This is a complicated story. It’s complicated because the U.S. and China have one of the most tangled relationships in the world: rivals, partners, customers, adversaries, all at once. And it’s complicated because most of what we’re talking about here isn’t shiny apps or headlines; it’s the invisible plumbing — databases, networks, and systems that ordinary people never see but governments depend on.
That’s why I want to walk slowly, piece by piece, through the moments that stood out to me as strange. What you’ll notice is a repeating pattern; the same three beats, over and over:
Crisis exploitation: a disaster hits, and surveillance suddenly becomes more acceptable.
Infrastructure positioning: Oracle slides in as the quiet backbone, not the visible face.
Regulatory arbitrage: they work through intermediaries, or position themselves as “trusted partners,” keeping plausible deniability if things go sideways.
Once you see the pattern, you’ll notice how often the timing feels just a little too close for coincidence.
Tiananmen, Oracle Bones, and the Power of a Name (1989–1991)
In June 1989, Beijing’s Tiananmen Square became the site of one of the most haunting events in modern memory. For weeks, students and workers gathered in the square, demanding change: freedom of speech, an end to corruption, a say in their country’s future. The atmosphere was electric, filled with hope.
The government’s answer was brutal. Tanks rolled in. Soldiers fired live rounds into unarmed crowds. By dawn, the square and the surrounding streets bore the marks of a massacre. The true number of dead has never been confirmed — hundreds, maybe thousands. Even today, the subject is so tightly censored that Chinese citizens grow up scarcely aware of it. If you’ve never learned the details, I’d encourage you to pause here and look it up. It is a reminder of just how far governments will go to protect their grip on power.
After Tiananmen, the Communist Party faced a profound legitimacy crisis. Their solution came in two parts. First, they launched a sweeping Patriotic Education Campaign, shifting focus from Marxism to China’s 5,000-year-old heritage. Suddenly the Great Wall, Confucius, and China’s ancient past were everywhere. Second, they invested heavily in modern surveillance systems — computers, databases, and networks to help police maintain stability and control dissent more efficiently.
And here’s where my research took an unexpected turn. Around this exact time (1989 to 1991) Oracle, the American database company with CIA roots, was quietly entering China. That timing alone is striking, given that most Western firms were pulling back after the crackdown. But then I noticed something even stranger.
At the same moment China was reviving oracle bones (cracked animal bones inscribed with the earliest Chinese writing), Oracle adopted the Chinese name 甲骨文, which literally means “oracle-bone inscriptions.”
At first, I thought it had to be coincidence. Oracle bones are archaeological artifacts; Oracle the company is a Silicon Valley database vendor. But digging deeper, I discovered that Oracle’s Taiwan agent, Feng Xingjun (a man steeped in Chinese history) had registered the trademark. A literal translation of “Oracle” into Chinese (“Prophecy Company”) would have sounded mystical or absurd. But 甲骨文? That carried cultural legitimacy at precisely the moment the government was pushing it.
So think about the alignment:
June 1989: A massacre shakes the regime.
Late 1989–1991: The Party revives ancient symbols like oracle bones to shore up legitimacy.
1989–1991: Oracle enters China, branding itself as the digital heir to the oldest Chinese records of statecraft.
2001: Twin Disasters, Twin Databases
In September 2001, the United States was struck by the 9/11 terrorist attacks. For Americans, it was a day that changed everything; a morning of fear and grief that hardened overnight into a political demand for security at any cost. And almost immediately, Larry Ellison stepped into that moment. He offered Oracle’s software for free to build a national ID system: one database to hold every American’s personal information, instantly accessible to government agencies. It was bold, it was controversial, and it was perfectly timed.
At first, I thought of this as a uniquely American story, born of a singular tragedy. But then I noticed what had been happening on the other side of the Pacific in that same year. China’s Golden Shield Project, which would become the technical backbone of the world’s most ambitious surveillance state, was approved just months earlier.
The dates are specific:
April 25, 2001: the State Council signs off on the project.
May 30, 2001: the Planning Commission issues its formal approval notice.
July 11, 2001: a joint directive sets implementation into motion.
By the time 9/11 happened in the U.S., Golden Shield was already underway. Just 62 days before the attacks, Beijing had given it marching orders.
And what was Golden Shield? Not a single tool, but an entire infrastructure: police databases linked across provinces, ID and household records digitized, internet activity folded into centralized monitoring. It was the skeleton of what became the Great Firewall; censorship and surveillance as one system.
Now, here’s the pattern. In the same calendar year, both the U.S. and China made decisive moves toward centralized surveillance databases. For the U.S., it took a single cataclysmic event to make Ellison’s national ID pitch sound plausible. For China, it was a slower convergence of pressures: separatist unrest in Xinjiang and Tibet, the banned Falun Gong movement, and an internet that spread faster than the censors could contain. Different triggers, same outcome.
