Paris Marx’s new book, Hyperscale: The Ambition and Excess of Big Tech’s Data Empires, arrives at a pivotal moment, coinciding with a surge in public and political scrutiny of the massive data centers that underpin modern digital life. As a longtime critic of Big Tech and host of the Tech Won’t Save Us podcast, Marx has spent years dissecting the burgeoning influence and environmental impact of these colossal infrastructure projects. His work comes to light as opposition to data center expansion intensifies globally, highlighting concerns over energy consumption, resource strain, and the concentration of power.
“I certainly didn’t predict that it was all going to kind of come to a head as my book was coming out,” Marx says with a chuckle, acknowledging the serendipitous timing. The book offers a comprehensive historical and contemporary look at the data center industry, tracing its evolution from the early days of computing to the current era of generative AI. Marx details how these facilities, often hidden from public view, are increasingly becoming central to political debates, influencing everything from election outcomes to national energy policies.
The book’s release precedes critical midterm elections in the United States, where the proliferation of data centers has emerged as a significant issue in several key races. However, Marx emphasizes that the struggles surrounding data centers are not new. Hyperscale takes readers on a journey through decades of technological ambition and the growing resistance it has sparked worldwide. The narrative spans from anti-war protests targeting American computer facilities during the Vietnam War to more recent grassroots movements in Ireland and Chile, and even explores alternative, community-driven models for data infrastructure in places like New Zealand.
In an interview, Marx elaborated on the genesis of his book and the escalating concerns surrounding the data center industry.
The Genesis of a Critical Examination
The decision to focus a book on data centers solidified for Marx around 2019, when countries like the Netherlands and Singapore began implementing moratoriums on new construction due to environmental and infrastructure concerns. The issue gained significant traction in 2022, particularly in Ireland, where the sheer scale of energy consumption by data centers placed an immense strain on the national grid.
“In 2022, there was a growing recognition of the demands of data centers in Ireland, on the energy grid in particular,” Marx explained. “There were huge quantities of electricity that data centers were starting to consume over there.” This observation was compounded by a separate, yet related, incident involving Google in The Dalles, Oregon. Google had actively sought to conceal its water usage figures, even lobbying the city council to maintain secrecy. For nearly a year, the community was kept in the dark until the actual numbers were revealed, showing that Google’s operations were consuming a staggering 29% of the community’s water supply.
The confluence of these events, coupled with the late 2022 release of ChatGPT and the subsequent explosion of interest in generative AI, made the need for more data centers apparent. Marx recognized that the challenges faced by communities in Ireland and Oregon were likely to become widespread. “It became apparent that there’s going to need to be a lot more data centers for generative AI,” he noted. “These places were already feeling the effects of significant data center construction, and it felt like that was going to start coming to a lot more communities very soon.”
Early Warnings Ignored: Global Precedents
The perception among many Americans, Marx observes, is that the data center boom is a recent phenomenon, largely driven by the current AI surge. However, his book highlights how other countries and communities served as crucial early warning signs. Ireland, in particular, offers a compelling case study of the unchecked expansion of data centers and its consequences.
Dublin enacted a moratorium on new data center construction in 2021, not due to environmental concerns initially, but because the national grid operator deemed it impossible to provide sufficient electricity to new facilities. “It was actually the grid operator that said, ‘Hey, there are too many data centers. We do not have the grid capacity to connect any more of them,’” Marx recounted.
The industry’s response was to circumvent these limitations. Unable to secure grid connections in Dublin, data center companies began looking for sites in other parts of Ireland. Furthermore, they explored alternative energy sources, including tapping into the gas network for generators and, eventually, constructing dedicated gas-fired power plants to exclusively power their facilities.
This shift has significant implications for public infrastructure. The Irish energy grid is a public utility, meaning only the public entity is authorized to build electricity lines. However, data center companies successfully lobbied for legislation that permits private entities to construct their own electricity cables. This development raises serious questions about equity and the potential for a bifurcated energy system, where critical infrastructure is increasingly privatized and tailored to the needs of a few large corporations.
The Gas Plant Gambit and Environmental Accountability
The trend of building private gas power plants to fuel data centers, initially observed in Ireland, is now becoming a prominent feature of the data center landscape in the United States. Marx sees a direct correlation between the events in Ireland and the current situation in the U.S.
“The gas stuff looks so prescient with what’s happening in the US now, where companies are racing to build these massive private gas power plants for data centers,” he stated. In Ireland, these power plants were often granted emissions licenses by the Irish Environmental Protection Agency, despite warnings that their operation would make achieving climate targets significantly more difficult. “The agency was warning that climate targets are going to be a lot harder to achieve now.”
