As I write this, I’m surrounded by handmade quilts in my parents’ neighbor’s art studio, borrowed at dawn because it’s the only quiet place to work. There’s something about documenting biological surveillance infrastructure while sitting among objects someone spent months stitching together with care. One woman makes quilts to keep people warm. I trace how your genome flows from your doctor’s office into databases that will predict your children’s futures before they’re born. It would be funny if it weren’t so terrifying.
But that’s the thing with power, right? It never announces itself. It doesn’t show up wearing a villain costume. There are just press conferences where tech executives stand next to a president and announce five hundred billion dollars in AI infrastructure investment. Cancer research. Competing with China. Good jobs. The kind of thing where objecting makes you seem unreasonable.
I think about the January 21st announcement a lot. I’ve written about it a couple of times on this platform already, tracking different threads of what Larry Ellison’s been building, but upon my nth revisit of that press conference footage, a few things clicked for me that I’d been missing. At first it just sounded like the usual tech hype, impressive business jargon about innovation and infrastructure and American competitiveness, the kind of thing you half-listen to while scrolling through other tabs. But somewhere around the fourth or fifth viewing, the scale of what they were actually discussing stopped being abstract corporate language and started feeling like something else entirely.
Nobody seems to be asking why you’d need ten gigawatts of power to ‘cure cancer’. That’s ten nuclear power plants running simultaneously. That’s not research infrastructure for a few thousand patients. That’s processing capacity for hundreds of millions of genomes.
And nobody remarked that standing next to Trump and Larry Ellison was Masayoshi Son, casually mentioning the UAE would invest through something called G42. A company Trump didn’t name. A company most Americans have never heard of. Which got me curious.
Because the thing about intelligence agencies, is they don’t pull names out of a hat. There’s usually some internal logic, some bureaucratic convention or inside joke (~Men~). So I looked up Project Stargate to see if the name had been used before. Turns out it had. Not just the 1994 movie about portals. There was a real Project Stargate, run by U.S. intelligence from 1972 to 1995.
The files are declassified now. Which is how I know the U.S. government spent twenty-three years and twenty million dollars trying to use psychics to spy on the Soviets. Remote viewing, they called it. The idea was you could train operatives to psychically perceive distant locations and predict what the enemy would do next.
They were completely serious. Real intelligence officers working on it at Fort Meade. And they got some results, though not consistently. A guy would psychically describe a Soviet facility and satellite photos would confirm details he couldn’t have known. The problem was accuracy rates hovering between fifteen and seventy-seven percent depending on who was evaluating. Expensive guessing, basically.
So in 1995, the CIA commissioned an independent evaluation that concluded the program was “not operationally useful.” They shut it down and declassified everything. But the goal didn’t die. The goal was always “see everything and predict what happens next.” They just needed better tools than psychics trying to astral project into Soviet military bases.
And sitting there at dawn surrounded by quilts, reading declassified documents about failed psychic espionage programs, it hit me. In 2025, those tools finally exist.
What’s Already Built
I’ve spent the last year documenting Larry Ellison’s empire. Media consolidation through his son’s acquisition of CBS and Paramount. The pursuit of Warner Brothers for CNN and HBO. Healthcare data through Oracle’s Cerner acquisition with 9.5 million patient records. Behavioral surveillance through TikTok ownership. Biometric data through facial recognition. Medical records through Oracle Health.
I thought I understood what he was building. Surveillance capitalism at scale. The kind Shoshana Zuboff documented, where companies harvest and sell predictions about your behavior. But Stargate revealed what I’d been missing. This isn’t about predicting what you’ll buy or watch. This is about genomic data itself. The permanent hereditary code that makes all other surveillance inescapable.
Because here’s what makes this different: You can delete your Facebook account. You can change your phone number and move cities. You can wear masks and avoid cameras and use cash if you’re determined enough. Behavioral surveillance is theoretically escapable.
But you cannot change your genome. Once it’s in a database, it’s there forever. And because it’s hereditary, you’re not just making a decision for yourself. You’re making it for descendants who haven’t been born yet, who will inherit both your genetic material and the surveillance infrastructure built around it.
Privacy scholars have shown that genomic data is uniquely reidentifiable even when stripped of obvious identifiers. Yves-Alexandre de Montjoye’s research found that just four data points from a genomic sequence can uniquely identify ninety-five percent of individuals. They “de-identify” the data by replacing your name with “Subject 15927,” but it remains scientifically identifying. You can hide a genome behind a number, but you can’t make it anonymous.
