There is nothing new under the sun: A perspective on data centres

Anyone who has spent the past two decades working on the edges of Australia's cities, or in its regional growth corridors, has watched the same landscape get rewritten several times over. Outer suburbs have absorbed a generation of new residents, often faster than the roads, schools, health services, and, not infrequently, the electricity network could keep pace with.

Old industrial land has emptied out as manufacturing has moved offshore or simply wound down, leaving behind sites and communities looking for a new economic purpose. Onto that same land, and often the same regions, a new generation of energy infrastructure has arrived: wind and solar farms transforming farmland into generation zones, batteries and pumped hydro changing what a substation site even is, and transmission and distribution networks stretched and augmented to carry it all to where the people are. Anyone who has been closely involved in any of this – planning it, financing it, delivering it, or living next to it – would offer the same honest assessment: governments and businesses got a good deal of it right, and got a meaningful part of it wrong.

Housing released well ahead of the infrastructure to support it. Generation and transmission buildouts that delivered the megawatts but not, in enough cases, the community consent that should have come with them. Land-use change that outpaced the institutions meant to manage it.

Data centres, in many ways, are simply the newest entrant into a landscape with a long, well-documented history of getting infrastructure-led change partly right and partly wrong. There is nothing structurally new in what is unfolding: concentrated, capital-intensive, infrastructure-hungry development landing in outer-suburban and regional communities that are still absorbing the last two or three waves of change.

What matters now is whether this wave gets built into that history as one of the things governments and industry got right, or one more entry on the list of things done in haste and regretted for many years thereafter. That is squarely a live choice, not a foregone conclusion, and it is one energy businesses are better placed than almost anyone else to influence.

For energy businesses specifically, doing this well rather than poorly comes down to a small number of disciplines.

1. Recognising the asymmetry they carry relative to their data centre customers

A developer can walk away from a poorly chosen site, a network post making supporting investments cannot, and so the bar for site and partner selection has to be set by the party left holding the consequence. Treating site selection itself as the single highest-leverage decision available, steering load toward precincts with latent capacity and existing infrastructure, rather than processing applications on a first-come basis and absorbing whatever social licence fallout follows.

2. Understanding that the community trust being spent on data centre connections is drawn from the same finite account as the trust needed for REZ and transmission delivery

Mishandling one makes the other, more important job harder. Applying the benefit-sharing and engagement discipline already built for renewable energy zones, rather than treating data centres as a purely commercial workstream sitting outside that experience. Making the fairness of cost allocation visible and legible to the communities and customers who will judge it, not just technically defensible to a regulator.

3. Recognising that doing this well is not the slower or more cautious path

It is rather the one path that avoids the delay, opposition and cost blowouts that have already slowed transmission and wind build-outs elsewhere, and that would slow this wave too if the same mistakes are repeated.

Start with the asymmetry, because it explains why this is a network problem in a way it isn't equally a developer problem. A data centre operator that chooses badly – the wrong site, the wrong community, the wrong level of engagement – has an exit available that no network shares. Projects get shelved, sold, relocated to a more accommodating jurisdiction, or quietly abandoned once the reputational cost outweighs the commercial case. The network underwriting that connection has no equivalent exit. The substation is built.

The same holds when a line is augmented. The community relationship, good or bad, belongs to the network for the decades the asset sits in the ground, long after the developer that triggered it has moved on to the next site or the next cycle of capital. That asymmetry is the reason networks cannot treat data centre connections as a standard commercial transaction and expect the risk to behave like one. The bar for judgement – on siting, on partner selection, on the terms attached to a connection offer – has to be set by the party that cannot leave, not the party that can.

If that asymmetry is accepted, the highest-leverage decision a network makes is not embedded in the connection agreement at all, but rather embedded in where the load is steered before an agreement is ever drafted.

The evidence on this is already unambiguous. Data centres proposed for brownfield sites carrying existing substations, existing grid connections, and a community accustomed to industrial use are clearing planning hurdles and avoiding public opposition with comparatively little friction. Proposals landing on greenfield sites, requiring fresh capacity to be built into communities that see only cost and disruption with no history of industrial tolerance, are the ones generating objections, local political campaigns and, in at least one state, legislative attempts at a moratorium.

