The Case against Capitalist Data Centres
By Gary Porter
The global data center boom—driven overwhelmingly by the artificial intelligence revolution—is being touted as progress. But beneath the glossy promises of digital innovation lies a stark reality: data centers are extractive, resource-devouring machines that impose severe costs on the communities that host them, while delivering meager local benefits. The facts demand scrutiny.
Marxists are not Luddites. We do not oppose new technology because we resist change or fear machines. Marxists advocate for massive productivity improvements that humanize and reduce labour, and increase the quality and abundance of goods and services that enrich the quality of human life for all, within the finite limits of the earth’s resources.
We know that the critical obstacle blocking us from attaining this goal is the private ownership, by a tiny parasitic class of capitalists, of the means of production. The parasites use these assets to accumulate capital and generate profit. They are the ones who stand between us and our most basic needs for decent housing, healthy food, higher incomes and more free time -- for quality lives in a healthy environment.
Among the most voracious of the capitalists, perhaps surpassed only by the fossil fuel industry, are the tech billionaires.
Based on the books and articles they write, and the speeches they give, a prominent network of tech billionaires is actively advocating for and building the infrastructure of what critics call an authoritarian future. They are often described as the “Nerd Reich”—a group that includes Peter Thiel, Elon Musk, Marc Andreessen, David Sacks, Brian Armstrong, Balaji Srinivasan, and Sam Altman.
Here is a look at their ideologies and actions:
· Peter Thiel (Palantir): The ideological architect. He bankrolls projects to bypass democratic oversight, and has stated “freedom and democracy are no longer compatible”. By “freedom” he does not mean our freedom. He means his freedom to do whatever he likes, regardless the consequences for us. His company, Palantir, now acts as the “operating system” for U.S. military and immigration enforcement.
· Elon Musk (Tesla/SpaceX/X): Champions “technocracy” over democracy. He has called elected leaders “tyrants”, and used X to support far-right figures. His companies are now critical to U.S. military infrastructure.
· Marc Andreessen (Venture Capitalist): His “Techno-Optimist Manifesto” argues an unregulated tech industry, not democracy, is key. He has been accused of making “authoritarian demands”.
· David Sacks (Venture Capitalist): Trump’s “AI and crypto czar”, who compares Trump’s return to “Caesar entering Rome”. He has tried to kill state-level AI laws, seeking to become the gatekeeper of U.S. AI policy.
· Brian Armstrong (Coinbase): Described as a “cryptocurrency titan” within the “Nerd Reich”.
· Balaji Srinivasan (Investor): Advocates for Silicon Valley’s “ultimate exit” from the U.S. and has become “stridently authoritarian,” proposing tech-governed cities run by loyalists.
· Sam Altman (OpenAI): Accused of privately working to weaken safety regulations while publicly supporting them, with a reported “pattern of lying” to consolidate power.
This shift is often described as the “Authoritarian Stack”—a system where tech oligarchs are merging with the state to provide “privatized control” over everything from AI and cloud platforms to military drones and surveillance. Their vision promotes “tech feudalism”—fiefdoms run by tech corporations—and a belief that democracy is “too slow and messy” to be trusted.
They are backing this vision with enormous financial and political capital, including $450 million to help elect Trump and secure billions in government contracts (like Palantir’s $10 billion Pentagon deal).
The Energy Crisis Hidden in the Cloud
Data centers are among the most electricity-intensive facilities on Earth. In 2025, global data center electricity consumption stood at 447 terawatt-hours (TWh). By the end of 2026, that figure is projected to reach 565 TWh—a 26 per cent increase in a single year. By 2030, consumption is expected to exceed 1,200 TWh, roughly equivalent to the annual power demand of Japan.
This surge is almost entirely attributable to AI. While conventional server electricity consumption grew by less than 1 per cent in 2025, AI-optimized servers saw consumption climb by more than 83 per cent—and are forecast to grow by a further 84 per cent in 2026. By 2027, AI servers will consume more electricity than all conventional servers combined.
These numbers are not abstract. Gartner, the research firm, states plainly that grid supply will be insufficient to meet the demands of future data center construction. Data center power demand is expected to reach 132 gigawatts (GW) in 2026 and 290 GW by 2030. This is not sustainable growth—it is on a collision course with physical reality.
Water: The Invisible Drain
Less discussed, but even more alarming, is water consumption. Data centers require vast quantities of water for cooling. A single 1 MW data center can use up to 25.5 million litres of water annually—equivalent to the daily water consumption of approximately 300,000 people.
Amazon alone reports that its global data center operations will use approximately 9.5 billion litres of water in 2025. While Amazon claims efficiency improvements, its figures exclude partner locations and construction-related consumption. Google’s data centers, by contrast, consumed an estimated 1.15 litres per kilowatt-hour in 2024—nearly ten times Amazon’s reported rate. Microsoft reported 0.27 L/kWh in 2025.
These are not victimless statistics. Over 45 per cent of data centers are located in water-stressed regions. Communities are already facing drought watch as their potable water supplies are diverted to cool server farms. In Chile, activists have successfully suspended data center permits over water concerns; in Mexico, a proposed $4.8 billion campus faces fierce opposition in a drought-stricken region.