Lining up the dates, the symmetry is uncanny:
Spring–Summer 2001: Golden Shield is approved in Beijing.
Fall 2001: 9/11 strikes, and Ellison pitches Oracle as the backbone of a national ID.
Coincidence? Maybe. But it looks more like a global pivot: the year both of the world’s great powers decided that the future of governance lay in centralized surveillance databases. Oracle wasn’t just a participant; it was positioning itself as the backbone, East and West.
The Export/Import Loop (2010s)
By the early 2010s, Oracle’s technology wasn’t just background plumbing. It was moving through a feedback loop: tested in one place, sold in another, then celebrated as “field-proven.” I didn’t see the pattern at first. It came into focus only after I traced a handful of receipts.
Start in 2011–2012. During the NATO Summit protests in Chicago, the Chicago Police Department pulled in Endeca (a search-and-analytics platform that Oracle had just acquired in 2011) to fuse police records with real-time social media. Tweets from organizers flowed into the same dashboards as 911 calls and crime reports. That’s not conjecture; it’s documented in investigations that surfaced internal materials and even presentation video describing how CPD used Endeca Information Discovery to merge protesters’ tweets with police data during the 2012 protests.
If you’re wondering how we learned that, two sources matter. First, a paper trail from Chicago’s own tech build shows the CLEAR program (Citizen and Law Enforcement Analysis and Reporting) was developed in partnership with Oracle, years before NATO, to knit CPD’s data into a central warehouse; the foundation that made rapid “situational awareness” possible. That partnership is spelled out in DOJ-funded research and evaluations. Second, the later leak and rediscovery of Oracle marketing slides made the protest-monitoring workflow visible to outsiders. Those materials tied Endeca, tweets, and policing together in a way the public could finally see.
Here’s where it turns. Within a short window, Oracle reps used the Chicago example to pitch Chinese police. Reports based on Oracle’s own China-facing decks and partner pages show company staff highlighting the Chicago protest deployment as a selling point. In at least one case, the materials claimed CPD used Endeca and other inputs to predict crime within a three-day window'; down to features like weather, call volume, and arrest data. And it was positioned as a ready-made public security solution for provincial PSBs.
By 2018, Chinese provincial police (Liaoning among them) were listed as users of Oracle analytics in company marketing. That arc matters: business software tested on U.S. protesters, then marketed to Chinese public security, then adopted by provincial departments. It’s not a straight export; it’s a loop.
A few technical notes round out the picture. Oracle’s Endeca Information Discovery wasn’t a niche toy; it was a hybrid search-analytics engine built to ingest structured and unstructured data at speed, with connectors and enrichment that made social streams, records, and logs clickable in one pane. In other words, it was purpose-built for exactly the kind of multi-source monitoring Chicago ran, and that Chinese police would want to scale. Oracle’s own documentation makes the capability set plain.
Put it all together and the pattern is hard to ignore:
2012: CPD fuses tweets and police data with Oracle/Endeca during NATO protests; DOJ-funded CLEAR reports confirm the Oracle backbone; leaked Oracle slides corroborate protest-specific workflows.
Mid-2010s: Oracle reps use the Chicago case in China-facing marketing, presenting Endeca as a policing solution.
2018: Provincial police in Liaoning and elsewhere appear as customers in Oracle marketing.
Once you see the receipts, Chicago and Liaoning stop looking like separate stories. They look like two stops on the same route.
COVID-19: The Health Data Pivot (2019–2020)
In 2019, months before COVID became a global headline, Oracle was already making a curious move in China. The company abruptly laid off hundreds of engineers in its Beijing and Shenzhen R&D centers. Nearly 1,600 employees were affected, and workers protested in the streets with banners that read: “High profits, why layoffs?” and “Keep politics out of technology.” For a company that had spent decades embedding itself in Chinese infrastructure, this was more than a cost-cutting exercise. It looked like retrenchment — a quiet step back.
At the same time, Larry Ellison was growing more vocal about China as a “threat” to U.S. competitiveness. Publicly, Oracle began reframing its China strategy, pivoting its emphasis away from local R&D toward cloud computing and government contracts elsewhere. On the surface, this looked like a business decision. But in hindsight, it feels like positioning. Oracle was slimming down its exposure in China just as the world was about to be plunged into a health crisis that would make data (specifically health data) more valuable than ever.
Then came COVID-19.
Almost overnight, China rolled out health-code apps. Ordinary citizens now lived by QR codes that governed whether they could travel, shop, or even leave their homes. These apps tracked testing results, quarantine orders, and location history. To outside observers, it looked miraculous — as if China had built a national bio-surveillance system in mere weeks. But the speed told another story. You don’t flip a switch like that unless the infrastructure is already there. The telecom integrations, the public security databases, the citizen ID backbones. All of it had been quietly prepared long before.
