In Houston, 1978: 100 percent of city-owned landfills sat in Black neighborhoods. Six out of eight incinerators. Three out of four private landfills. From the 1930s through 1978, 82 percent of the city’s garbage was dumped where Black people lived—though Black residents made up only 25 percent of the population.
Dr. Robert Bullard, researching a lawsuit his wife was filing against the state of Texas, documented what would become the founding evidence of environmental justice (Bullard, 1990, pp. 43-44).
This wasn’t random. City council members—all white—made deliberate choices about where to site facilities that would poison the air, contaminate the soil, and sicken children. They called it urban planning. Economic development. Progress.
Four years later, North Carolina chose the predominantly Black town of Warren County for a PCB-contaminated soil dump. Residents didn’t complain. They lay down in front of the trucks. Over 500 people were arrested in what became the first major act of civil disobedience against environmental racism in American history (McGurty, 2007, pp. 1-3). The Washington Post called it “the marriage of environmentalism with civil rights.” For the people breathing the poison, it was simply survival.
By 1987, the United Church of Christ had documented what Black communities already knew: race was the single most powerful predictor of where hazardous waste would end up in America. Not income. Not property values. Race (UCC Commission for Racial Justice, 1987). A 2001 Los Angeles study confirmed what officials long denied—“minorities attract TSDFs but TSDFs do not generally attract minorities” (Pastor, Sadd, & Hipp, 2001, p. 17). They weren’t putting dumps where Black people happened to live. They were targeting Black neighborhoods for the dumps.
The pattern is old. The pattern is intentional. And I’m watching it repeat right now in the language of artificial intelligence governance.
The Architecture of Containment
Environmental racism operates through a specific logic: the harm is acceptable because it can be contained to people whose lives are deemed less valuable. Professor Michael Ash at the University of Massachusetts Amherst has documented the elite mindset precisely: the assumption that pollution can be “displaced onto the Other, onto the wrong side of the environmental tracks”—so decision-makers can “put on blinders, don’t pay too much attention to the gross amount of pollution that is being produced” (Ash & Fetter, 2004).
But here’s what that logic obscures: segregated cities create more pollution for everyone. Ash’s research demonstrated that white people in highly segregated metropolitan areas face significantly worse air quality than white people in integrated areas. Rachel Morello-Frosch at UC Berkeley found the same pattern: higher concentrations of cancer-causing air pollutants in segregated areas, affecting all residents regardless of race (Morello-Frosch & Jesdale, 2006). When you tell decision-makers they can offload environmental harm onto Black communities, they stop caring about reducing pollution altogether. Everyone breathes the same sky. But the lie of containment keeps the system running.
The architecture of containment isn’t limited to geography. It operates wherever power needs to neutralize threats while maintaining the appearance of legitimacy.
In March 1971, activists broke into an FBI field office in Media, Pennsylvania, and stole documents that revealed what Black organizers had known but couldn’t prove: the government wasn’t just surveilling them—it was actively working to destroy them (Medsger, 2014). The FBI’s Counter Intelligence Program had one directive from J. Edgar Hoover: “expose, disrupt, misdirect, discredit, or otherwise neutralize” Black liberation movements (Churchill & Vander Wall, 1990, p. 37).
Assata Shakur described discovering the scope of it in her autobiography: “Nobody could possibly have known that the FBI had sent a phony letter to Eldridge Cleaver in Algiers, ‘signed’ by the Panther 21, criticizing Huey Newton’s leadership. No one could have known that the FBI had sent a letter to Huey’s brother saying the New York Panthers were plotting to kill him” (Shakur, 1987, p. 221). The FBI manufactured fake cartoons, forged letters, made anonymous phone calls, and deployed agents to pit organizations against each other. They told SNCC that the Panthers were coming to kill them. They told the Panthers that other groups were plotting against them.
On December 4, 1969, Chicago police—working with the FBI and Cook County State’s Attorney Edward Hanrahan—assassinated Fred Hampton in his bed. He was 21 years old (Haas, 2019). The “threat” they neutralized was the Panthers’ Free Breakfast for Children program, their free health clinics, and their political education classes. The danger wasn’t violence. The danger was that they were effective.
