The 13F landed on November 14. And just like that, Michael Burry — the guy who called the greatest housing short in modern history — was done with Microsoft. Done with Oracle. Both positions, gone. Zero. Nada.
Now, here's the kicker: this isn't 2008. This isn't subprime mortgages. This is the AI trade — the trade that has carried the S&P 500, the NASDAQ, and basically every tech ETF on the planet since ChatGPT went mainstream.
Wall Street's reaction? Crickets. Microsoft shares actually ticked up about 2.5% from late September. Oracle climbed roughly 8%. The market shrugged. But I've been staring at on-chain data and capital flows long enough to know that when a guy with Burry's track record quietly exits two of the biggest AI infrastructure plays on Earth, you don't ignore it. You dissect it.
Because here's the thing nobody's saying: Burry isn't just selling two stocks. He's telling you the AI capex supercycle is running on fumes. And the market's non-reaction? That's exactly when the rug gets pulled.
Let me explain why.
The Setup Nobody's Talking About
Let's rewind a bit. Since 2023, the entire US equity market narrative has been AI, AI, AI. Microsoft invested over $13 billion into OpenAI. Oracle pivoted from legacy database software into a cloud infrastructure giant, competing with AWS and Azure for AI workloads. Together, they represent two distinct bets on the same thesis: that AI is a real, monetizable revolution, not a glorified autocomplete.
I've been tracking this convergence since I audited an AI-driven prediction market protocol back in 2025 — and found a critical vulnerability in its oracle mechanism just before mainnet launch. That experience taught me something: in this market, the narrative always runs ahead of the fundamentals. Red candles don't lie. But the stories people tell themselves? They lie all the time.
For the past two years, the story was: AI capex is a moat. Companies build massive data centers, buy millions of GPUs, and lock in their competitive advantage. Microsoft's capital expenditures went parabolic. Oracle's too. And the stock prices followed, because in a zero-to-low-rate environment, the market rewards spending on the future.
Here's the problem nobody wants to admit: capex is not revenue. And revenue is not profit.
The Core: Burry's Uncomfortable Math
Let's be real about who Michael Burry is. He's not a crypto Twitter personality. He's not a macro influencer with a newsletter. He's the guy who read the subprime mortgage documents, modeled the defaults, and shorted the entire housing market years before the crash. He has a track record of seeing structural fragility before the crowd does.
And now, he's looked at Microsoft and Oracle — two of the most heavily owned tech stocks in the world — and said: I'm out.
Let's unpack the math. Microsoft's stock has a price-to-earnings ratio hovering around 35-40x. Oracle's P/E is even steeper, often north of 40x forward earnings. Those multiples assume not just that AI revenue grows — but that it grows at an accelerating, uninterrupted pace for years. They assume the capex spiral is rational. They assume every dollar spent on Nvidia GPUs translates into durable, profitable services.
Burry's exit is a vote against that assumption.
Consider what a company like Microsoft actually needs to justify its valuation. If you're paying 35 times earnings, you're betting that the company's earnings will roughly triple over the next decade. That's a huge bet on AI-specific revenue streams — Copilot subscriptions, Azure AI services, OpenAI's compounding losses, and whatever else Satya Nadella can monetize. It's a high-conviction, high-multiple bet on one of the most competitive and rapidly commoditizing markets in history.
I've watched this pattern before. In DeFi Summer 2020, I saw protocols with zero revenue hit billion-dollar valuations. I saw yield farmers call it alpha when it was really exit liquidity. The digital casino was open 24/7, and the house always won — because the house knew something the farmers didn't: yield wasn't generated; it was subsidized. It was a Ponzinomics treadmill propped up by new entrants.
The AI trade has uncomfortably similar vibes. Not because AI isn't real — it absolutely is. But because the investment thesis has shifted from "this is valuable" to "this is worth 40x earnings." And that gap is where the carnage happens.
Now, I'm not saying Burry is definitely right. Far from it. His track record includes big winning shorts — but also early, painful losses. The man shorted the market in 2020 and got run over by the recovery. He's been early before. Being early means being wrong for a long time in this game.
But here's the point that matters: Burry isn't a momentum trader. He's a structural-value analyst. When he dumps two AI mega-caps simultaneously, it's not a tactical move. It's a thesis-level statement.
The Contrarian Angle: The 13F Is a Lie (Or at Least a Lobby)
Here's where it gets interesting. Everyone is reading this news as "Burry thinks AI is overvalued." And that's the obvious take. But I want to flip this around — because in my line of work, the obvious read is usually the planted one.