How It Actually Happens
So here’s how it works in practice. You go to your doctor for a routine checkup. Blood work, maybe a liquid biopsy for early cancer detection. Your doctor hands you a consent form. Twenty pages of legal jargon in eight-point font. Buried somewhere in paragraph fourteen: “Sample may be used for research purposes.”
Research sounds noble. Finding cures, advancing science. You sign because you need the test and nobody expects you to read twenty pages before a medical procedure. But “research purposes” has become infinitely elastic. Your genome gets extracted and digitized, aggregated into commercial databases, analyzed by AI you’ll never see, used to generate predictions about your health and behavior and productivity, and sold to insurance companies and employers and anyone willing to pay.
Your data can be used to make inferences about relatives who never consented. Your children will inherit genetic profiles in databases before they’re born. You cannot delete it later. You cannot opt out retroactively. The consent was theater. You thought you were helping cure cancer. You were authorizing permanent biological surveillance.
The International Angle
What fascinates me about Stargate is how it launders surveillance through international partnerships. Follow the thread. Israeli military intelligence develops genomic extraction technology through companies like Imogene AI, founded by IDF Unit 8200 officers. Larry Ellison personally invested twenty-three million dollars in Imogene. Not Oracle. Him personally. Then the Abraham Accords create legal frameworks for Israeli technology to transfer to UAE state apparatus.
And operating on the receiving end is G42. A UAE company whose chairman is the country’s National Security Adviser. Not a former official. The current guy running UAE intelligence also runs this “AI company.” Their history reads exactly like you’d expect. They created ToTok, a messaging app the New York Times exposed as comprehensive surveillance. They learned genomic collection from BGI, a Chinese company now on the U.S. Entity List for national security concerns. They operated DarkMatter Group, hiring former NSA operatives to build targeting systems against journalists and activists.
This is the company operating Stargate UAE, the largest announced site. This is who has a board seat in the joint venture. This is who Trump very carefully didn’t mention by name during that press conference.
American companies provide the infrastructure. Oracle controls health records. OpenAI provides AI processing. Microsoft invested 1.5 billion in G42 before quietly backing away when the surveillance history became too toxic, though the technical integration remains fully operational. The whole thing gets marketed as healthcare innovation. Cancer research. American competitiveness. Perfect marketing, really. Who’s against curing cancer?
But this is how you build discriminatory systems nobody questions. You hide them inside objective technical infrastructure. Except now we’re operating at the level of genetic code itself. A transnational biological surveillance alliance where Israeli technical expertise plus Emirati willingness to deploy plus American infrastructure equals a global system no single government can regulate and no population can reject.
What It Becomes
Genome-wide association studies have identified thousands of genetic variants associated with complex traits. Intelligence, psychiatric conditions, behavioral tendencies. The associations are scientifically valid, though the predictive power isn’t perfect. But it doesn’t need to be perfect. It just needs to be profitable.
So AI analyzes your genome and generates polygenic risk scores. An insurance company buys these predictions without telling you. They charge ten times normal premiums or deny coverage based on “actuarial risk assessment.” An employer’s algorithm filters you out based on genetic markers for high healthcare costs. You never know why. You get the generic rejection email. A bank denies your mortgage because AI predicted high lifetime medical expenses. You ask what factors. They say proprietary trade secret. You have no recourse.
GINA, the Genetic Information Nondiscrimination Act, fails to protect against this because it only covers explicit discrimination based on known genetic information. Not AI-derived predictions. Not algorithmic discrimination. The law was written for a previous era. This creates a biological caste system. Optimal genomes get best rates, hired first, approved for loans. High-risk genomes become uninsurable, unemployable, denied capital. A permanent biological underclass created through algorithmic exclusion hidden in proprietary code nobody’s allowed to examine.
What I Can’t Prove
I’ve spent hundreds of hours on this investigation. Read thousands of pages of documents. Traced money through corporate filings. Mapped infrastructure through patents and contracts. But I cannot answer the one question that matters most. Are they doing it yet? Is Oracle Health data currently flowing into Stargate AI training? (I’m afraid of the answer).
If this is happening, they’re hiding it carefully behind NDAs and corporate privacy policies. What I can prove is that every piece of necessary infrastructure exists and is operational. Oracle owns the health records through Cerner. G42 operates the processing facilities. OpenAI provides the algorithms. Tempus aggregates data from forty million patients. The technical integration would be trivial.
The legal loopholes are in place. The financial incentives align perfectly. Every company involved has a documented history of doing exactly this kind of surveillance when they thought they could get away with it. They built the pipes. And nobody builds pipes for fun. Nobody spends five hundred billion dollars on infrastructure they don’t intend to use.
