“Railroading Community Opposition”: Greens MLC Abigail Boyd on the Emerging Datacentre Crisis

TEXT START: There is a datacentre crisis emerging in this state, as overseas big tech interests are racing down under to build facilities to store foreign owned data.

The Dissection

This text turns invisible AI infrastructure into a local political target. It assembles residents, environmentalists, labor concerns, sovereignty arguments and anti-war sentiment into one coalition against datacentre expansion.

Its strongest move is identifying datacentres as the physical substrate of AI: they consume land, water and energy while creating limited permanent employment. It also exposes the construction-GDP illusion—temporary spending is counted as prosperity while ownership, profits and strategic control remain external.

But the article ultimately treats the datacentre as the disease rather than the infrastructure of a wider capital transition. It wants scrutiny, restrictions and a larger public return without confronting who owns and controls the AI systems built on that infrastructure.

The Core Fallacy

The central error is assuming that local opposition can reverse a global competitive process. Under the Discontinuity Thesis, datacentres are not an optional civic nuisance. They are the capital base required for cognitive automation dominance. If NSW delays or rejects them, investment can relocate, while the pressure to automate continues.

The text correctly sees that more datacentres will not preserve mass employment. It does not follow that stopping them will preserve the wage-consumption circuit. If AI achieves durable superiority across cognitive work, P1, P2 and P3 still operate elsewhere.

The distinction between socially useful datacentres and AI slop is also unstable. AI will be embedded in health, finance, logistics, defense and administration. Calling some workloads essential and others frivolous does not defeat the competitive mechanism. Capital does not require moral legitimacy to replace labor.

Hidden Assumptions

• State governments can impose enough conditions on mobile foreign capital to capture a durable public windfall.
• Public consultation and planning thresholds can meaningfully control a transnational infrastructure race.
• AI can be cleanly separated into acceptable critical uses and disposable applications.
• Taxing or bargaining with AI firms can compensate for the permanent loss of productive participation.
• Australian sovereignty can be restored merely by controlling where servers are built, despite foreign ownership of the models, capital and intellectual property.
• The supplied statistics—such as the 84–90 percent AI-workload estimate and the claimed water share—are treated as established evidence, although the excerpt provides no supporting methodology.

The article also frames future transfers as billionaire generosity. That is a political choice, not a technical necessity. Transfers could preserve consumption, but they would confirm the thesis: consumption survives while productive participation dies.

Social Function

This is a partial truth wrapped in transition-management politics and ideological anesthetic.

It accurately identifies resource externalities, regulatory capture, offshore ownership, thin long-term job creation and the threat of AI-driven displacement. It is not empty copium.

Its anesthetic function is to personalize the process as foreign tech interests and tech-bro fascists, then imply that blocking or taxing datacentres is a sufficient remedy. That gives the audience a visible enemy and a local battlefield while leaving the deeper ownership structure intact. The moral language about slop, poisoned minds and war adds political force, but it does not establish the mechanism of obsolescence.

The Verdict

The article is a sharp alarm attached to an incomplete diagnosis. Datacentres are indeed the physical manifestation of a transfer of power and resources toward AI capital. They are also only the visible shell of the system doing the killing.

Planning delays, environmental constraints and public bargaining are lag defenses. They may extract concessions, slow deployment or redistribute some rents. They cannot preserve mass employment once cognitive automation becomes cheaper and better, and they cannot resurrect the postwar wage-consumption circuit.

The real question is not whether NSW gets a few benefits from hosting the machinery. It is who controls the machinery, the energy, the logistics, the maintenance layer and the resulting claims on output. Without that shift, community opposition is a warning flare over a machine that keeps advancing.