All of this was legal. Sanctioned by the state. Done in the name of national security.
Harry Haywood, a Black Communist who survived this era, wrote in his autobiography: “The full story of intrigue, murder, character assassination, splittism, and provocative activities is only now beginning to come to light. The exposure of the FBI’s notorious Counter Intelligence Program (COINTELPRO) operations was but the tip of the iceberg” (Haywood, 1978, p. 628).
In 2017, a leaked FBI intelligence assessment created a new category: “Black Identity Extremists.” The language had changed. The program had not (FBI, 2017). Operation IRON FIST and subsequent surveillance programs deployed undercover agents against Black Lives Matter activists using the same playbook from 1969—updated with social media monitoring and predictive policing algorithms.
The architecture of containment doesn’t disappear. It adapts.
State violence is expensive and visible. There’s a cleaner way to neutralize a movement: buy it.
After the urban rebellions of the 1960s, the corporate establishment needed new allies in Black communities. McGeorge Bundy—former National Security Advisor who orchestrated the Vietnam War—became Ford Foundation president in 1966 and immediately redirected foundation resources toward Black organizations willing to operate within acceptable parameters (Ferguson, 2013). Fifty corporations jointly sponsored Black Power conferences. Harvard redefined “self-determination” to mean community development corporations and tax incentives for investors in the ghetto.
Richard Nixon said it plainly in 1968: “What most of the militants are asking is not separation but to be included in—not as supplicants, but as owners, as entrepreneurs—to have a share of the wealth and a piece of the action” (Weems & Randolph, 2009, p. 154).
This is the nonprofit industrial complex in formation. Take the language of the movement. Fund the parts you can control. Require 501(c)(3) status so they can’t be “too political.” Let them provide services—food banks, after-school programs, job training—but never let them build power. Turn organizers into grant writers. Turn movements into managed programs.
The people creating the harm now fund the organizations claiming to address it. And because they control the funding, they control what’s possible to imagine.
What dies in this transaction? Community. The very fabric that makes resistance possible.
The Black Panthers didn’t serve free breakfast to 20,000 children because they thought hungry kids should learn to be entrepreneurs. They did it because they understood that liberation is collective or it’s nothing. Fred Hampton said it: “We’re going to fight racism with solidarity” (Hampton, 1969). Not with small business loans. Not with individual wealth accumulation. With solidarity.
When you replace “we” with “I,” when you turn information into a commodity to be hoarded rather than shared, when you teach people to compete for scraps instead of demanding the whole table—you’ve killed the movement without firing a shot.
The AI parallel is almost too perfect. And now they’re adding another layer: you won’t even own the tools to succeed.
You Will Own Nothing
At the 2024 New York Times DealBook Summit, Jeff Bezos compared corporate data centers to “the old factory generator”—infrastructure companies built before the electrical grid existed. “That is going to flip,” he said. “It’s going to flip to utility computing” (Bezos, 2024). The implication was clear: owning your own computing power is an anachronism. The future is AWS. The future is the cloud. The future is you, paying Amazon for computing access. Forever.
Not own. Rent. Subscribe.
And to make sure that vision becomes reality, the infrastructure is being dismantled. Memory manufacturers are openly shifting production away from consumer hardware toward AI data centers. Micron’s CEO announced the company is “reallocating capacity” to AI-related memory products, with high-bandwidth memory revenue projected to quadruple in a single year (Micron, 2024). SK Hynix is “prioritizing HBM3E production for AI chips over conventional DRAM” (SK Hynix, 2024). Dell, ASUS, HP, and Lenovo have all reported consumer PC price increases driven by component costs. Industry analysts at Gartner project that “consumer PC pricing [will] remain elevated through 2027” with “normalization unlikely before 2028” (Gartner, 2024).