The Window
So here we are. By 2029, Stargate’s full capacity will be operational according to their own timeline. Once that happens, once data flows at scale and predictions integrate into decision-making systems across every sector, dismantling becomes nearly impossible. Not because the technology can’t be shut down, but because too many powerful interests will depend on it continuing.
This is the moment where we still have choices. We could expand GINA to cover all genomic uses and AI predictions. We could require algorithmic transparency. We could ban foreign government control of American health data. We could establish patient ownership of genomic data as a fundamental legal principle.
Or we can do nothing. We can assume best intentions from companies with documented surveillance histories. We can trust that having the capability doesn’t mean they’ll use it. And then we can wake up in five years living in a biological caste system we didn’t consciously choose but somehow collectively accepted.
The surveillance won’t announce itself. There won’t be a press conference. It’ll be gradual, invisible. Denied insurance for unexplained reasons. Kids filtered by hiring algorithms identifying genetic risk markers. Grandchildren inheriting genomic scores determining their opportunities before birth.
And maybe you’ll trace it back to that consent form you signed in 2025, when you thought “research purposes” meant finding cures rather than building a permanent biological surveillance state.
A Note To Readers
If you made it this far, thank you. This investigation has consumed months of my life. Every claim is sourced with links to primary documents, academic research, and investigative journalism. I encourage you to verify this yourself. Don’t trust me. Look at the documents. Read the scholarship. Follow the money. See if the pattern holds.
And if you think this matters, please share it. Most people don’t know this is happening. They hear “AI infrastructure” and “cancer research” and assume it’s benign. The power of independent journalism is that we can say things institutional media won’t. But that only works if people actually see the work.
This is the threshold. This is the moment where choices still exist. Once we cross it, there’s no going back. You can’t delete data from databases after it’s been copied and distributed across jurisdictions and integrated into decision-making systems too powerful to overcome.
(Happy New Year!)
Sources Cited
Key Documents & Evidence
CIA – STARGATE Declassified Archive (1972–1995 U.S. remote viewing program)
https://www.cia.gov/readingroom/collection/stargateCIA – STAR GATE PROJECT: AN OVERVIEW (Program structure, cost, evaluation)
https://www.cia.gov/readingroom/docs/CIA-RDP96-00789R002800180001-2.pdfMicrosoft – “Microsoft invests $1.5 billion in Abu Dhabi’s G42 to accelerate AI development and global expansion” (Official press release)
https://news.microsoft.com/source/2024/04/16/microsoft-invests-1-5-billion-in-abu-dhabis-g42-to-accelerate-ai-development-and-global-expansionReuters – “Microsoft to invest $1.5 billion in Emirati AI firm G42 for minority stake” (Board seat, security assurances, DarkMatter links)
https://www.reuters.com/markets/deals/microsoft-invest-15-bln-emirati-ai-firm-g42-new-york-times-reports-2024-04-16Globes – “Precision medicine co Imagene AI raises $23m led by Larry Ellison” (Ellison’s investment, Israeli origin, precision medicine focus)
https://en.globes.co.il/en/article-precision-medicine-co-imagene-ai-raises-23m-led-by-larry-ellison-1001514613Ropes & Gray – “Bureau of Industry and Security Imposes Export Restrictions on BGI Group, the World’s Largest Genomics Company” (BGI entities on U.S. Entity List)
https://www.ropesgray.com/en/insights/alerts/2023/04/bis-imposes-export-restrictions-on-bgi-group-the-worlds-largest-genomics-companyTempus – multimodal data library (over 40 million research records and linked genomic data)
https://mlq.ai/research/tempus-ai-tem-deep-diveTempus – “Tempus Introduces Fuses, A Program Designed to Transform Clinical-Genomic Research Collaboration” (40M+ research records stat)
https://finance.yahoo.com/news/tempus-introduces-fuses-program-designed-123000980.htmlAl Jazeera – “Calling app ToTok used as ‘spying tool’ by UAE: Report” (Mass surveillance, DarkMatter/PAX AI links)
https://www.aljazeera.com/news/2019/12/23/calling-app-totok-used-as-spying-tool-by-uae-reportNature Communications – Rocher, Hendrickx & de Montjoye, “Estimating the success of re-identifications in incomplete datasets using generative models” (re-identification risk building on Sweeney’s work)
https://www.nature.com/articles/s41467-019-10933-3Notre Dame Law Review – Jessica L. Roberts, “The Genetic Information Nondiscrimination Act as an Antidiscrimination Law” (limits of GINA)
https://scholarship.law.nd.edu/ndlr/vol86/iss2/3What is Stargate? President Trump announces $500 billion investment in AI Infrastructure project
