Emerging Issues Brief: Data Centre Backlash and the Risks for Police

Data centres, the physical infrastructure facilities that house networked computers, servers, and data storage systems for modern digital economy, have existed for decades. However, the Artificial Intelligence (AI) boom has created a ravenous new demand for computing power, and as a result Big Tech hyperscalers are investing historic sums in new, substantial data centre construction globally.

Australia and New Zealand are part of that build-out. Australia’s Commonwealth National AI Plan anticipates roughly six gigawatts of new capacity (worth approximately A$150 billion), and a number of projects have been announced across the country. New Zealand has seen a similar investment in data centre construction with a number of global players announcing plans to build data centre infrastructure.

As with any large-scale infrastructure development, the construction of data centres in Australia and New Zealand is not an apolitical affair. These developments have generated substantial negative sentiment, and this sentiment carries policing implications.

Why data centres draw opposition

Worries over the construction of data centre cluster on a few key issues. One of those concerns revolves around the energy demands of AI and the subsequent environmental costs.

On average, a single query run through the generative AI chatbot ChatGPT needs nearly 10 times more electricity than a standard Google search.

UN researchers say data centres consumed an estimated 448 terawatt-hours of electricity (which equates to approximately three times Australia’s annual electricity use) in 2025. If treated as a nation, that would make data centres the world's 11th largest electricity consumer. Alongside their electricity demand, data centres also need a huge amount of water for their cooling systems.

The UN found that data centres globally consumed 4.5 trillion litres of water in 2025, enough to meet the needs of more than 600 million people in Sub-Saharan Africa.

While Big Tech has invested in renewable technologies, such as hydropower, geothermal plants, and solar panels to help meet their energy needs, currently the vast amount of energy for these data centres come from non-renewable fossil fuels.

Such concerns, among others, mean that data centres are increasingly being reframed by community and environmental groups, not as objections to a particular development, but as opposition to an entire category of industrial infrastructure that competes with households for scarce energy and water while degrading local amenity. Such reframing impacts policing, because it shifts opposition from site-by-site planning disputes potentially toward sustained, networked campaigning of the kind that produces protest activity, repeated planning challenges, and direct action across multiple sites at once.

The geography of data centre development compounds this risk. Hyperscale data facilities target land near green spaces, agricultural areas, or peri-urban communities where developers can secure affordable sites with existing electricity access. In other words, rural and remote locations are particularly vulnerable. Importantly, these are precisely the areas where policing is thinnest; where the least amount of policing resources are available compared to more urban environments, where the local social fabric is most exposed to disruption from large industrial development, and where the gap between the scale of community concern and the scale of available policing response is widest. This creates a risk of public-order pressure generated by these sites falling on the parts of the policing system least equipped to absorb it. In order to help address these potential challenges, police across Australia and New Zealand may need to consider a relocation of existing police resources should data centre construction generate substantial political activism in more rural and remote areas.

Beyond data centres

At the same time, it should be noted that unease over data centre construction is one part of a much broader anti-technology/anti-AI sentiment that runs across economic, political, and personal dimensions. While data centres provide the most visible focus for community opposition, underlying grievances point well beyond any single piece of infrastructure.

The wider anti-AI sentiment is driven by several factors. The most widely held factor is a concern around anticipated economic displacement. Direct evidence of large-scale AI-driven job loss is currently limited, but the underlying technological progress is itself unprecedented. AI systems are tackling far more complex coding tasks than was predicted a year ago. The number of AI agents has expanded sharply. Enterprise spending on AI is rising dramatically. There is no evidence yet in the labour market data of AI replacing many jobs but given the pace at which the technology is improving it would be too early to dismiss its possibility. Society may be on the verge of a profound reallocation of resources and political upheaval.

As a result, Silicon Valley itself has issued warnings that state intervention, and perhaps a universal basic income, will be necessary to provide a social safety net that cushions the fall. While automation has destroyed jobs before, white-collar workers now threatened by AI have a much stronger political and social voice than in the past. Even a small number of high-visibility layoffs could provoke a backlash disproportionate to the scale of the loss.

The fear of displacement coexists with the extraordinary wealth concentration the AI boom has produced. The annualised recurring revenue of Anthropic, a leading model-maker, is projected to reach roughly A$70 billion by the end of June 2026, up from A$12 billion at the close of 2025 and A$123 million at the start of 2024. Nvidia, the dominant AI chip-maker, has become the world's most valuable company, with first-quarter 2026 revenue of approximately A$114 billion. This trend could result in targeted attacks on individuals, corporations and their leaders in response to economic division and environmental concerns.