Carbon Emissions: A Growing Liability
In 2025, approximately 450 TWh of data centre power demand generated 0.2 billion tonnes of CO₂ emissions—over 0.5 per cent of the world’s total. While this may seem modest relative to heavy industry, the growth trajectory is what matters. US data centers already operate at 548 kg CO₂ per MWh—48 per cent above the national grid average.
The proposed solution is often natural gas with carbon capture, but this raises costs by $15–$45 per MWh, bringing total power costs to approximately $115/MWh. These costs, of course, are passed on—either to ratepayers or through higher prices for digital services. Moreover, carbon capture is unproven at the scale required. In fact so far, carbon capture is a unicorn solution, a lie. In Alberta, a proposed $70-billion data center designed to run on 7.5 GW of gas-fired power could raise provincial greenhouse gas emissions to levels not seen since the coal era.
The Economic Shell Game
Perhaps the most deceptive claim made by data center proponents is that of job creation. The real numbers tell a different story.
A Business Insider analysis found that even the largest data centers generally employ fewer than 150 permanent workers, and some have as few as 25. Construction jobs are temporary—sometimes lasting less than a year—while tax breaks often last a decade or longer. The analysis found that tax breaks can amount to over $2 million for every permanent, full-time job created—roughly eight times the average per-job subsidy in other economic development deals.
According to fDi Markets, for every million dollars invested in US and Canadian data centers, a mere 0.15 jobs were created. This is a staggeringly inefficient use of public incentives.
Meanwhile, the costs to the public continue to mount. Data centers drive up electricity costs as utility operators invest billions in new grid infrastructure. In Virginia, households could see electricity bills rise by $37 per month by 2040 to pay for grid upgrades driven by data centers. Seventy per cent of electricity price increases across the US have occurred in locations near significant data center activity. Walmart has publicly stated that surging electricity bills are imperiling its expansion in states such as Ohio and Virginia.
The Human Toll
Beyond economics and the environment lies a deeper concern: quality of life and public health.
Communities near data centers report constant low-frequency vibration that residents say is ruining their health and homes. Lawsuits have been filed specifically about data center noise. Residents also raise concerns about heat generation, air quality, and impacts from backup generators.
The health costs are measurable. Research from Harvard’s T.H. Chan School of Public Health estimated between $53 million and $99 million in annual health damages from air pollution associated with a single large data center facility. A recent model indicates that US data centers in 2030 could contribute to nearly 1,300 deaths annually, resulting in a public health burden of more than $20 billion. In the most affected counties, the estimated per-household health burden can reach seven times the national average.
A Movement in Motion
Communities are fighting back—and winning. Anti-data center groups more than doubled from 396 at the end of 2025 to 833 across 49 US states by March 2026. In the first quarter of 2026 alone, grassroots opposition blocked or delayed 75 projects worth a combined $130 billion—matching the total for all of 2025.
In Canada, large mobilizations are pressuring Hamilton city council to enact a moratorium on data center development. In Spokane, Washington, a year-long moratorium was approved after a potential 500-megawatt data center—equal to half the combined power used by all residential and business customers in Spokane County—came to light.
Seventy per cent of Americans now oppose a data center being built near their home. The issue has shifted from local planning disputes to a broader political reckoning.
Socialist demands
Data centers are not benign infrastructure. They are voracious consumers of energy and water, significant emitters of carbon, poor creators of local employment, and genuine threats to community health and quality of life. The benefits—primarily to a handful of trillion-dollar technology corporations—are concentrated. The costs—higher electricity bills, depleted water supplies, degraded air quality, and diminished neighbourhoods—are borne by the communities that host them.
This is not a Luddite argument against technology. It is a demand for accountability, transparency, and fairness. If the AI revolution is to proceed, it must do so on terms that do not sacrifice the well-being of the communities upon which it depends. The facts are clear: the current model is unsustainable, inequitable, and deserves the fiercest opposition.
The management of power generation, water use, carbon pollution, and AI, is logically public. These resources must be socialized and managed by the workers and communities they purport to serve. Private ownership of these very social resources and problems makes no sense at all. In fact, the insanity of the existence of billionaires while the vast majority struggle to make ends meet demands a clear resolution.
AI data processors must become much more efficient in their power usage, maximizing the percentage of the electricity that goes directly to computing rather than cooling and other overheads. China is ahead of the west on power usage efficiency under strict state regulation. Vastly more efficient chips must be developed. Data facilities must be banned from using fresh water to cool inefficient processors. Only sea water should be used until operating efficiency no longer produces heat waste, and even sea water must be recirculated.
Avi Lewis has demanded a moratorium on the development of AI data centres until the vast global theft of human creators product is fully redressed and new additions to knowledge bases are performed only after agreement by the creator and reasonable compensation paid.
Workers and their unions must be able to veto new implementation of AI that costs jobs unless adequate provision is made to cut hours with no reduction in total pay or benefits (such as 28 hours work for 35 hours day). As Karl Marx once said, the beginning of freedom is the reduction of labour time.
Why not tax billionaires until their shoes fall off. Expropriate physical and financial assets, luxury yachts, homes and the private aircraft of the billionaires. They are arrogant, autocratic and completely useless. AI and Data centers are a symptom of authoritarian class rule by a small class of capitalist bloodsuckers. Join the rising movement against them.