Paris Marx’s new book, Hyperscale: The Ambition and Excess of Big Tech’s Data Empires, arrives as opposition to massive server farms reaches a fever pitch in the United States. The longtime critic of Big Tech and host of the Tech Won’t Save Us podcast notes that the timing was unexpected. He did not predict that the issue would come to a head precisely as his book was being published.
While data centers are currently influencing US elections, with the issue potentially deciding key midterm races, the struggle is not new. Marx’s work traces a global history of resistance, starting with antiwar protests targeting American computers during the Vietnam War. It moves through recent conflicts in Ireland and Chile, and concludes with an alternative vision emerging in New Zealand.
Energy Grid Strain in Ireland
Ireland serves as a primary case study for the infrastructure challenges facing the industry. In 2022, local authorities recognized the immense demands data centers placed on the national energy grid. Dublin passed a moratorium in 2021 after the grid operator stated there was no capacity to connect additional facilities.
Industry response involved seeking alternative power sources. Unable to secure grid connections in Dublin, companies looked toward other parts of Ireland and began tapping into the gas network. This led to the construction of dedicated gas power plants to fuel the servers, bypassing the public utility system. The Irish energy grid is publicly owned, but data center operators lobbied for legislation allowing private companies to build their own electricity cables, raising concerns about equity.
Operators obtained emission licenses from the Irish Environmental Protection Agency to burn fossil fuels. The agency warned that these activities would make achieving national climate targets significantly harder. This pattern mirrors current developments in the United States, where companies are racing to build private gas power plants for their facilities.
Water Use and Community Impact
Marx highlights the case of Google in The Dalles, Oregon, as an early warning sign. The company attempted to keep its water consumption figures secret, even pressuring the city council to maintain confidentiality. After a year of resistance, the community obtained the data and discovered that Google was using 29 percent of the local water supply.
The release of ChatGPT in late 2022 accelerated these tensions. It became clear that generative AI would require a vast expansion of data center infrastructure. Communities already feeling the effects of construction realized that the scale of the problem was about to increase dramatically.
Marx argues that many Americans believe data centers appeared suddenly. However, the global history shows that local communities have long served as early warning signs for the industry’s expansion. The scale of the current build-out is staggering, but the tactics used by tech companies to secure resources are familiar.
Alternative Models and Future Risks
The book concludes with an example from New Zealand, where a community-led effort created a speech-recognition model for the Māori language. They used local cloud providers to host the data, offering a non-extractive alternative to the dominant tech model. Marx suggests that governments should implement firmer regulations on both data center development and the AI products they support.
He questions whether the current infrastructure build-out is justified. The industry is planning for an AI boom that may not materialize as expected. If the AI bubble collapses, the physical infrastructure, including gas plants and server farms, may become stranded assets. Unlike the dotcom era, where surviving companies like Equinix bought up excess capacity, today’s technology cycles so quickly that data centers planned for current AI needs may be obsolete before demand returns.
Marx urges a reassessment of how much computation is actually necessary. He argues that the financial, social, and environmental costs of the current trajectory cannot be justified. As Europe faces similar collisions between economic ambitions and climate goals, the global debate over data center expansion is likely to intensify.
Source: Wired
