The Drey Dossier

The Drey Dossier

Playback speed
×
Share post
Share post at current time
0:00
/
0:00
Transcript

Tony Blair and The Uncle Larry Doctrine

The Drey Dossier's avatar
The Drey Dossier
Jan 13, 2026

The Tony Blair Institute’s plan to build AI tools for governments isn’t exactly a “new” venture. It’s the execution of a long-standing doctrine I call: the Uncle Larry Doctrine. It’s a strategic play where the goal is not necessarily to create the best product, but to become the indispensable floor you’ve built beneath everyone’s feet – including your partners.

I’ve been covering Ellison and Oracle for months, so when a Google alert hit my inbox three days ago about Tony Blair’s Institute building AI tools to “rival Palantir,” I was filled with that familiar sense of dread that only pattern recognition against your will can give you.

The story comes from Democracy for Sale, and their reporting reveals that the Tony Blair Institute for Global Change has set up an “AI Incubator” targeting government clients in the Middle East and Asia. According to TBI’s financial filings, they posted a $4.3 million loss in 2024 while pouring money into this pivot, and insiders say the products are basically PowerPoint slides right now, but none of that matters because the products were never really the point.

To understand why, you need to know about the $348 million that Ellison has pumped into Blair’s Institute since 2021. Former employees say working at TBI felt like “doing tech sales for Oracle,” and there’s a story from Kenya where officials got so exhausted from Oracle pitches they started calling Blair “Uncle Larry.” That £257 million was never a donation; it was strategic capital deployed through a former Prime Minister who looked at George Bush retreating to Texas to paint veterans as a penance for the Iraq War and thought to himself: I could do so much more.

Blair’s motivation isn’t complicated: relevance. The man ran Britain for a decade, led his country into a war based on lies that killed hundreds of thousands of people, and instead of facing any meaningful accountability, he built a 45-country operation that lets him walk into ministries and presidential palaces as a peer rather than a pariah. He currently serves as chair of the European Council on Tolerance and Reconciliation’s initiative on Gaza peace, which is the kind of appointment that would be darkly comedic if it weren’t actually happening. Ellison’s money keeps Blair’s machine running, and in exchange, Blair opens doors that Oracle’s sales team could never unlock on their own, particularly in the Gulf states where Blair’s relationships run deep and where the real surveillance budgets are flowing.

This is the part that doesn’t get discussed enough: Ellison isn’t just funding Blair for access to governments, he’s funding Blair for access to Middle Eastern money. The Gulf states are spending billions on AI and surveillance infrastructure, and they’re not buying from American tech companies directly because the optics are complicated. But a former British Prime Minister running a nonprofit focused on “global change” and “AI governance”? That’s a much easier handshake. Blair is Ellison’s gateway to sovereign wealth funds and defense ministries that would otherwise be difficult to reach, and that access is worth far more than $320 million.

The real story starts in 2016, when Ellison tried to buy Palantir and Peter Thiel told him no. Oracle already had surveillance tech from their Endeca acquisition, but Palantir was ahead on AI, and Ellison wanted that edge. You have to understand something about this man: it’s not enough that he wins, everyone else must lose. So when Thiel rejected him, Ellison went quiet, and for five years it looked like he’d moved on, which is exactly what he wanted everyone to think.

Three things happen simultaneously in this period: Ellison starts funding Blair’s Institute, Oracle builds a nine-story underground data center in Jerusalem serving as Israel’s military and intelligence hub, and Oracle launches a classified infrastructure project with the Israeli Air Force. Nobody connected these dots because Ellison had spent 48 years teaching people not to watch him, to think of Oracle as a boring enterprise database company, reliable and unsexy and utterly beneath concern.

The trap springs in April 2024 when Oracle and Palantir announce a partnership. On paper it looks mutually beneficial, with Palantir getting access to sovereign cloud infrastructure in strategic regions, especially Israel where they’d just signed a major defense partnership, and Oracle getting access to Palantir’s refined AI targeting systems. Thiel agrees to move Palantir’s entire Foundry platform onto Oracle’s cloud, and this is where I believe he made his strategic miscalculation.