An additional concern relates to the societal and personal fabric. AI is associated with the erosion of community, identity, trust, and potentially civilization. Stories centred around chatbot-linked psychiatric crises and suicides and the proliferation of AI-generated deepfake abuse have become all too commonplace. At the most extreme, the societal fear that AI develops consciousness, supplants human intelligence and iterates upon itself outside of human intervention, a phenomenon known as recursive self-improvement, has emerged as an acute concern. Taken together, rapid AI development has produced a cocktail of grievances that interact with a broader social environment characterised by declining trust in authority and institutions, elevated generalised anxiety, and a more confrontational information environment. The result is that disruption to the security and status of many people could translate into unrest, political activism, and at the extreme, into violence.

From grievance to action

Although Australia has not yet experienced widespread political activism and violence motivated by technological distrust and anxiety, globally we have seen a number of disruptive and violent instances motivated by this set of grievances.

In November 2025, an insurrectionary anarchist cell set fire to the construction site of the AI Campus at the Equinix data centre in Meudon, France, with the subsequent communiqué citing Equinix's alleged complicity in mass surveillance and genocide.

In April 2026, a Rome resident running a 200,000-follower anarcho-primitivist Instagram account was arrested for allegedly preparing Kaczynski-inspired attacks on major institutions. He allegedly spread extremist material promoting violence, including how-to guides to build 3D-printed weapons and explosives, while encouraging attacks on major institutions.

On 7 April, thirteen shots were fired at the front door of the Indianapolis home of city councillor Ron Gibson, with a "No Data Centers" note left on the doorstep; the attack came days after Gibson supported the rezoning approval for a data centre development in his district.

On 10 April, OpenAI CEO Sam Altman was the target of a firebombing by a suspect carrying a manifesto on AI's existential risks who had previously called for "Luigi'ing" (a reference to Luigi Mangione, the man accused of the murder of the UnitedHealthcare CEO in 2024, who has since garnered a significant following and is revered as a ‘saint’ in some extremist circles) tech executives.

What this means for policing

The proposed data centre builds across Australia present emerging implications for Australian and New Zealand policing. The most immediate is the fact that data centre construction has generated both opposition and anxieties. This opposition can manifest itself in both relatively mundane forms of political activism, such as planned protest, but as we have seen internationally, this opposition can likewise present itself in violent or extremist action.

The Australian threat picture sits between the international evidence and a domestic environment that already shares many of its preconditions. Data centre opposition in Australia is active and growing: a A$1.1 billion development in Hazelmere outside Perth was withdrawn in 2026 after 1,900 public submissions opposed it; AirTrunk's proposed A$5 billion Western Sydney facility, which would be Australia's first gigawatt-scale data centre, has drawn organised opposition from residents and schools. These are not yet sites of political violence, but they are the kind of locations and decisions around which the international cases have crystallised. Additionally, this threat needs to be understood as one that exists beyond direct threats to data centres.

As seen internationally, the threat can emerge in action pointed towards local government officials who approve data centre projects as well as perceived representatives of the tech industry. The expanded target set changes the protective security footprint from a small number of fixed sites to a dispersed set of people in their everyday lives. Importantly, while the threat to data centres themselves can be reduced through the implementation of safety features, councillors at home, planning officials at community meetings, and tech industry employees in their workplaces and residences cannot be, at least not at any reasonable resource cost.

Risk management considerations therefore will need to include both infrastructure-focused threats as well as person-specific threats when managing a public safety response. This significantly broadens the population of potential targets, while also creating a category of person-specific risk that current Australian and New Zealand policing arrangements are not configured to manage.

Three implications follow. Local councillors and planning officials in jurisdictions hosting major data centre projects need a defined policing contact for threat assessment, particularly in the period immediately following controversial planning decisions, which the international evidence identifies as the window of greatest exposure.

The public information environment surrounding planning decisions, including published home addresses, meeting attendance, and voting records, needs to be understood as a potential target identification pipeline, with corresponding advice provided to local government regarding managing exposure during periods of heightened risk.

Intelligence collection arrangements need to extend beyond the infrastructure-focused indicators to capture person-specific radicalisation directed at named officials following planning decisions.

In conclusion, the construction of data centres across Australia and New Zealand represents a site of emerging policing risk. It should, however, be seen as one element in a broader anti-technology/AI ecosystem that has the potential to generate grievances and motivate action beyond any single piece of infrastructure.

BACK TO BLOG