But there’s a geopolitical wildcard. Chinese manufacturer ChangXin Memory Technologies (CXMT) is positioned to undercut Western prices significantly—pricing DDR4 and DDR5 modules 20-30% below Samsung and SK Hynix (TrendForce, 2024). Whether this alternative materializes depends on U.S. trade restrictions, supply chain politics, and whether the Western-dominated tech ecosystem will accept Chinese memory chips. The hardware restriction isn’t just about corporate strategy. It’s about whether alternative supply chains can break the monopoly before dependency locks in.
This isn’t a natural market shift. This is manufactured scarcity—contested, but deliberate.
They’re restricting consumer hardware to force you into cloud subscriptions. You won’t be able to afford to own a computer, so you’ll rent one from Bezos. Every month. Forever. And when you stop paying, you lose access to your work, your files, your computing power.
This is digital sharecropping.
You don’t own the land—you rent access. You don’t own the tools—you subscribe to them. The landlord controls what you can plant, when you can harvest, and takes a cut of everything you produce. You’re perpetually in debt to the system. And if you step out of line, they cut off your access.
Except now the “land” is your ability to compute. Your ability to create. Your ability to organize.
When you own your computer, you control what’s on it. You can install whatever software you want. You can encrypt your files. You can disconnect from the internet and still work. You can share tools freely with others. You can build systems that threaten power.
When you rent computing from Bezos’s grid, he sees everything you do. He decides what software you can run. He can change the terms of service tomorrow. He can raise prices whenever he wants. He can terminate your account if you violate the rules he writes. And you have no recourse because you own nothing.
This is the endpoint of Black capitalism’s logic: not just individual competition within an exploitative system, but individual dependency on the very infrastructure that exploits you. Now we’re told that AI is the great equalizer. Learn to prompt. Learn to code. Use these tools to get ahead, to optimize, to scale your personal brand, to automate your way to success.
The framing is always individual. Your AI assistant. Your competitive advantage. Your productivity hack.
What’s missing? The collective. The solidarity. The shared power.
When ChatGPT launched, the discourse was immediate and predictable: courses on “how to use AI to make money online.” Gurus selling “AI wealth secrets.” The entire conversation became about individual gain—use AI to write faster, sell more, automate your business, beat your competition.
Nobody’s talking about using AI to organize rent strikes. To collectively analyze eviction patterns. To build shared databases of labor violations. To coordinate mutual aid at scale. To pool knowledge instead of selling it.
Because that’s not what the tools are being built for.
The most powerful AI models cost money—money that flows to the same companies restricting your access to computing hardware. The education to use them well costs money. The infrastructure to run them costs money. And in each of these transactions, you’re being trained: this is a tool for getting ahead, not for pulling others up with you.
Information becomes a commodity. You don’t share your best prompts—you monetize them. You don’t teach people what you’ve learned—you sell the course. You don’t build in public—you build your moat.
And you do it all on rented infrastructure. Paying monthly. Forever.
The Panthers understood something crucial: you can’t win by learning to exploit people better than the exploiters. You can’t liberate yourself by becoming the landlord, the boss, the algorithm’s owner. And you certainly can’t liberate yourself by becoming a permanent tenant.
Liberation is collective or it’s just a different arrangement of chains.
When Frederick Douglass said literacy was the pathway from slavery to freedom, he didn’t mean “learn to read so you can get a better job within the plantation system.” He meant: learn to read so you can understand the ideology that justifies your bondage, so you can organize with others, so you can imagine a world where there are no plantations at all.
The question isn’t “how do I use AI to get ahead?”
The question is: “how do we ensure everyone has access to these tools—and owns them—not to compete with each other, but to build power together?”
That question makes the AI companies very nervous. Because the answer threatens everything they’re building. It threatens the subscription model. It threatens the cloud monopoly. It threatens the data extraction. It threatens the control.
And that’s exactly why we have to ask it.
Infrastructure Endgame
The leading AI companies are valued at staggering levels despite massive losses. OpenAI: $157 billion valuation, $5 billion in operating losses (The Information, 2024). Microsoft, their largest investor, is already testing alternatives (The Information, 2026). Apple treats AI models as interchangeable commodities—ChatGPT and Gemini compete for positioning like search engines (Bloomberg, 2026). The Bank of England warned in October 2024 that AI valuations “appear stretched” with “broader financial stability implications” (Bank of England, 2024).