You know what’s not a miscalculation? Subscribing to the Rough Riders tier for $5/month

When you host someone’s platform on your cloud, you’re not just storing their data; you can observe how their systems perform under stress, see their computational patterns, watch how they handle massive real-time datasets. You’re not stealing code, that would be illegal, but you gain unprecedented visibility into their technical architecture. From January 2024 through mid-2025, Palantir refined their AI targeting systems in an active conflict, with systems like Project Lavender and The Gospel getting optimized in real-world military operations, and all of it ran on Oracle’s servers: every algorithm iteration, every optimization flowing through infrastructure where Oracle engineers could observe.

Oracle technically can’t build a Palantir competitor because they’re in a global partnership with contracts, probable non-compete provisions, and shared customers. But Tony Blair can, because he runs an “independent nonprofit think tank.” If Blair’s Institute builds AI platforms for governments while being funded entirely by Larry Ellison and leveraging everything Oracle learned from hosting Palantir’s systems, that’s not Oracle competing with Palantir; that’s just a think tank doing think tank things.

It’s plausible deniability with a British accent, and it means Peter Thiel spent years perfecting technology, tested it in a live war, worked out the bugs, and built all the training data on Oracle’s infrastructure while Oracle’s man was quietly building the alternative.

Thiel is not stupid, and he is not passive. The man bankrolled a decade-long legal campaign that destroyed Gawker Media over a grudge; he does not take betrayal lightly. I would be surprised if Palantir’s lawyers aren’t already reviewing every clause of that Oracle partnership, looking for exits and violations. But the problem with infrastructure dependency is that you can’t leave quickly even when you see the trap, and unwinding from Oracle’s cloud would take years while competitors built on that same cloud eat your lunch.

The deeper truth is that Ellison doesn’t actually need TBI to beat Palantir at the technology level. Think about it like this: Palantir makes the cars while Oracle builds the highways. In 2024, Ellison convinced Palantir to move all their cars onto his highways, and now he’s building his own cars through Tony Blair. Even if his cars are never as good, he still wins, because everyone, Palantir, TBI, whoever comes next, they’re all driving on Oracle’s roads.

You can swap out cars easily, replace software, migrate to competitors, but you can’t move highways because that infrastructure takes decades to build and costs tens of billions to replace. Oracle doesn’t care who wins the AI surveillance race because whoever wins has to pay Oracle to run their systems. Once you’re on someone else’s infrastructure, you’re not their competitor; you’re their tenant paying rent.

Here’s why this should matter to you even if you’ve never thought about cloud infrastructure in your life: Oracle runs the systems that determine whether you get your government benefits, whether your healthcare records transfer correctly between providers, whether your tax return processes on time, whether the police database that pulls up your name during a traffic stop is accurate. This isn’t abstract technology; it’s the plumbing of modern life, and one company is quietly becoming the landlord of all of it.

The thing about being a landlord is you don’t need to run the best restaurant in the building, you just need to own the building where all the restaurants operate. TBI only needs to be good enough that governments feel like they have options, because as long as they’re choosing between Palantir and TBI, they’re both choosing Oracle.

The scenario that keeps me up at night isn’t even Thiel and Ellison competing; it’s what happens when they realize they don’t need to. TBI handles certain governments, Palantir handles others, Oracle infrastructure runs both, and suddenly you’re not looking at competition anymore, you’re looking at a cartel controlling the surveillance infrastructure of dozens of nations with no meaningful alternative in sight.

That’s where the money is flowing. That’s what the partnership structures enable. That’s what a $348 million investment in a former Prime Minister’s “think tank” actually buys you: a war criminal chairing peace initiatives while selling surveillance tools to the highest bidder, all running on infrastructure controlled by a man who never forgives rejection.

An 82-year-old is still suing companies over slights, still playing 20-year revenge games, still building traps that won’t fully close for another decade. If someone can be that patient, that strategic, that ruthless at 82, we should be paying very close attention to what he builds next, because by the time most people notice, they’ll already be living inside it.

SOURCES CITED:​​

Discussion about this video

User's avatar

Ready for more?

© 2026 Audrey Henson · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture