May 2026. Michael Burry just added to his short positions against Nvidia and Oracle. The market yawned. It shouldn't.
I've spent 23 years watching markets misread signals. This one is different. Burry isn't a perma-bear. He's a forensic analyst who reads balance sheets like autopsies. When he moves, it's because the numbers scream. And right now, the numbers on AI are screaming something the crowd refuses to hear.
Let's cut through the noise. The AI trade is the largest liquidity sink in financial history. It's not just stocks. It's a macro bet on interest rates, fiscal policy, and the durability of a supply chain that's more fragile than a DeFi protocol after a governance attack. Burry's short is a canary in a coal mine—and crypto traders should be watching it like a hawk.
The Hook: A Short That Speaks Volumes
On May 12, 2026, Burry's latest 13F filing revealed increased put positions on Nvidia (NVDA) and Oracle (ORCL). The filing didn't disclose exact sizes—typical for Burry—but the signal is unambiguous. He's betting against the two most prominent AI infrastructure plays. The market's reaction? A shrug. NVDA barely moved. ORCL dipped 0.3%. The complacency is the story.
Burry doesn't short for fun. He shorted subprime mortgages in 2008 when everyone called him crazy. He shorted Tesla in 2020 and lost—but he also shorted the housing bubble and won big. His track record is mixed, but his methodology is consistent: he finds where the market's assumptions diverge from hard data. Right now, the hard data on AI is flashing red.
Context: The Macro Backdrop That Burry Is Betting On
To understand Burry's move, you need to see the macro canvas. The Federal Reserve hiked rates aggressively from 2022 to 2024, pushing the federal funds rate to 5.25%-5.50%. Then they started cutting in September 2024, but only by 100 basis points total, landing at 4.25%-4.50% by early 2025. The market initially priced in four to five cuts for 2025. Reality? Inflation is sticky. Core CPI is stuck at 3.0%-3.5%. The Fed has revised expectations down to one or two cuts. This is the "higher for longer" regime that Burry thrives on.
Here's the math that matters: AI companies are long-duration assets. Their valuations are based on discounted future cash flows. A 25-basis-point change in rates can swing Nvidia's valuation by hundreds of billions. When rates stay high, the discount rate crushes those future earnings. Burry is essentially shorting the duration of the AI trade. He's betting that the Fed won't cut aggressively because inflation won't cooperate.
But it's not just rates. It's the fiscal picture. The U.S. federal debt has blown past $35 trillion. The 2024 deficit was 6.4% of GDP. Interest payments now exceed defense spending. The Treasury is flooding the market with bonds, absorbing liquidity that would otherwise flow into risk assets. This is the "fiscal dominance" scenario—where fiscal needs keep long-term rates elevated, regardless of what the Fed does. Burry sees this. He knows that high rates and high debt are a toxic mix for speculative growth stocks.
Core: The Three Pillars of the AI Bubble
Let's break down why Burry's short is more than a contrarian bet. It's a data-driven indictment of three structural flaws in the AI trade.
1. The Valuation Disconnect
Nvidia's revenue grew 122% in fiscal 2025. Impressive. But the stock trades at a forward P/E of over 40. Oracle's P/E is even higher. The market is pricing in decades of hypergrowth. But here's the problem: the growth is decelerating. Nvidia's data center revenue growth has slowed from triple digits to double digits. The law of large numbers is kicking in. You can't sustain 100% growth when your revenue base is $100 billion.
I've seen this pattern before. In 2020, DeFi protocols were posting insane APYs. Everyone thought it was the future. But when you model the token emissions, you realize the yields were subsidized by new entrants. The moment inflows slowed, the whole house of cards collapsed. AI is no different. The capex cycle is being driven by a handful of hyperscalers—Microsoft, Google, Meta, Amazon. They're spending billions on data centers and chips. But where's the revenue? AI services are still largely unprofitable. The downstream applications—chatbots, image generators, code assistants—are burning cash. The profit is concentrated upstream in chipmakers like Nvidia. That's a classic "price scissors" problem: upstream margins are fat, downstream margins are razor-thin. This imbalance is unsustainable.