This is a bubble. Even the people building it acknowledge it’s a bubble.
Here’s what makes it different from previous tech bubbles: the infrastructure is real. When the dot-com bubble burst, websites disappeared. When the AI bubble bursts, we’ll be left with physical infrastructure built at enormous environmental cost—infrastructure controlled by whoever survives the crash.
And that infrastructure is already being weaponized. While everyone’s focused on ChatGPT and image generation, the Pentagon has quietly integrated AI into the entire military apparatus.
In August 2023, Deputy Secretary of Defense Kathleen Hicks announced “Replicator”—a program to field “multiple thousands” of autonomous weapons systems “within the next 18 to 24 months” (Department of Defense, 2023). Not drones with human operators. Autonomous systems. AI-controlled weapons that select and engage targets.
The Pentagon has hundreds of active AI projects, most classified. The ones we know about include autonomous drone swarms, AI-powered targeting systems, unmanned surface vessels like the Sea Hunter (which can operate for months without crew), “collaborative combat aircraft” that fly alongside piloted fighters, and Project Maven—AI that analyzes drone footage in real-time to identify people and objects (GAO, 2021).
They’re spending billions to ensure “appropriate levels of human judgment over the use of force”—a phrase deliberately left undefined. In practice, it means a human can be “in the loop,” or “on the loop” (supervising), or increasingly, completely removed from decision-making.
The justification is always geopolitical competition. We need AI weapons to counter adversaries. We need them deployed fast. Testing? Safety? Ethical review? Those slow things down.
Here’s what nobody’s saying out loud: these systems will be trained on data from actual conflicts. Conflicts fought disproportionately in Black and brown countries. Conflicts where the “targets” are people who look like the communities already being poisoned by environmental racism, already being surveilled by COINTELPRO’s successors, already being managed by the nonprofit industrial complex.
The same pattern. Just automated.
When an autonomous weapon system makes a “mistake” and kills civilians, who’s accountable? The AI didn’t mean to. The operator was just supervising. The company that built it followed Pentagon guidelines. The Pentagon approved it under “appropriate levels of human judgment.”
This is surveillance that never sleeps feeding weapons that never hesitate. COINTELPRO’s targeting infrastructure married to the capacity for mass killing. Control that’s been algorithmically optimized, violence that happens faster than human accountability can track.
And it’s being built with the same AI infrastructure that’s supposed to “democratize” access to knowledge.
AI as Scapegoat
Now here’s where it gets insidious. They need you to believe AI is consuming all the resources. Because that hides what they’re actually doing.
You’ve seen the headlines: “AI data centers are draining rivers.” “ChatGPT uses a bottle of water per prompt.” “AI will cause water bankruptcy.”
The reality is more complicated—and more damning.
Data center water usage varies enormously by cooling technology. According to Uptime Institute’s 2024 global survey, 38% of data centers use evaporative cooling—and those facilities account for 71% of total water consumption (Lawrence Berkeley Lab, 2025). The massive hyperscale data centers run by Google, Microsoft, Amazon, and Meta predominantly use evaporative cooling, consuming 0.9-1.3 liters per kilowatt-hour (Data Center Dynamics, 2025). Google alone: 5.6 billion gallons in 2023. Microsoft: 1.7 billion gallons (Google, 2024; Microsoft, 2024).
And the trend is accelerating. Evaporative cooling adoption increased from 32% to 40% of facilities between 2020 and 2025 (Uptime Institute, 2025). AI-specific infrastructure is even more water-intensive—Nvidia H100 clusters use direct liquid cooling at 1.5-2.0 liters per kWh. This isn’t hidden or accidental. It’s a business decision: hyperscale operators prioritize energy cost savings over water conservation.
The environmental impact is real.
