cooled and where it is built — both of which are choices, not fixed facts.

Why water comes up

Many cooling systems use evaporation to shed heat, and evaporative cooling consumes water. At scale, that can be significant, which is why “how much water?” is a fair and increasingly common question. The concern sharpens in water-stressed regions, where any large new draw on potable supply is hard to justify.

It’s a design and siting choice

Cooling sits on a spectrum. Evaporative systems use more water but less energy; closed-loop and air-cooled systems use little or no water but can use more energy, especially in hot climates. There is a genuine water-versus-energy trade-off to manage. Crucially, water source matters as much as volume: using recycled, non-potable or rainwater for cooling is very different from drawing on drinking-water supplies. And siting matters most of all — water-intensive cooling belongs where water is genuinely abundant, not in catchments already under stress.

What good looks like

A few principles keep water use defensible. Be transparent: disclose expected water use and its source, so communities and regulators can judge it. Avoid potable water for cooling wherever possible, favouring recycled or non-potable sources. Match cooling to climate: choose designs suited to local conditions rather than a one-size template. And account for water in approvals, alongside energy, rather than treating it as an afterthought. None of this is exotic; it is simply planning the resource in from the start.

Tasmania as a worked example

Tasmania’s position is favourable here. A cooler, wetter climate reduces cooling demand in the first place and eases the water-stress concerns that dominate on the mainland — and cooler ambient temperatures make low-water cooling designs more practical. That does not remove the need for care: transparency about water use and source still matters, and local catchment conditions still deserve attention. But it does mean Tasmania can pursue digital infrastructure without the acute water trade-offs facing hotter, drier jurisdictions — provided the choices are made deliberately.

A manageable concern

Water is a legitimate question, but it is a manageable one. The volume a data centre uses is not destiny; it is the product of decisions about cooling technology, water source and location. Made openly, and made with the local context in mind, those decisions turn water from a flashpoint into a non-issue — and protect the community trust that every project ultimately depends on.

Part of the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure.

A data centre is, in effect, a building that turns electricity into heat — and then spends more energy throwing that heat away. Captured and planned for, it could instead warm homes, pools, greenhouses and farms. The waste stream is real; so is the opportunity to make it a public benefit.

Almost all the electricity a data centre consumes ends up as heat. Today, most of that heat is simply rejected to the atmosphere through cooling systems — a double inefficiency, since energy is spent both creating the heat and getting rid of it. As data centre capacity grows, so does the volume of warmth being discarded. The question worth asking is not how to dispose of it, but how to use it.

What the heat can do

Data centre heat is low-grade — warm rather than hot — which shapes where it is useful. But there are many uses that suit exactly this kind of heat, especially with a heat pump to lift the temperature where needed:

• Community heating. Warming public buildings, social housing, schools or leisure centres through a local heat network, displacing gas or electric heating.
• Aquatic centres and pools. Swimming pools are energy-hungry to heat and an obvious, year-round match for a steady low-grade heat source.
• Greenhouses and horticulture. Protected cropping runs on warmth; recovered heat can extend growing seasons and lift local food production.
• Aquaculture. Maintaining water temperatures for fish and shellfish is a natural fit for a constant supply of warm water.
• Light industry. Drying, washing and pre-heating processes that need warmth rather than high heat.

Across Europe and Scandinavia, district-heating networks already draw on data centre heat to warm thousands of homes. The technology is proven; what is usually missing is the planning to connect supply with demand.

The catch: proximity and planning

Heat does not travel far economically. Unlike electricity, you cannot pipe warmth across a region without losing most of it. That makes heat reuse a planning problem more than a technical one: the data centre and the heat user have to be close, and the connection has to be designed in early. Retrofitting heat recovery onto a facility built without it in mind is difficult and often uneconomic. Capturing this opportunity means siting data centres near potential heat offtakers — or planning new public facilities, greenhouses or housing alongside them — and bringing operators, councils and utilities together before the concrete is poured.

Tasmania as a worked example

Tasmania is unusually well placed for this. Its cooler climate means genuine, year-round demand for heat — for buildings, for protected horticulture, and for the aquaculture industry that already anchors parts of the state’s economy. The same cool climate makes data centre cooling more efficient in the first place. Pairing data centres with local heat users — warming a community pool, supporting a greenhouse operation, or feeding an aquaculture facility — would turn a discarded by-product into tangible local value, and give host communities something they can see and feel.

A visible answer to “what’s in it for us”

Heat reuse matters beyond efficiency. One of the hardest questions any data centre proposal faces is what the host community actually gains. Recovered heat is one of the most concrete possible answers: a warmer public pool, lower energy bills for a community facility, a longer growing season, jobs in local food production. It converts an abstract piece of digital infrastructure into a direct contribution to local life.