Marx expressed a lack of surprise at the escalating situation in the U.S., having witnessed the pattern unfold previously. However, he acknowledged the sheer scale and the rapid pace at which these developments have escalated. The ambition to rapidly expand data center capacity, particularly for AI, is driving a significant increase in fossil fuel reliance, creating a direct conflict with global climate objectives.
The Specter of Industry Collapse: Infrastructure After the Bubble
The rapid build-out of data centers and associated power infrastructure, especially in anticipation of sustained AI demand, raises questions about the long-term viability of these investments. Marx posits that the current fervor surrounding AI could represent an “AI bubble,” and its potential collapse carries significant implications for the physical infrastructure being erected.
“I think we’re in an AI bubble. But what does the collapse of that bubble look like?” Marx pondered. The plans for numerous data centers and dedicated gas power plants are predicated on an ever-increasing demand for AI computation. However, supply chain issues, including difficulties in acquiring specialized turbines, chips, and other components, could delay or even prevent the completion of these projects before a potential reassessment of AI demand.
Marx draws a parallel to the dot-com boom and bust of the late 1990s and early 2000s. While the excess build-out of fiber optic capacity is a well-documented consequence, there was also a significant, albeit earlier-stage, build-out of data centers. Many data center companies failed during that period, but survivors like Equinix were able to acquire distressed assets at low cost, positioning themselves for future growth.
The current situation, however, presents a unique challenge. If data centers are being constructed with the assumption of a short lifespan for AI-specific hardware – with chips needing to be cycled out within a few years – the value of the physical facility might diminish significantly by the time demand rebounds. Even if the data center structure remains, its core components, crucial for its immediate utility, might become obsolete.
“Inevitably, even if the data centers don’t come, then those gas plants will exist,” Marx warned. These facilities, built to serve a specific, potentially ephemeral, demand, will likely seek new customers to justify their continued operation. The intention, he believes, is not to build them only to shut them down, but their long-term purpose becomes uncertain in a scenario where AI demand falters.
An Alternative Vision: Community-Driven Data Futures
In contrast to the extractive and environmentally taxing model of hyperscale data centers, Marx’s book concludes with an inspiring example from New Zealand. There, a community-led initiative focused on developing a more accurate speech-recognition model for the Māori language, utilizing local cloud providers to host the data. This approach represents a vision of a more equitable and sustainable way to harness the power of AI.
“I would like to see governments go really hard on actually putting firmer regulations on the development of data centers,” Marx stated, advocating for stronger governmental oversight. However, he stressed that these regulations should extend beyond infrastructure to encompass the AI products themselves.
Marx points to the unfulfilled promises of generative AI and the unforeseen harms that have emerged. He observes a disconnect between the narrative promoted by tech leaders and the reality of these technologies’ impact. “Our governments do not seem to be so willing to actually take that on. They’re buying the narrative that generative AI is still what Sam Altman sold it to be in early 2023 rather than what it has become today.”
He advocates for a fundamental re-evaluation of our technological trajectory, especially in light of the immense resources being poured into enabling generative AI. “We should be asking, how much computation do we really need? How many resources are actually justifiable to put into this project? What are we actually getting out of it, right?” Marx questioned. He believes that a rigorous assessment would reveal that the current scale of investment is not justifiable, either financially, socially, or environmentally.
The Global Confrontation: Data Centers and Climate Goals
The tension between the burgeoning demand for data center capacity and the urgent need to meet climate targets is a global phenomenon. In Europe, this conflict is becoming increasingly acute. My colleague Joel Khalil reports on how the European Union’s insatiable appetite for AI is colliding with the climate goals of its member nations. Khalil predicts that data centers could become a significant political flashpoint in Europe, mirroring the situation in the U.S., as countries approach elections next year.
“The fascinating thing about the European data center build-out is the collision of competing interests,” Khalil observes. “National-level policy butts up against local priorities, ambitions for economic growth stand in conflict with climate objectives, and planning bureaucracy foils attempts by AI companies to rapidly expand their fleets.” He concludes that “It’s hard to see how Europe could possibly have it all.”
The implications of this clash are far-reaching. The construction of massive data centers, often powered by fossil fuels, directly undermines efforts to transition to renewable energy sources. The immense water requirements of cooling these facilities also strain local resources, particularly in water-scarce regions. Furthermore, the concentration of data processing power in a few large corporations raises concerns about digital sovereignty, data privacy, and the potential for monopolistic control over information.
As the digital world expands at an unprecedented pace, the physical infrastructure underpinning it – the data centers – is coming under intense scrutiny. Paris Marx’s Hyperscale provides a critical lens through which to understand this complex and rapidly evolving landscape, urging a more deliberate and responsible approach to technological development and its societal and environmental consequences. The choices made today regarding data center expansion will undoubtedly shape our future, impacting everything from our energy grids to our climate.