This is close to a binary outcome, and it is a variable that networks have real influence over. A first-come, first-served approach to connection applications treats siting as the developer's problem to solve. A precinct-led approach – of the kind that Ausgrid, Endeavour Energy and Essential Energy have already begun advocating for – treats siting as the network's opportunity to shape, directing capital toward land that already carries the infrastructure and the social tolerance to absorb it, and away from land that doesn't.

It is tempting to treat data centre engagement and renewable energy zone engagement as separate workstreams, run by different teams, answering to different parts of the business. They are not separate. Both draw on the same underlying asset: the willingness of communities and their elected representatives to accept large infrastructure, disruption and change in exchange for a case they find credible.

That willingness is not unlimited, and it does not reset cleanly between projects. A community that feels a data centre connection was imposed on it without consultation or visible benefit will bring that experience, and that scepticism, to the next transmission line or renewable energy zone proposal put in front of it, regardless of which part of the network's business is asking.

Networks spending down that trust for a data centre connection today are borrowing against the harder, longer-dated project that the energy transition still needs from the same communities tomorrow. Handled well, the reverse is also true: a data centre program that visibly delivers local benefit can build, rather than deplete, the reserve of trust the rest of the transition depends on.

The reassuring part of this argument is that none of it requires networks to invent a new capability. The discipline of community benefit-sharing, local workforce development, and genuine early engagement has already been built, tested, and refined through a decade of renewable energy zone delivery. (We have written about the necessary step change in community engagement and social value creation – what we call ‘Community Engagement 3.0’ – previously.) Apprenticeships and training commitments tied to construction and operations phases, transparent engagement at the earliest stages of project conceptualisation, and benefit-sharing arrangements that give a community a tangible stake in a development it is being asked to host.

This is not theoretical best practice, it is the accumulated, hard-won experience networks already hold from REZ work. The task is not to design a new model for data centres. It is to insist, deliberately and as a matter of policy rather than case-by-case judgement, that the REZ model is the default starting point for data centre engagement too, rather than allowing the commercial urgency of a data centre connection to bypass discipline that took years to build elsewhere in the business.

There is a version of this argument that positions social licence as a tax on speed. The careful, consultative path against the fast, commercially aggressive one. The evidence does not support that framing. The delay, cost blowout and political risk already visible in transmission and wind development in Australia did not come from moving too carefully.

It came from moving without sufficient regard for the communities being asked to host the change, and paying for that shortfall later in the form of legal challenges, moratoria, and projects stalled indefinitely in the planning system. Data centres are already showing the same early symptoms: legislative moratorium attempts, local members of parliament intervening on behalf of residents, community coalitions setting public conditions for acceptance before a shovel is in the ground.

None of that is the fast path. The fast path, over any realistic multi-year horizon, is the one that spends the time up front to make a project legitimate in the eyes of the community hosting it, because that is the version of the project that does not get stopped halfway through.

Put together, doing this well is not a matter of goodwill or corporate messaging. It is a small number of concrete operating choices: a precinct-led connection strategy rather than an application-led one, benefit-sharing and local workforce commitments built into standard connection terms rather than negotiated ad hoc under commercial pressure, cost allocation that is not just regulator-defensible but genuinely legible to the households and businesses who will judge whether the boom helped or hurt them, a governance discipline that treats data centre engagement as drawing on the same finite account of community trust as every other piece of transition infrastructure a network is trying to deliver.

Again, none of this is new thinking. It is the accumulated lesson of the past two decades of outer-suburban and regional infrastructure change, applied to the newest chapter of it. The networks that treat it that way will come out of this boom with both a stronger balance sheet and a stronger social licence for everything that still needs to be built after it.

The ones that do not will get the growth and lose the trust, and discover, as others have before them, that the second loss is by far the more expensive one.

Get in touch to discuss how networks can turn data centre demand into durable growth without socialising the risk, losing community trust or crowding out the infrastructure on which the energy transition depends.

Connect with Simon Gutmann on LinkedIn.