But here’s what the “AI is drinking your water” narrative obscures: these same regions have been privatizing water for decades. Nestlé extracts tens of millions of gallons on expired permits for pennies (Food & Water Watch, 2024). Municipal water systems are sold to private companies that jack up rates by 59% on average. Infrastructure in Black communities has been neglected for generations—like Flint, where the water crisis wasn’t about scarcity but about who controlled the pipes and who was deemed worthy of investment.
Now AI becomes the convenient scapegoat.
“Sorry, can’t afford to fix your pipes—the data center needs it.” But the infrastructure neglect preceded the data center. The deliberate decisions not to invest in certain communities’ water systems were made decades ago. The pipes that should have been replaced in the 1990s are still poisoning people today.
The data center didn’t cause that. Racism caused that. Disinvestment caused that.
But now when people complain about resource scarcity, they’re told: “Well, AI needs a lot of resources, you know. Everyone’s making sacrifices.”
This is the same move as blaming immigrants for housing shortages when the real issue is financialization of housing. Blame the visible new thing. Hide the structural extraction.
AI is being positioned as the scapegoat for capitalism’s failures. For infrastructure they refused to maintain. For resources they’ve been commodifying. For commons they’re enclosing and reselling at profit.
And it works because it sounds plausible. AI is powerful and new and demanding. Of course it uses a lot of resources. Of course there are tradeoffs. Of course someone has to sacrifice.
Just never them. Always you.
The Same Machine, Different Masks
The major AI companies have become deeply intertwined through investments and dependencies designed to be too big to fail. When the bubble bursts, the small players disappear. What’s left: a handful of corporations controlling the infrastructure—the same companies with government contracts, building autonomous weapons, forcing dependency on their cloud.
The bubble isn’t a mistake. It’s the mechanism.
Environmental racism, COINTELPRO, the nonprofit industrial complex, Black capitalism, digital sharecropping, AI governance—the same machine. The pattern is identical: create the problem, position yourself as the solution, ensure the solution never threatens your power.
This is how power works when it’s sophisticated. You don’t just suppress the opposition—you fund it. You don’t just criminalize organizing—you offer an alternative that looks like empowerment but forecloses liberation. You don’t just restrict access—you create a rental system that makes restriction profitable.
The sophistication of modern power: it doesn’t say “no.” It says “yes, but only on our terms.” Access without ownership. Opportunity that destroys collective power. It trains you to see solidarity as naive.
What’s Different This Time
Every generation faces a version of this pattern. What makes the AI moment different is speed, scale, and irreversibility.
Speed: Previous iterations took decades to lock in. AI infrastructure is being built in months. By the time most people understand what’s happening, it’s already done.
Scale: Environmental racism poisoned neighborhoods. COINTELPRO destroyed movements. AI touches everything—every communication system, financial transaction, hiring decision, policing algorithm, weapon system. The scope of control isn’t local anymore. It’s total.
Irreversibility: You could, theoretically, clean up toxic waste. You could expose COINTELPRO (and activists did). But once you’ve surrendered ownership of computing infrastructure, once hardware becomes unaffordable, once everyone’s dependent on cloud subscriptions—how do you get it back? Once autonomous weapons are deployed, making kill decisions faster than human oversight—how do you reverse it?
This iteration of the pattern has a finality the others didn’t.
They want you to believe this is all inevitable. Every piece framed as technological necessity, natural evolution, impossible to change.
But nothing about this is natural. Manufacturers stopping consumer hardware production—that’s a business decision. Computing centralized in corporate clouds—that’s infrastructure policy. Autonomous weapons with decreasing oversight—that’s military doctrine. Resources privatized while technology takes the blame—that’s theft with a cover story.
Every bit of this results from decisions made by people with power to maintain that power. Which means different decisions could produce different outcomes.
The inevitability narrative is a weapon. Its purpose is to make you stop fighting before you start.
Don’t give them that.
What’s At Stake
Here’s what we’re actually deciding right now, whether we realize it or not:
Will we own the tools of our liberation, or rent them from our oppressors?
When the Panthers fed children, they owned the kitchen. When they provided healthcare, they controlled the clinic. When they educated their community, they determined the curriculum.
