Why Data Centers Have Become the Center of Big Tech's Backlash

Data centers are the physical key to understanding tech giants' power and backlash.
In Hyperscale, Paris Marx argues that understanding Big Tech requires focusing on data centers, not just algorithms. Hyperscale is both a technical term and a model of resource concentration and power expansion: data centers consume vast electricity and water, enjoy tax breaks, with costs borne by local communities while benefits flow to a few giants. The AI boom accelerates expansion and intensifies tensions, turning data centers into political issues. Tech companies ignoring these tensions face greater backlash and stricter regulation.
Editor's note: This article is based on tech journalist Paris Marx's new book Hyperscale, exploring the central role data centers play in the public and policy backlash against Big Tech.

The overlooked "hyperscale" infrastructure
When we talk about the power of Big Tech, attention tends to focus on algorithms, app interfaces, or privacy policies. But in his new book Hyperscale, journalist Paris Marx offers a more fundamental perspective: data centers are the key to understanding the power structure of tech giants. These sprawling, energy-hungry facilities are not only the physical substrate of cloud computing and AI, but also the intersection where tech companies collide with society, the environment, and policy.
In the interview, Marx unpacks the reality behind the concept of "hyperscale"—it is far more than a technical term, but a model of resource concentration and power expansion. The siting, construction, and operation of data centers involve land, electricity, water, and tax incentives, and these costs are often borne by local communities while the benefits flow mainly to a handful of tech giants.
Hyperscale in industry parlance typically refers to facilities with hundreds of thousands of servers per site, campus-scale footprints, and hundreds of megawatts of power capacity. There are currently only about a thousand such facilities worldwide, yet they carry a substantial share of cloud and AI workloads. It is also an operating model: companies drive unit compute costs to a minimum through self-built campuses, custom chips, and highly automated operations, creating a scale threshold that latecomers struggle to match. For this reason, "hyperscale" is both a technical description and an implicit statement about market concentration—only a handful of players can afford this level of capital expenditure, which provides a physical footnote for understanding the structural power of tech giants.
The root of the backlash: a mismatch between costs and benefits
The backlash against Big Tech is no accident. Data centers require enormous amounts of electricity and water, and their construction is often accompanied by local tax breaks and infrastructure investment. Marx points out that this exchange was initially seen as a success story of "attracting investment," but as facilities grew in scale, communities began to realize that the environmental and fiscal costs they bore were far higher than expected.
On the other side, the AI boom has further driven up demand for compute, accelerating data center expansion and intensifying the tensions. Residents' concerns about noise, land use, and energy prices stand in stark contrast to the "innovation dividend" proclaimed by tech companies. This mismatch between costs and benefits is the core source of backlash sentiment.
One frequently overlooked metric in data center operations is power capacity and grid interconnection queues. In places like the United States, connecting a large data center to the grid often requires years of waiting, forcing companies to either build their own generation facilities or secure special electricity rates and priority interconnection rights—directly driving up electricity bills for nearby users and crowding out resources for grid upgrades. Alongside this is water consumption: many facilities use evaporative cooling, sparking competition with agriculture and residential water use in water-scarce regions. In addition, local governments often grant property tax abatements lasting more than a decade to attract investment, shrinking the tax base for schools and public services. Together, these mechanisms explain why the benefits of "attracting investment" are often reassessed only years later.
From physical facility to political issue
Marx's analysis reveals an important shift: data centers have evolved from purely technical infrastructure into an object of public policy and political debate. Local hearings, protests by environmental organizations, and disputes over power allocation all show that tech companies can no longer treat data centers as "neutral" business decisions.
This shift also challenges the tech industry's long-standing narrative—that technological progress inevitably brings broad social benefits. As the physical footprint of data centers becomes impossible to ignore, the public is beginning to ask: who is paying for these facilities? Who benefits from them? The answers to these questions are reshaping the relationship between tech companies, regulators, and communities.
NIMBY (Not In My Backyard) is a key framework for understanding this shift. The noise from data centers (cooling equipment and backup generators), nighttime lighting, transmission lines, and land occupation are all typical NIMBY facility characteristics. But unlike traditional NIMBY projects, the benefits flow mainly to shareholders rather than local employment—automated operations mean a single campus often creates only a few dozen long-term jobs. This combination of "high resource use, low local employment" weakens community acceptance and has turned local hearings from routine procedures into genuine political battlegrounds, with lawsuits and referendums over zoning permits, water rights, and electricity cost allocation becoming commonplace.
Implications for the industry
The perspective offered by Hyperscale reminds us that understanding Big Tech's power cannot be limited to the software layer. As physical infrastructure, data centers embody multiple tensions around resource allocation, environmental costs, and policy battles. For the tech industry, ignoring these tensions could mean even greater backlash and stricter regulation.
Marx's discussion also sounds a warning for the compute race in the AI era: as companies rush to expand data center capacity, social acceptance is not a given. How to balance compute expansion with community interests will become an unavoidable challenge for tech companies.
Source: Interviews related to tech journalist Paris Marx's new book Hyperscale.
One-line Summary
Data centers are the physical key to understanding tech giants' power and the backlash against them.
Paragraph Summary
In Hyperscale, Paris Marx argues that understanding Big Tech's power requires looking beyond algorithms and apps to data centers as physical infrastructure. "Hyperscale" is not just a technical term but a model of resource concentration and power expansion: data centers consume vast amounts of electricity and water, enjoy tax breaks and infrastructure investment, with costs borne by local communities while benefits flow to a few giants. The AI boom drives up compute demand, accelerating expansion and intensifying tensions, turning data centers from business decisions into political issues debated in hearings, protests, and power allocation disputes. The article argues that if tech companies ignore these tensions, they will face greater backlash and stricter regulation, making the balance between compute expansion and social acceptance an unavoidable challenge.
Related articles

Win a Car for Getting Close to 50 Pounds? The Whatnot Marketing Behind a YouTube Challenge
An analysis of the YouTube challenge "Get Close to 50 Pounds and Win a Car," exploring how it drives traffic to livestream platform Whatnot and why weight-guessing content attracts viewers and converts commercially.

The Binding of Isaac Developer Removes AI Music, Steam Review Bombing Subsides
The Binding of Isaac: Repentance+ faced Steam review bombing over AI music, and the developer has removed the tracks. This article reviews the incident, analyzes player anxiety about AI replacing human creation, and explores transparency and trust costs around AI content in the game industry.

Astra 6.2 and Sol 6.2 Unexpectedly Flop on Simple Bench
Astra 6.2 and Sol 6.2 underperformed on the Simple Bench benchmark, sparking heated discussion on Hacker News. This article analyzes the background, Simple Bench's testing logic, and implications for evaluating model capabilities.