Let's talk about the 13F disclosure itself. You know what a 13F actually is? A quarterly snapshot filed up to 45 days after the quarter ends. That means Burry's "fresh" November 14 disclosure could be transmitting trades he executed in August or September. During that window, Microsoft and Oracle both rallied — which means he may have sold into strength, not weakness. The market didn't react to his exit because, by the time the news hit, the trade was ancient history.
But here's the thing that really matters, and it's the angle nobody's covering: the 13F doesn't tell you the full story. It doesn't show you the puts. It doesn't show you the short positions. It doesn't show you the swaps.
Remember what Burry did in 2008. He didn't just sell banks. He bought credit-default swaps. He didn't short the obvious instrument — that would have been too expensive. He found the cheapest, most leveraged way to express his thesis. The 13F only shows long equities. If Burry had taken a massive short position in Microsoft or Oracle through options or total-return swaps, you wouldn't see it in this filing at all.
Actually, news of his new puts on the broader market broke recently. He bought put options on the S&P 500 and the NASDAQ. So while the financial press is fixated on his tech stock exits, I'm looking at this from a different angle: he's not just exiting positions. He's actively positioning for a crash.
That's the real story. The AI capex cycle is peaking, and the smartest contrarian in the room is building a portfolio that pays off when the mainstream crowd finally catches up.
And here's where we need to bring it home to crypto. This is 2025. The AI narrative has fused with crypto. We've got AI agents launching tokens, decentralized compute networks, "decentralized AI" — I've audited some of these protocols myself, and let me tell you, the technical quality is as variable as a pump-and-dump on Twitter. Some of these projects have genuinely cracked important problems around verifiable inference and decentralized training. But many are just selling tokens on the AI buzzword.
Burry's signal matters here because the AI trade is interconnected across public equities and crypto. If Microsoft and Oracle correct, the AI narrative cools. If the AI narrative cools, then the AI-agent tokens and the GPU DePIN plays and the "inference marketplace" coins are next in line. It's the same exit liquidity game, just with different tickers and more decimal places.
Wash trading is everywhere in crypto. But now, wash trading has evolved into "narrative trading" in public equities. Hedge funds buy the AI story, push the weights up, sell to latecomers, and rotate. Burry sees it. He's calling it. And he's positioning for the post-crash world.
The deeper question here is about AGI itself. Let me tell you what I actually think: AI is a powerful technology, but it's not an instantaneous business. The gap between technological capability and economic monetization is enormous. And in that gap lies the risk. The market has been treating AI like a linear, predictable revenue machine. But the reality is messier.
I've spent my career looking at the gap between narratives and fundamentals. I broke the story on ICOs in 2017 with zero GitHub commits, just because I cross-referenced whitepapers against code. I watched DeFi yield farmers lose everything to impermanent loss in 2020. I traced NFT whale dumps in that 40% floor collapse in early 2022. In every case, the catalyst for the crash wasn't new information. It was exactly the news cycle shifting from hope to doubt. And the first sign of that shift is always the smart money exiting quietly.
Here's the reality: the quarterly earnings doesn't need to be bad for the AI trade to crumble. It just needs to stop being better than expected. The market has priced in perfection in Microsoft and Oracle earnings. If their capex guidance comes in 10% below consensus — if Mark Zuckerberg or Jensen Huang or Nadella starts saying the words "efficiency" or "capital discipline" — the whole trade unwinds.
The Takeaway: What to Watch Next
Look, I'm not telling you to sell everything and buy puts. That's what Burry does — and he's been wrong before, for long stretches. The market is a machine that can stay irrational longer than you can stay solvent; I've lost trades to that truth, and I'll lose again.
But the signal is real, and it's worth respecting.
What I am saying is this: the red candles on Microsoft and Oracle are already flickering. The exit liquidity is beginning to walk. If Burry's conviction begins to spread through other hedge funds — and I know from watching the 13F filings that it's starting to — the AI trade could get ugly, fast.
Watch the next round of earnings. Watch Microsoft's and Oracle's capex guidance like a hawk. Watch whether other whales quietly follow Burry out the door. Watch whether the AI-convergence crypto sector starts bleeding LPs.
Because when the smartest bear in the room positions for a crash, the crash doesn't care whether you believe in AI. It just cares whether you saw the exit door.
Red candles don't lie. The question is whether you'll read the signals before they engulf your portfolio.