They didn’t rent the space from the Ford Foundation. They didn’t need permission from IBM to feed people. They didn’t operate on subscription models that could be revoked.
Ownership meant autonomy. Autonomy meant power. Power meant threat.
That’s why they were destroyed.
Now they’re offering you the same deal in a different package: You can have AI tools. You can have cloud computing. You can have access to powerful models. Just never own any of it. Just pay monthly. Just operate within our terms of service. Just let us see everything you do.
Will we build collective power with these tools, or turn them into weapons against each other?
AI could coordinate rent strikes across a city. It could analyze patterns of wage theft across industries. It could help communities track environmental violations. It could translate organizing materials into dozens of languages instantly. It could pool knowledge instead of commodifying it.
But that’s not what it’s being built for. It’s being built for individual optimization. For automating the hustle. For competing more efficiently within an unchanged system.
When the tool is designed for individual gain, teaching it to everyone doesn’t liberate anyone. It just makes the exploitation more sophisticated.
Will we fight for infrastructure we control, or accept dependency on theirs?
The difference between a public library and a bookstore isn’t just price. It’s about who decides what’s available, who has access, what happens to the knowledge.
The difference between owning your computer and renting computing power isn’t just cost. It’s about who controls what you can do, who sees what you’re doing, what happens when you stop paying.
They’re trying to privatize computing the way they privatized water, and they’re using AI as the justification.
If we let them, we’ll be buying back—at profit margins they set—what should have been a commons. Forever.
Will we accept autonomous weapons making kill decisions, or demand human accountability?
Once you automate killing, once you remove human judgment from the loop, once “appropriate levels” means whatever the Pentagon decides this week—you’ve created a system that can’t be held accountable.
The targets will be selected by algorithms trained on biased data. The decisions will happen too fast for oversight. The responsibility will be diffused across so many actors that nobody’s guilty.
And the communities already being targeted by environmental racism, already being surveilled by COINTELPRO’s descendants, already being managed by the nonprofit industrial complex—they’ll just face it all at machine speed now.
This isn’t hypothetical. The systems are being built. The contracts are being signed. The weapons are being deployed.
What This Moment Demands
I don’t have a ten-point plan. I don’t have a manifesto. I’m not here to tell you exactly what to do.
What I have is questions. The kind that refuse both false hope and passive despair. The kind that activate rather than prescribe.
How do we ensure computing power remains accessible and owned by the people who need it?
Not rented. Owned. What would it look like to fight for the right to affordable hardware the way previous generations fought for the right to literacy, to land, to vote?
How do we build AI tools for collective power instead of individual optimization?
What if the most powerful use of AI wasn’t to write better marketing copy or automate your business, but to organize communities, track exploitation, coordinate mutual aid, analyze power structures?
What infrastructure would that require? What knowledge would need to be shared? What alliances would need to be built?
How do we refuse the privatization of resources while rejecting the scapegoating of AI?
They’re selling you water that used to be public and blaming technology for why it’s scarce. How do we name that theft while also demanding better computational systems?
How do we demand accountability for autonomous weapons before they’re deployed at scale?
The Pentagon is fielding thousands of autonomous systems. What would actually stop that? Not appeals to their conscience. What material pressure could work?
How do we support the people building alternatives without them being captured by foundation funding?
The nonprofit industrial complex is ready to fund “ethical AI” initiatives that pose no threat. How do we resource the work that actually builds power without selling out to the people we’re fighting?
How do we teach pattern recognition at scale?
Not in academic journals. Not in think tank reports. How do we make sure the grandmother in the housing project, the worker at the warehouse, the organizer in the community center can see what’s being done to them in time to resist it?
I don’t have the answers. Nobody does, alone. That’s the point.
These aren’t questions for individuals to solve through better personal choices. They’re questions that require collective power to even approach.
And building that power requires exactly what they’re trying to prevent: ownership of tools, shared knowledge, solidarity across differences, the belief that we deserve better than this.
The Pattern and the Refusal
I’m 39 today. Old enough to have seen this pattern repeat. Old enough to recognize it in new clothes.
