Michael Burry Warns Of AI Bubble As Global Stocks Sell Off

The AI Bubble Warning: Michael Burry’s $1.1 Billion Bet

Michael Burry, the investor who famously predicted the 2008 housing crash, has placed a massive $1.1 billion bet against some of the most hyped AI stocks. His weapon this time: put options on Nvidia and Palantir—contracts that earn a profit if the share price falls. As disclosed in SEC filings, Burry’s bearish wager reflects a conviction that the artificial intelligence frenzy has pushed valuations to unsustainable levels.

Burry is not alone in his caution. He recently warned that the upcoming SpaceX initial public offering, alone raising $75 billion, could be seven times larger than the biggest dot-com-era IPO. When combined with expected offerings from OpenAI and Anthropic, he argues the total haul could exceed the $44 billion raised by 446 tech IPOs before the 2000 crash. “Is AI creating the biggest liquidity drain in market history?” he asked in a public note. That question hits at the core of today’s anxiety.

Meanwhile, Warren Buffett’s Berkshire Hathaway is sitting on nearly $400 billion in cash—a record hoard that suggests the Oracle of Omaha sees few attractive investments at current prices. Yet the optimists have their own rejoinder: Goldman Sachs has raised its S&P 500 target to 8,000, and AI-linked companies are generating real revenue. Anthropic, for instance, notched $30 billion in annual recurring revenue. As we explored when the AI bubble debate first intensified, the bulls counter that genuine revenue changes the calculus. The debate has never been more divisive.

Global Stock Sell-Off: Broadcom, Nvidia, and the Chip Rout

The catalyst for the latest downswing was a disappointing guidance update from Broadcom, whose shares tumbled 15% in a single session—a move we detailed in our latest market summary. The rout then spread across the semiconductor sector, as we explained in our analysis of the chip meltdown, dragging down memory chip stocks like Micron and SanDisk (see our SanDisk coverage). The pain quickly enveloped the broader tech space.

Nvidia, the poster child of the AI boom, saw its $5.5 trillion market capitalization wobble. While the chip maker still dominates the market for AI training hardware, any hint that demand might decelerate sends shockwaves through a market that has become deeply dependent on AI-linked equities. The sell-off was not just about earnings—it was about the realization that concentration risk has reached historic extremes.

Concentration Risk: S&P 500 Weights Approach Dot-Com Levels

According to the International Monetary Fund’s October 2025 Global Financial Stability Report, the information technology sector now makes up 35% of the S&P 500—the same weight it held at the peak of the dot-com bubble. Even more striking: just seven stocks, the so-called Magnificent Seven, account for 33% of the entire index. A concentration measure known as the Herfindahl-Hirschman Index, which tracks how evenly the index is distributed, has climbed more than 20 percentage points higher than during the dot-com era.

When a handful of names drive most of the market’s gains, the index becomes fragile. A stumble by any one giant can pull down everyone else. That’s exactly what happened when Broadcom’s forecast disappointed. The IMF warns that current US equity valuations “appear stretched relative to fundamentals,” a polite way of saying prices have outrun plausible profit growth.

Not everyone agrees. Bulls point to the genuine technological transformation underway. Artificial intelligence is not a fad—it’s reshaping industries. But the question is whether that transformation justifies a $5.5 trillion price tag for Nvidia or a $30 billion recurring revenue figure for a still-unprofitable AI startup like Anthropic. The bulls see a revolution; the bears see a replay of 1999.

AI Adoption: Real Revenue or Hype?

Here the picture gets messy. A Federal Reserve analysis released in April 2026, using multiple surveys, found that only 18% of US firms had adopted AI when measured by simple firm count—yet an employment-weighted measure jumped to 78%. The disparity shows that large companies are going all in, while smaller businesses lag far behind. The bar chart below (Figure 1) visualizes this split, as well as how many workers personally use generative AI tools: 41% of the labor force, with 54% employed at firms that have adopted the technology.

In other words, the AI revolution is real, but it’s uneven. The revenue that companies like Nvidia and Anthropic count on depends heavily on a relatively small set of giant customers—the “hyperscalers” that are projected to spend a combined $725 billion on AI infrastructure in 2026. If those big spenders ever pause or trim budgets, the knock-on effect could be severe.

The accompanying data table lays out the stark figures: Burry’s $1.1 billion in puts, Nvidia’s $5.5 trillion market cap, and the 35% tech weighting all sitting side by side with an adoption rate that is still, in many corners, nascent. The disconnect between sky-high valuations and the ground-level reality of uneven AI integration is the core tension that Burry’s bet exploits.

Conclusion

Michael Burry’s $1.1 billion wager is not a prediction that AI will fail. It’s a bet that the market has priced in too much success, too quickly. The warning signs are written in plain data: concentration risk that exceeds the dot-com era, valuations that the IMF calls stretched, and an adoption pattern that, while accelerating, still favors a few giant firms over the broad economy.

The stock sell-off sparked by Broadcom is a reminder that sentiment can turn on a dime. When one giant coughs, the whole index catches a cold. Yet the counter-argument—real revenue, surging infrastructure spending, and a technology that seems poised to reshape the world—is not easily dismissed. Both sides have data that supports their view.

For investors and observers alike, the prudent course is to watch the numbers, not the narrative. The data table and chart above provide a framework for cutting through the noise. Whether this is a bubble or a boom, history suggests that extremes never last. The only question is how long they’ll run before reality reasserts itself.