Like much of this agenda, it only happens by design — when data centres and the communities around them are planned together rather than in isolation. Treated as an afterthought, the heat is wasted. Planned for from the start, it becomes a public good.

Part of the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure.

Data centres are usually cast as a problem for the grid — a heavy, growing load that strains the system. The more interesting story is the opposite: handled well, flexible data centres can stabilise the grid and make more renewable energy viable, by being the buyer for clean power that would otherwise be thrown away.

As data centre demand grows, the default narrative is one of strain: more load, more pressure on the grid, more competition for power. That concern is legitimate, and we have written about getting the energy balance right. But it tells only half the story. The other half is that large, controllable electricity loads are exactly what a renewable-heavy grid needs — and data centres, especially those running AI workloads, can be among the most flexible loads on the system.

The problem: clean energy we throw away

On a sunny, windy day with modest demand, renewable generation can exceed what the grid needs. When that happens, wholesale spot prices fall — sometimes below zero. Negative prices are a signal of abundance: there is more clean energy available than there are customers to use it. Faced with paying to generate, wind and solar operators curtail — they switch off. That curtailed energy is clean power, already built and paid for, simply spilled. Worse, the prospect of being curtailed and earning nothing weakens the business case for building the next wind or solar farm. Oversupply, paradoxically, can slow the renewable build-out we need.

The opportunity: flexible demand that soaks up the surplus

This is where flexible data centres change the picture. A large load that can ramp up when energy is abundant and cheap, and ease back when energy is scarce and expensive, is precisely the counterpart a variable renewable grid is missing. When prices go negative, a flexible data centre can absorb the surplus the grid cannot otherwise place — becoming, in effect, the customer of last resort for energy that would have been curtailed. The renewable operator earns revenue instead of switching off; the clean energy is used instead of wasted; and the data centre runs on genuinely surplus, low-cost, low-emissions power.

Not all computing is flexible — latency-sensitive services need to run around the clock. But a meaningful and growing share is interruptible or shiftable, particularly AI model training and other batch workloads, which can be scheduled to follow cheap, clean supply. On-site batteries and backup systems add a further layer, able to provide fast grid services such as frequency support. The result is a load that doesn’t just consume from the grid, but actively helps balance it.

A virtuous cycle for renewables

Put together, this flips the relationship. Flexible data centres improve grid stability by absorbing oversupply and providing demand response. By giving curtailed energy a buyer, they improve the economics of existing renewables — and strengthen the case for building more. More firm offtake means more renewable investment; more renewables, paired with flexible demand, means a cleaner and more stable grid. Designed deliberately, data centre growth and renewable growth reinforce one another rather than compete.

Tasmania as a worked example

Tasmania is well suited to demonstrate this. Its hydro system provides flexible, firm capacity, and its wind resource is strong and growing — a combination that produces exactly the kind of surplus-and-firming dynamic where flexible compute adds value. Data centres operated as interactive grid assets could absorb surplus wind when it would otherwise be curtailed, lean on hydro firming when needed, and help underwrite the next wave of renewable build — turning Tasmania’s clean energy advantage into a deeper, more durable one.

Planning for it on purpose

None of this happens automatically. It requires the right market signals and tariffs that reward flexibility, workloads designed to be shiftable, and planning that treats data centres as participants in the energy system rather than passive loads bolted onto it. That is a coordination challenge — connecting energy policy, market design and digital infrastructure so the incentives line up. Get it right, and one of the most-cited risks of the AI era becomes one of the most useful tools for the clean energy transition.

Part of the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure.

Data centres, AI factories, connectivity and renewable energy are usually planned in separate lanes, one project at a time. Treated as a single, coordinated system, they could deliver far more — for the economy, the grid and the communities that host them.

Australia is building the pieces of a modern digital economy at pace: data centres and AI compute, the networks that connect them, and the renewable energy that should power them. The trouble is that we are largely building these pieces in isolation. A data centre proposal is assessed on its own merits. A transmission upgrade follows its own timeline. A renewables project answers to a different process again. Each may be sound in isolation, yet together they can pull against one another — competing for the same grid capacity, the same land, and the same public goodwill.

The cost of planning in silos

Fragmented planning has predictable consequences. Infrastructure that should be co-located ends up scattered, so clean energy and the demand that could use it never quite meet. Connectivity lags behind the facilities that depend on it. Communities face a series of separate proposals with no overall picture of what is being built or why, which understandably breeds suspicion rather than support. And the synergies that make the whole more valuable than the sum of its parts — compute sited next to firm renewables, networks planned alongside load, benefits bundled and shared — are simply left on the table.

What coordination actually requires

Coordinated strategy is not central control; it is deliberate alignment. In practice it means a few concrete things. A shared picture: a view of where compute, energy and connectivity are headed together, so investment and approvals can be sequenced rather than collided. Policy settings that pull in the same direction: planning, energy and digital policy designed to reinforce one another instead of working at cross purposes. Genuine collaboration: government, industry and communities at the same table early, not in sequence after decisions are effectively made. And benefit by design: structuring projects so that hosting digital infrastructure delivers visible, lasting value to the regions that carry it.

Tasmania as a worked example

Tasmania is well placed to model this. With a strong renewable base, a defined set of regions, and growing interest in digital infrastructure, it is small and coherent enough to plan as a system rather than a scatter of unrelated projects. The opportunity is to coordinate renewable generation, compute demand, network capacity and community benefit into a single, intentional strategy — turning a natural advantage into a deliberate one. Done well, Tasmania could demonstrate what coordinated digital infrastructure planning looks like in practice, and what it yields: sustainable growth that communities support because they share in it.

The work ahead

Moving from competing projects to coordinated strategy is not a technical problem so much as an organising one. The capabilities exist; what is missing is the shared framework, the policy alignment and the convening that bring them together. That is deliberately where the Digital Infrastructure Institute focuses — connecting research, policy and the people who have to make these decisions, so that Australia’s digital infrastructure is planned as the coordinated, beneficial, sustainable effort it needs to be.

The final piece in the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure. Read the earlier posts on national infrastructure, sovereign AI, and the energy balance.

Data centres are among the fastest-growing users of electricity in the developed world. Handled well, that demand can drive investment and clean energy growth. Handled poorly, it strains grids, raises emissions and erodes public trust. The difference is planning.

The rise of AI has turned data centres from a niche concern into a headline energy issue. Training and running large models requires enormous, sustained computing power, and that power has to come from somewhere. As a result, communities and governments are increasingly asking a fair question: is this growth good for us, or is it simply being done to us? The honest answer is that it can be either — and which one depends almost entirely on how it is planned.

The real tension

The opportunity is genuine. Data centre investment brings construction activity, long-term operational roles, and demand that can underwrite new clean energy generation. But the pressures are just as real. Large facilities can place significant new load on electricity grids, compete for water where they use it for cooling, and concentrate benefits with operators while concentrating concerns — power prices, noise, land use — with the surrounding community. Ignoring either side of that ledger is how good projects become local controversies.

What “getting the balance right” looks like

Several principles separate well-planned growth from the alternative. Site for energy, not just land: build where clean, firm power is genuinely abundant, so new demand pulls clean generation forward rather than straining existing supply. Plan with the grid, not around it: coordinate facilities with energy planners early, so transmission and generation keep pace. Set efficiency expectations: hold new facilities to strong standards for energy and water use rather than accepting whatever is cheapest to build. And share the benefit visibly: make sure the communities hosting this infrastructure see real, lasting value — in jobs, skills, and contributions to local energy and amenity.

Tasmania as a worked example

Tasmania shows how the balance can tip towards opportunity. Its renewable energy base means new computing load can, in principle, be met with clean power rather than fossil generation — the single biggest factor in whether data centre growth is sustainable. That is a real and unusual advantage. But it is not automatic. Capturing it means coordinating data centre demand with energy planning, being transparent about water and grid impacts, and ensuring local communities share in the gains. The renewable advantage creates the possibility of beneficial growth; planning is what turns it into the outcome.

A solvable problem

The energy footprint of digital infrastructure is often framed as an unavoidable cost of progress. It is better understood as a planning challenge — one we already have the tools to manage. Where facilities are built, how they are powered, how efficiently they run, and how their benefits are shared are all choices. Made deliberately, and made together, they turn a source of anxiety into a source of advantage.

That word — together — is the crux. Data centres, energy, connectivity and community outcomes cannot be optimised one project at a time. Getting the balance right at the level of a single facility is necessary; getting it right across the whole system is what actually delivers sustainable growth — and that requires coordination by design.

Part of the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure.

“Sovereign AI” is becoming a fixture of policy conversations. Behind the term is a practical question: how much of the capability shaping our economy and society should Australia be able to build, run and govern itself?

Artificial intelligence is increasingly described as something nations need to hold “sovereign” capability in. The phrase can sound abstract, even nationalistic. It is neither. At its heart, sovereign AI is about a country’s ability to develop, deploy and govern AI using infrastructure, data and skills it can rely on — under its own laws and values — rather than depending entirely on systems owned and controlled elsewhere.