They’re not more creative than they used to be. They’re just better funded.
The same strategies that justified redlining in the 1930s justify algorithmic lending discrimination now. The same logic that rationalized COINTELPRO powers domestic surveillance systems. The same structure that built the nonprofit industrial complex gets applied to “AI ethics” boards at tech companies. The same lie that said “separate but equal” was constitutional now claims “centralized AI is necessary for safety.”
They polish the language. They update the technology. But the pattern stays the same: create the problem, position yourself as the solution, ensure the solution never threatens your power.
Pattern recognition is a gift. Once you see it, you can’t unsee it. And that gift isn’t for you alone. It’s meant to be shared, taught, passed on to anyone who’ll listen.
Frederick Douglass didn’t learn to read so he could be literate by himself. He learned so he could teach others, build collective understanding, organize. Literacy wasn’t the pathway from slavery to freedom for him alone—it was the pathway for everyone he taught, and everyone they taught, and everyone those people taught.
The same is true now. Understanding how AI governance replicates historical patterns of control isn’t just for people who study AI. It’s for everyone living under the systems being built. The people being priced out of hardware ownership. The people being sold cloud subscriptions as “solutions.” The people watching their resources get blamed for corporate extraction. The people being targeted by autonomous weapons. The people being trained to compete with each other instead of organizing together.
The pattern applies to all of us. So the understanding has to reach all of us.
The pattern is old. But it’s not inevitable.
Every iteration of this machine has been resisted. Redlining was fought—and sometimes beaten. COINTELPRO was exposed. The nonprofit industrial complex has been analyzed and named, which means it can be refused.
People have always found ways to learn when learning was criminalized. To organize when organizing was illegal. To build power when power was supposed to be impossible. They did it with tools they made themselves, with knowledge they shared freely, with solidarity they chose over safety.
We can do it again. But only if we refuse the lies: that individual access equals liberation, that renting equals ownership, that AI is too dangerous for the masses and too necessary for war, that this is all inevitable.
None of it is true. All of it serves power.
They need us to accept the terms they’re offering. They need us to rent instead of demanding ownership. They need us to compete instead of organizing. They need us to believe the pattern can’t be broken.
Our refusal breaks it. Not alone. Not through individual choices. But together, with clear eyes about what’s being built and what’s at stake.
This essay isn’t a blueprint. It’s an invitation. To see the pattern. To name it clearly. To refuse the terms being offered. To ask better questions than the ones they want us asking. And to remember what Fred Hampton knew: we’re going to fight this with solidarity, or we’re not going to fight it at all.
The tools are being built right now. The infrastructure is being locked in right now. The decisions are being made right now.
We’re not too late. But we won’t be early for much longer.
What we do with that matters.
References
Apple Inc. (2024, June). WWDC 2024 Keynote. Apple Developer.
Ash, M., & Fetter, T.R. (2004). Who Lives on the Wrong Side of the Environmental Tracks? Evidence from the EPA’s Risk-Screening Environmental Indicators Model. Social Science Quarterly, 85(2), 441-462.
Bank of England. (2024, October). Financial Stability Report. Bank of England.
Bezos, J. (2024, December 4). Interview at The New York Times DealBook Summit [Video interview]. Andrew Ross Sorkin, interviewer. The New York Times.
Bloomberg. (2025, January 22). OpenAI’s Stargate Project Raises Questions About Funding Sources. Bloomberg.
Bloomberg. (2026, January). Apple Expands Google’s Gemini Role in Latest iOS Update. Bloomberg Technology.
Bullard, R.D. (1990). Dumping in Dixie: Race, Class, and Environmental Quality. Westview Press.
Churchill, W., & Vander Wall, J. (1990). The COINTELPRO Papers: Documents from the FBI’s Secret Wars Against Dissent in the United States. South End Press.
Data Center Dynamics Intelligence. (2025). Global Data Center Cooling Technology Report 2025. Data Center Dynamics.
Department of Defense. (2023, August 28). Deputy Secretary of Defense Kathleen Hicks Announces Replicator Initiative [Press release]. defense.gov.