And here’s the part we can’t gloss over: for surveillance to truly scale, you need more than phone records or internet histories. You need biological data. Medical records. Test results. Health movements that are intimate and deeply personal. COVID gave governments the perfect cover to collect it — and once that door is opened, it’s very difficult to close.
Meanwhile, in the United States, Ellison made his move. In March 2020, he personally called President Donald Trump and offered Oracle’s databases to manage the government’s pandemic response. Oracle then built a national COVID registry in record time, tracking vaccine trial data and treatments. The speed was uncanny. It felt less like a new project and more like a template ready to be deployed; a U.S. mirror of the systems that had just gone live in China.
So line up the dates:
2019: Oracle retreats from China’s R&D sector, shifts tone on Beijing.
Early 2020: China launches health-code apps with astonishing speed.
Spring 2020: Oracle pivots in the U.S., building a national COVID database almost instantly.
At first glance, these are parallel stories. But step back, and the pattern is harder to ignore. Oracle reduced its footprint in China just before the crisis, then reemerged in the U.S. as the architect of pandemic health surveillance. China wanted health data for social control; Ellison wanted health data as the new frontier of governance infrastructure. The crises were different, but the playbook was the same.
What looks like coincidence begins to feel like preparation.
TikTok: The Present-Day Test Case (2020–2025)
TikTok wasn’t always treated as a national security threat. It started as an app for fun: in 2016, ByteDance launched Douyin in China; in 2017, it created TikTok for global markets; and in 2018, it bought Musical.ly, giving the platform a huge U.S. audience. At the time, it was mostly about lip-syncs, dance routines, and quick comedy. But as we’ve seen before, tools that seem lighthearted can, under the right pressures, become something far more serious.
In October 2020, Beijing updated its export-control law. By December 1, 2020, algorithms for personalized recommendations (like the one behind TikTok’s “For You” feed) were officially added to the restricted list. That meant China’s government had to approve any transfer of the technology outside its borders. The timing wasn’t random. In Washington, the Trump administration was threatening to ban TikTok unless ByteDance sold its U.S. operations. By locking the algorithm, Beijing ensured the most valuable part of TikTok could not be handed over.
The June 2022 Leak and Oracle’s Role
On June 17, 2022, BuzzFeed News published leaked audio of TikTok employees saying, “Everything is seen in China.” That story landed hard in Washington. And then, on that very same day, TikTok announced Project Texas. From that point forward, all new U.S. user data would live on Oracle Cloud Infrastructure, and older data in Virginia and Singapore would be erased. Oracle would also audit TikTok’s source code and moderation systems.
To put that plainly: Oracle did not buy TikTok, but it took custody of where American data is stored and gained access to how TikTok decides what content to promote or remove. Normally, shifting an app of TikTok’s size onto a new cloud provider would take months or even years. The fact that this was announced the same day the leak appeared suggests the plan was already waiting in the wings. The leak created the pressure; Oracle provided the solution.
On the night of January 18, 2025, TikTok went offline in the United States. For nearly twelve hours, the app was unusable. It came back on January 19 — the very day a U.S. law required TikTok to divest or face a nationwide ban. When the app returned, TikTok issued a public statement thanking the President of the United States for his support.
It looked less like a glitch and more like a signal. The blackout landed exactly when the ban was supposed to take effect, and its quick restoration showed the decision wasn’t about technical capacity but political permission. For Oracle, already the custodian of U.S. data, this episode reinforced why its role mattered: TikTok’s future in America now depended on the presence of a trusted American partner in the middle.
In August 2025, TikTok introduced new Community Guidelines. The updates included expanded categories for “misinformation,” “foreign influence,” and “national security.” Account bans spiked, and creators began noticing content being removed far more quickly than before.
A month earlier, in July 2025, TikTok had hired Erica Mindel as its Public Policy Manager for Hate Speech. Mindel had previously served as an instructor in the Israeli Army and later worked as a contractor for the U.S. State Department. Her role included shaping antisemitism policy, engaging with legislators, and monitoring hate speech trends. The hire came after Israeli President Isaac Herzog met with TikTok leadership in 2024, urging stronger action on antisemitic content. These changes didn’t alter Oracle’s technical oversight, but they did show TikTok layering new kinds of political pressure and external expertise into its governance at the very moment regulatory scrutiny was at its peak.
The Pattern
Line up the dates and the pattern is clear:
Dec 2020: Beijing locks TikTok’s algorithm under export-control law.
June 17, 2022: BuzzFeed leak hits; Project Texas is announced the same day.
Jan 18–19, 2025: TikTok blackout coincides with a divest-or-ban deadline.
July–Aug 2025: TikTok hires security-linked policy staff and rewrites guidelines under political pressure.
