Why are AI data centres facing a backlash? | The Economist

Key Concepts

• Hyperscale Data Centers: Massive facilities designed for high-performance computing and AI training, requiring immense electricity and cooling infrastructure.
• Energy Intensity: The rapid increase in electricity consumption by data centers, which now account for 30% of total usage in Ireland.
• "Vibes" (Political Sentiment): The cultural and political perception of AI, heavily influenced by the alignment of tech leaders with specific political ideologies.
• Disemployment: The economic concern that AI automation will replace human labor, leading to job losses.
• Marginal Extra Use: The concept that AI data centers represent a net-new demand for electricity that does not replace existing fossil fuel consumption, unlike the transition to electric vehicles.

---

1. The Escalation of Data Center Infrastructure

Data centers have transitioned from local infrastructure projects to regional and national concerns due to their unprecedented scale.

• Scale: A proposed project in Utah covers 61 square miles (roughly the size of the country of Andorra or twice the size of Manhattan).
• Energy Demand: The Utah project aims for a 9-gigawatt capacity, which exceeds the current total electricity consumption of the entire state of Utah.
• Operational Impact: These facilities operate 24/7 and require massive cooling systems, leading to noise pollution and land-use conflicts as they expand into agricultural or green-belt areas.

2. Political Justifications and Economic Arguments

Proponents of data center expansion, particularly at the local government level, rely on several key arguments to justify these projects:

• Property Tax Revenue: These facilities provide a significant, consistent tax base that funds local services like schools.
• Economic Growth: Proponents argue that data centers stimulate broader economic development and job creation.
• National Security: In the context of the "AI race" against China, data centers are framed as essential national security infrastructure.
• Utility: The widespread adoption of tools like ChatGPT is used to justify the existence of the underlying server infrastructure, connecting the "helpful query" to the "worring warehouse."

3. The Case Against Data Centers: The "Three Pillars" of Opposition

Critics, particularly within the American left, categorize their opposition into three distinct areas:

A. Labor Impact

• The Paradox: While economists argue that AI might augment productivity, the industry’s own leaders (e.g., Sam Altman) frequently discuss a future with mass unemployment, necessitating Universal Basic Income (UBI).
• Displacement: The fear is that AI will replace human workers, and the economic benefits of this efficiency will not be distributed equitably.

B. Climate Impact

• Net-New Consumption: Unlike the transition to electric vehicles, which shifts energy sources, AI data centers represent a "marginal extra use" of electricity.
• The Efficiency Argument: Critics reject the notion that AI efficiency (e.g., using ChatGPT instead of an illustrator) justifies the massive energy expenditure, noting that it forces a trade-off between essential human energy needs and AI-generated content.

C. The "Vibes" (Political Alignment)

• Silicon Valley Politics: The perception of AI is tainted by its association with a specific "right-wing Silicon Valley tech class."
• Political Alliances: Figures like Elon Musk and Peter Thiel are seen as aligned with the extreme right, while Sam Altman’s perceived proximity to Donald Trump has alienated the American center-left, making opposition to AI infrastructure a politically strategic move for Democrats.

4. Global Context and Regulatory Trends

• Ireland: Data centers have surged from 5% of national electricity use in 2015 to 30% today, serving as a warning for other regions.
• Maine: The state legislature has moved to ban new hyperscale data centers.
• Utah: Residents are utilizing referendums to challenge large-scale projects.
• Memphis: Controversy has erupted over the use of gas turbines at XAI’s "Colossus" data center, highlighting the tension between rapid AI deployment and local environmental/regulatory standards.

---

Synthesis

The rapid expansion of hyperscale data centers has moved beyond a technical challenge to a significant political and social flashpoint. While proponents emphasize tax revenue and national security, the opposition is fueled by concerns over massive, net-new energy consumption, the threat of labor displacement, and a deep-seated distrust of the political leanings of the tech elite. As these facilities grow to the size of small countries, the conflict between the demand for AI compute and the preservation of local environments and economic stability is likely to intensify.