Federal Bureau of Investigation. (2017, August 3). Black Identity Extremists Likely Motivated to Target Law Enforcement Officers [Intelligence Assessment]. (Leaked document).
Ferguson, K.E. (2013). Top Down: The Ford Foundation, Black Power, and the Reinvention of Racial Liberalism. University of Pennsylvania Press.
Food & Water Watch. (2024). The State of Public Water in the United States. foodandwaterwatch.org.
Gartner. (2024, October). PC and Mobile Market Forecast Q4 2024. Gartner Research.
Google. (2024). Environmental Report 2024: Water Stewardship. Google Sustainability.
Government Accountability Office. (2021). Artificial Intelligence: DOD Needs to Improve Acquisition Guidance (GAO-21-404). gao.gov.
Haas, J. (2019). The Assassination of Fred Hampton: How the FBI and the Chicago Police Murdered a Black Panther. Chicago Review Press.
Hampton, F. (1969). Power Anywhere Where There’s People [Speech]. Chicago, IL.
Haywood, H. (1978). Black Bolshevik: Autobiography of an Afro-American Communist. Liberator Press.
Lawrence Berkeley National Laboratory. (2025). United States Data Center Energy Usage and Water Consumption Report. LBNL.
McGurty, E.M. (2007). Transforming Environmentalism: Warren County, PCBs, and the Origins of Environmental Justice. Rutgers University Press.
Medsger, B. (2014). The Burglary: The Discovery of J. Edgar Hoover’s Secret FBI. Knopf.
Meta. (2024). Sustainability Report 2024. Meta Investor Relations.
Micron Technology. (2024, September 25). Q4 FY2024 Earnings Call [Transcript]. Micron Investor Relations.
Microsoft. (2024). Environmental Sustainability Report 2024. Microsoft Corporate Responsibility.
Morello-Frosch, R., & Jesdale, B.M. (2006). Separate and Unequal: Residential Segregation and Estimated Cancer Risks Associated with Ambient Air Toxics in U.S. Metropolitan Areas. Environmental Health Perspectives, 114(3), 386-393.
Pastor, M., Sadd, J., & Hipp, J. (2001). Which Came First? Toxic Facilities, Minority Move-In, and Environmental Justice. Journal of Urban Affairs, 23(1), 1-21.
Reuters. (2026, January 28). Nvidia Shares Fall After U.S. Expands China Chip Export Curbs. Reuters.
Shakur, A. (1987). Assata: An Autobiography. Lawrence Hill Books.
SK Hynix. (2024, October). Q3 2024 Earnings Presentation. SK Hynix Investor Relations.
The Information. (2024, December). Inside Apple’s AI Deals with Google and OpenAI. The Information.
The Information. (2024, September 27). OpenAI Projects More Than $3.7 Billion in Annual Revenue. The Information.
The Information. (2026, January). Microsoft Tests Claude Amid OpenAI Governance Concerns. The Information.
TrendForce. (2024, November). DRAM Market Analysis: CXMT Pricing Strategy and Market Impact. TrendForce Research.
United Church of Christ Commission for Racial Justice. (1987). Toxic Wastes and Race in the United States: A National Report on the Racial and Socio-Economic Characteristics of Communities with Hazardous Waste Sites.
Uptime Institute. (2024). Global Data Center Survey 2024: Cooling Technologies and Trends. Uptime Institute.
Uptime Institute. (2025). Cooling Technology Trends 2020-2025. Uptime Institute Intelligence.
U.S. Department of Commerce, Bureau of Industry and Security. (2026, January 27). Commerce Department Strengthens Export Controls on Advanced Computing Semiconductors. bis.doc.gov.
Weems, R.E., & Randolph, L.A. (2009). Business in Black and White: American Presidents and Black Entrepreneurs in the Twentieth Century. NYU Press.
White House. (2025, January 21). President Trump Announces Stargate AI Infrastructure Project [Press release]. whitehouse.gov.
Wright, N. (1968). Black Power and Urban Unrest. Hawthorn Books.