Each crisis produces the same response: Oracle moves in, the arrangement tightens, and what was supposed to be temporary begins to look permanent. TikTok may still be known as the app for short videos, but at this point it functions as a test case in how U.S.–China rivalry, corporate power, and surveillance politics converge. China keeps the algorithm; Oracle keeps the data; Washington keeps moving the deadlines. And with each delay, the patchwork solution becomes harder to see as temporary at all.
Looking at TikTok in this light, you start to see more than just an app under pressure; you see the familiar outline of Larry Ellison’s playbook. A crisis creates urgency, Oracle positions itself as the indispensable backbone, and regulators are offered a compromise that looks safer but quietly cements Oracle’s power. The question that follows naturally is whether this is coincidence or conviction.
Larry Ellison’s Philosophy of Surveillance
If the TikTok story shows Ellison’s playbook in action, then his own words reveal why the playbook looks the way it does. For decades, Ellison has been unusually candid about the world he envisions, and when you line up his statements with the crises of each era, the alignment is difficult to ignore.
2001: A National ID for America
In the weeks after September 11, 2001, when fear of terrorism was still raw, Ellison gave interviews offering Oracle’s database software free of charge to the U.S. government. His proposal was sweeping: build a national ID system, one database containing every American’s personal details, available instantly to law enforcement. He framed it as a patriotic duty, a simple act of modernization. Critics saw it differently — a dangerous step toward total surveillance. Yet Ellison was unapologetic, insisting that the real risk was not collecting enough data.
2013: Defending the NSA
When Edward Snowden’s leaks in 2013 exposed mass surveillance programs at the National Security Agency, many technology leaders distanced themselves from government spying. Ellison did the opposite. In an interview with CBS News, he defended the NSA, asking, “Who’s ever heard of this abuse? Name one example.” To him, the problem wasn’t overreach; it was that people didn’t appreciate the security benefits of centralized monitoring.
2024: The AI Panopticon
Most recently, in September 2024, Ellison addressed a crowd at Oracle’s annual conference and spoke about artificial intelligence in almost utopian terms — not as a tool for convenience, but as a tool for behavior control. “AI will help keep everyone on their best behavior,” he said, describing a future where constant monitoring becomes the foundation for order. What others might call a panopticon, Ellison described as progress.
The Pattern of Belief
Put these moments together and a philosophy comes into focus. For Ellison, databases are not neutral tools; they are the backbone of governance. He has pitched them after crises, defended them when their use was exposed, and celebrated them as instruments of social discipline. The trajectory is consistent: collect more, centralize more, monitor more.
And when you look back at the story we’ve just walked through — from Oracle’s symbolic entry into China after Tiananmen, to Golden Shield, to the health-data pivot during COVID, to Oracle’s custodianship of TikTok today — it reads less like coincidence and more like conviction. Ellison has been telling us for decades what kind of world he wants to build. The only question left is whether we are willing to believe him.
And while this chapter has focused on China and the United States, Israel has already appeared at the edges of the story. In the next part of this series, we will shift to that relationship directly — one that is less about cultural symbolism and more about raw power. Much of what I’ve uncovered connects Oracle to Israel’s experiments with social surveillance and the biotech tools developed within the IDF. If China was the proving ground for Ellison’s ability to embed quietly in the background, Israel shows what happens when the strategy becomes explicit.
Government Docs
Beijing Oracle Software Systems Ltd. Corporate Registry (1991)
Oracle’s Chinese Trademark “甲骨文” (Oracle Bone Inscriptions) — China Trademark Office
Reporting
Associated Press: How U.S. Tech Companies Helped Build China’s Surveillance State (Dec. 2024)
BuzzFeed News Leak: “Everything is seen in China” (June 17, 2022)
Wall Street Journal: TikTok to Route U.S. Data to Oracle (June 2022)
South China Morning Post: TikTok US Leadership Shake-Up (Apr. 2025)
The New Arab: TikTok Hires Former Israeli Soldier for Hate Speech Policy (July 2025)
The Intercept: Chicago NATO Protests, Oracle Endeca Surveillance (2012)
Chicago Tribune / DOJ Reports: CPD CLEAR Database Built with Oracle (2004–2007)
Background & Context
Direct Quotes (Smoking Guns)
“Everything is seen in China.” — TikTok staffer in leaked audio (BuzzFeed, June 2022).
“We are going to have supervision … Citizens will be on their best behavior because we’re constantly recording and reporting everything that’s going on.” — Larry Ellison, Sept. 2024 conference, as reported by Ars Technica.
“Who’s ever heard of this abuse? Name one example.” — Ellison defending NSA surveillance after the Snowden leaks (CBS News, 2013).
“Oracle will provide the software free of charge … a national database of American citizens.” — Ellison’s offer after 9/11 (2001 interviews, archived).