2. The Supply Glut
Remember the 2000 fiber-optic bubble? Companies laid enough cable to circle the earth multiple times. Then demand didn't materialize. The same thing is happening with AI chips. In 2023, H100 GPUs had a 36-52 week lead time. By 2025, that dropped to under 20 weeks. By 2026, it's likely under 8 weeks. Supply is flooding the market. TSMC is ramping up advanced packaging. Samsung and Intel are entering the fray. The shortage is over. When supply catches up to demand, prices fall. Nvidia's gross margins—currently around 70%—will compress. That's a direct hit to earnings.
I've audited enough tokenomics to know this pattern. When a project's token emission schedule accelerates, the price dumps. The same logic applies to physical supply. The AI chip market is about to face a supply shock. Burry sees it. He's shorting the companies that are most exposed to this glut.
3. The Profit Distribution Trap
Look at the AI value chain. Upstream: chip designers (Nvidia, AMD), foundries (TSMC), equipment makers (ASML). Downstream: model developers (OpenAI, Anthropic), application providers (Salesforce, Adobe). The upstream players are minting money. The downstream players are bleeding. OpenAI reportedly loses billions annually. The only reason they survive is because of massive venture capital infusions. This is not a healthy ecosystem. It's a pyramid scheme where the top layers extract value from the bottom.
In crypto, we saw this with the ICO boom of 2017. The infrastructure providers—exchanges, miners—made fortunes. The actual projects? Most died. The same dynamic is playing out in AI. The "picks and shovels" narrative is overhyped. When the bubble bursts, the downstream will collapse first, and then the upstream will follow as demand evaporates.
Contrarian: The Unreported Angle—AI Is the New DeFi Summer
Here's what the mainstream financial press isn't telling you. The AI trade is structurally identical to the DeFi yield farming mania of 2020. In both cases, you have a narrative-driven boom, massive capital inflows, and a fundamental disconnect between hype and revenue. In DeFi, the yields were subsidized by token emissions. In AI, the growth is subsidized by corporate capex. Both are forms of financial engineering that create the illusion of sustainable growth.
I remember the summer of 2020. I was modeling Curve Finance's emission rates. I saw that the APYs were unsustainable—they were paying out more in tokens than they were generating in fees. I published a warning three weeks before the crash. My subscribers got out. The same math applies to AI. The hyperscalers are spending billions on AI infrastructure, but the return on that investment is unclear. If the ROI doesn't materialize, they'll cut capex. That's the trigger for the bubble to burst.
There's another angle: the fiscal-monetary policy mix. The U.S. is running a massive fiscal deficit while the Fed is tightening. This is the "unholy trinity" of high debt, high rates, and high inflation. It's unsustainable. Something has to give. Burry is betting that the AI trade is the first domino to fall. He's not just shorting Nvidia; he's shorting the entire macro regime that has propped up risk assets for the past decade.
And here's the kicker: the AI trade is a proxy for all risk assets, including crypto. When AI stocks correct, the contagion will spread. Crypto has already shown correlation with tech stocks. A 20% drawdown in Nvidia will trigger a 30% drawdown in Bitcoin. The liquidity that's been chasing AI will flee to safety. The same money that poured into AI tokens—like FET, RNDR, AGIX—will exit just as fast.
Takeaway: What to Watch Next
Burry's short is a signal, not a prophecy. He could be early. He could be wrong. But the data points are clear. Here's what I'm watching:
- Nvidia's quarterly earnings: If data center revenue growth drops below 50%, the bubble is cracking.
- The Fed's dot plot: If they signal zero cuts for 2026, long-duration assets are toast.
- TSMC's monthly revenue: A month-over-month decline would be the first sign of demand destruction.
- Burry's next 13F: If he increases his puts, the conviction is real.
For crypto traders, the play is simple: don't be the last one holding the bag. The AI trade is the canary. When it dies, the whole risk-on complex follows. Static is a choice. Data over destiny. Speed is the only moat.
I've been through 2017 ICOs, 2020 DeFi, 2021 NFTs, and 2022's Terra collapse. Every time, the pattern was the same: euphoria, overinvestment, and a brutal reckoning. AI is no different. The only question is timing. Burry's short tells me the clock is ticking. Are you listening?