What sovereignty actually rests on

Sovereign AI is not one thing but four, stacked together. Compute: the data centres and specialised “AI factories” that train and run models. Data: the information that feeds them, and the rules governing its use. Talent: the people who build, operate and oversee these systems. Governance: the standards, policy and oversight that decide how AI is used and who is accountable. A country strong in one layer but absent in the others does not have sovereign capability — it has a dependency with extra steps.

Why it matters now

Three reasons make this pressing rather than theoretical. Resilience: if the compute, models and expertise a society depends on all sit offshore, that society is exposed to decisions, prices and disruptions it cannot influence. Security: critical AI capability is, increasingly, critical national capability, with all the protection that implies. And value: the economic returns from AI flow disproportionately to those who own the infrastructure and skills, not only those who use the applications. Sovereignty is, in large part, about keeping a fair share of that value — and that capability — onshore.

Not autarky — strategic capability

Sovereign AI does not mean building everything ourselves or walling off global technology. That would be neither possible nor wise. It means making deliberate choices about which capabilities are too important to outsource entirely, and investing in those: domestic compute powered by reliable clean energy, strong data governance, a deep talent pipeline, and clear rules that reflect Australian expectations. It is the difference between participating in the global AI system on our own terms and simply consuming it on someone else’s.

Tasmania as a worked example

The infrastructure layer is where this becomes tangible — and where Tasmania is interesting. Sovereign compute has to be powered by something, and the most defensible place to build it is where clean, firm energy is genuinely abundant. Tasmania’s renewable strengths make it a credible candidate to host sovereign AI infrastructure that is sustainable rather than simply expedient. But, as with all digital infrastructure, the advantage is only captured if it is planned for deliberately — connecting energy, networks, skills and governance rather than treating each as a separate problem.

The choice ahead

Australia will not decide whether AI matters — that is settled. It will decide how much of the capability behind it to build and govern at home, and how thoughtfully. Sovereign AI is the name for taking that choice seriously: not as a slogan, but as a question of infrastructure, energy, skills and policy that we plan for on purpose.

Part of the Digital Infrastructure Institute’s series on planning Australia’s digital infrastructure.

Data centres, networks and the compute behind artificial intelligence have quietly become as fundamental to Australia’s future as roads, ports and the electricity grid. Our planning needs to catch up.

When we picture national infrastructure, we think of physical things: highways, railways, ports, transmission lines, water. These are the systems a country deliberately plans, funds and protects, because everything else depends on them. Digital infrastructure — the data centres, fibre and mobile networks, and computing capacity that now sit beneath almost every part of modern life — belongs in exactly that category. The difference is that we have not yet started treating it that way.

The foundation beneath everything else

Banking, health, government services, logistics, energy markets and education all now run on digital infrastructure. When it is unavailable, the economy stops; when it is slow or insecure, everything built on top of it suffers. Artificial intelligence raises the stakes again. AI is, at its core, an infrastructure story: it depends on large concentrations of computing power, the data that feeds it, and the energy that runs it. As demand for that capacity accelerates, the systems that provide it move from the background of the economy to its foundations.

Why “national” is the right word

Three things make this a national-interest question rather than a purely commercial one. First, energy: large-scale computing reshapes electricity demand and, with it, the planning of grids and generation. Second, place: decisions about where this infrastructure is built influence regional investment, jobs and development for decades. Third, sovereignty and security: who owns and operates critical compute, and under whose rules, is a question no serious country can leave entirely to chance. These are precisely the considerations we apply to other national infrastructure — and precisely why digital infrastructure deserves the same strategic attention.

Tasmania as a test of the idea

Tasmania makes the point concretely. Its renewable energy advantage — abundant hydro and wind — gives it a genuine opportunity to host digital infrastructure that is both economically valuable and genuinely sustainable, a combination that is hard to find elsewhere. But that advantage is only realised if it is planned for: if energy, connectivity, land use and community benefit are considered together, and early. Left to chance, the same advantage can be squandered through poorly sited projects, grid strain and lost public trust. The opportunity is real; so is the need to plan for it.

From individual projects to national planning

Today, digital infrastructure is largely approached one project at a time — assessed, debated and approved in isolation. That made sense when these were marginal facilities. It makes far less sense now that they shape energy systems, regional economies and national capability. Treating digital infrastructure as national infrastructure means planning it deliberately: coordinating it with energy and connectivity, weighing the public interest alongside private investment, and giving communities a clear stake in the outcome.

That shift — from competing projects to coordinated strategy — is the harder, more important conversation, and it is the one Australia now needs to have. At the Digital Infrastructure Institute, it is the conversation we exist to advance.

This is the first in a short series on planning Australia’s digital infrastructure. The next pieces look at sovereign AI, the energy balance, and what coordinated strategy actually requires.