The rapid proliferation of data centers, once the invisible backbone of the internet, has moved to the forefront of global political and environmental discourse as the demand for generative artificial intelligence reaches an all-time high. Paris Marx, a prominent technology critic and host of the Tech Won’t Save Us podcast, has released a comprehensive examination of this phenomenon in his new book, Hyperscale: The Ambition and Excess of Big Tech’s Data Empires. The publication arrives at a critical juncture, as local communities, national governments, and environmental advocates increasingly challenge the physical footprint of the digital world.
The Physical Reality of the Virtual Cloud
For decades, the concept of "the cloud" was marketed as an ephemeral, weightless space for data storage. However, as Marx details, the infrastructure required to maintain this digital ecosystem is composed of massive, resource-intensive industrial complexes. These data centers require vast amounts of land, reliable high-voltage electricity, and millions of gallons of water for cooling.
The tension between tech expansion and local resources has shifted from a niche concern to a primary issue in regional governance. In the United States, data centers have become a pivotal topic in midterm elections, particularly in states like Virginia, where "Data Center Alley" in Loudoun County houses the world’s largest concentration of these facilities. Voters are increasingly concerned about the impact on local utility rates, noise pollution, and the strain on the aging electrical grid.
A Chronology of Global Resistance
The current wave of opposition is not a sudden development but the culmination of a decade of friction between Silicon Valley and the physical world. Marx’s research identifies several key milestones that signaled the approaching conflict:
- 2019: The First Moratoriums. Singapore and the Netherlands were among the first to implement temporary bans on new data center construction. Singapore, a land-constrained island nation, recognized that data centers were consuming an unsustainable share of its total electricity supply.
- 2021: The Dublin Grid Crisis. EirGrid, Ireland’s state-owned electric utility, issued a warning that it could no longer guarantee connections for new data centers in the Dublin region. This led to a de facto moratorium that forced tech giants to look elsewhere.
- 2022: The Transparency Battle in Oregon. In The Dalles, Oregon, a legal battle erupted over Google’s water consumption. The city initially sided with the tech giant to keep usage figures secret, but public pressure eventually revealed that Google was consuming nearly 30% of the community’s total water supply during a period of prolonged drought.
- 2023–Present: The Generative AI Explosion. The launch of ChatGPT and subsequent AI models dramatically increased the demand for high-performance computing. Unlike traditional servers, AI-specialized chips generate significantly more heat and require even more intensive power and cooling infrastructure.
Environmental and Infrastructure Data: The Hidden Costs
The scale of resource consumption by hyperscale data centers—facilities owned by giants like Amazon, Google, Microsoft, and Meta—is staggering. According to data from the International Energy Agency (IEA), data centers currently account for approximately 1% to 1.5% of global electricity use. However, in concentrated hubs, this figure is much higher.
In Ireland, figures from the Central Statistics Office (CSO) show that data centers consumed 21% of the nation’s total metered electricity in 2023, surpassing the total consumption of all urban households combined. This has placed Ireland’s climate goals in jeopardy, as the government struggles to balance its commitment to reducing carbon emissions with its economic reliance on the technology sector.
Water usage is equally contentious. A standard large-scale data center can consume upwards of 500,000 gallons of water per day for evaporative cooling. In water-stressed regions of the Western United States and Chile, this has led to direct conflict with agricultural interests and local residents. In 2024, a Chilean court partially blocked a Google data center project over concerns regarding the impact on a sensitive aquifer, forcing the company to redesign its cooling system to use air instead of water.
The Shift Toward Private Energy Generation
As public grids reach their limits, the tech industry has begun a controversial pivot: building its own power plants. In Ireland and parts of the United States, data center operators are increasingly seeking permits to build private, gas-fired power plants on-site. This allows them to bypass grid constraints but directly undermines national and international efforts to transition away from fossil fuels.
Critics argue that this creates a two-tiered energy system. While the public is encouraged to conserve energy and transition to renewables, private corporations are locking in decades of fossil fuel reliance to power AI algorithms. Marx notes that the Irish Environmental Protection Agency has warned that these private licenses will make national climate targets significantly harder to achieve.
Analysis of the AI Investment Bubble
A central theme of Marx’s critique is the possibility that the current massive build-out of infrastructure is predicated on an "AI bubble." This echoes the fiber-optic boom of the late 1990s, where thousands of miles of "dark fiber" were laid before there was sufficient demand to use them.
However, the AI bubble presents unique risks. The specialized chips (GPUs) used for AI training have a relatively short shelf life, often becoming obsolete within three to five years. If the commercial demand for generative AI fails to meet the current hype, the industry may be left with massive, specialized warehouses and private gas plants that serve no broader social purpose.
"Inevitably, even if the data centers don’t come, then those gas plants will exist," Marx argues. "They will be looking for new customers to justify their existence. I’m sure that people are not building them with the intention to just shut them down." This raises the prospect of "stranded assets"—infrastructure that remains environmentally damaging even if its original economic justification has evaporated.
Seeking Alternative Visions: The Māori Example
Despite the critique of the "extractive" nature of Big Tech, Marx highlights alternative models that prioritize community needs over corporate expansion. One such example is found in New Zealand, where the Māori people have engaged in a "sovereign AI" project.
Rather than handing over their linguistic and cultural data to global corporations to be processed in offshore data centers, Māori groups have worked with local cloud providers to host their own speech-recognition models. This approach ensures that the data remains under the control of the community and that the infrastructure used to process it is scaled to actual need rather than speculative growth. This model offers a template for how technology can be deployed in a way that respects local resources and cultural autonomy.
Regulatory Outlook and Future Implications
The collision between technology expansion and physical limits is forcing a regulatory reckoning. In the European Union, new directives are being considered that would require data center operators to publish detailed reports on their environmental performance, including water usage and waste heat recovery.
In the United States, the political fallout is becoming more localized. Planning boards and city councils that once welcomed tech investment with open arms are now implementing stricter zoning laws and requiring companies to contribute to the upgrading of local electrical infrastructure.
As countries go to the polls in the coming years, the "politics of the data center" are expected to intensify. The fundamental question remains: how much computation does a society actually need, and at what cost to the environment and local equity?
"We should be asking, how much computation do we really need?" Marx suggests. "Does what we’re getting from what Silicon Valley is creating really justify the resources that we are collectively sacrificing in order to realize it? If we actually did the assessment, I think that we would very clearly find that there’s no way to justify what we are doing right now—from a financial perspective, but also from a social and environmental one."
The outcome of this global struggle will determine not only the future of the technology industry but also the viability of national climate strategies and the autonomy of communities living in the shadow of the hyperscale empire.
About the Author
