Michael Burry’s Bet Against AI Justified? Logic Explained

Artificial Intelligence has gone from a futuristic concept to corporate obsession. Every major company, from Nvidia and Microsoft to Meta and Oracle, is betting big on AI, pouring billions into data centres, GPUs, and cloud infrastructure.

The outcome? Soaring valuations. In just two years, AI-linked companies have added trillions of dollars in market capitalization. Nvidia alone recently crossed $4 trillion, making it one of the most valuable firms in the world.

But one man thinks this story sounds all too familiar. Michael Burry, the same investor who famously predicted the 2008 financial crisis, has just placed an enormous bet that this AI frenzy is unsustainable. And his reasoning digs deeper than hype, it's about accounting.

The Big Short on AI

Michael Burry’s fund, Scion Asset Management, has reportedly allocated nearly 80% of its portfolio to bets against two AI poster children, Nvidia and Palantir. He has done this through put options, contracts that gain value when the stock price falls. In simple terms, Burry is positioning for an AI correction. His logic? The profits driving AI optimism might not be as solid as they look.

It is worth mentioning that Michael Burry has reportedly de-registered Scion Asset Management from the U.S. Securities and Exchange Commission, as shown in an image he shared on social media. He also hinted at a “big release” on November 25, though details remain unclear. The move suggests Burry may be transitioning Scion into a family office, which would no longer require SEC registration.

The Depreciation Trick

AI infrastructure isn’t cheap. A single Nvidia GPU used in data centres can cost between US $25,000 and US $40,000. These machines are cutting-edge but have short lifespans, typically three years before new models make them obsolete. Yet, Burry claims that several hyperscalers including Meta, Oracle, and others, are changing the rules. They’ve extended the “useful life” of this equipment on their balance sheets from 3 years to 5 or 6 years. That simple tweak dramatically boosts reported profits. 

Source: Company SEC Filings

Here’s an analogy to understand this better:

If you buy a ₹1 lakh machine and plan to use it for 3 years, you record ₹33,000 as an expense each year. Stretch that to 6 years, and your annual expense drops to ₹16,600. Profit looks higher, but nothing about the machine’s real life changed.

According to Burry’s analysis, this accounting shift could inflate total reported earnings by about US $176 billion between 2026 and 2028. If corrected, Oracle’s profits could fall by 27%, and Meta’s by 21%.

In his view, markets are valuing these companies on numbers that might not hold up, the classic setup for a bubble.

A Familiar Pattern: Lessons from Past Bubbles

Michael Burry’s warning echoes several past episodes where technological excitement outpaced real profits.

1. The Dot-Com Bubble (1999–2001)

Back then, the internet was the “next big thing.” Startups with no clear revenue model saw their valuations skyrocket. When profitability failed to materialise, the NASDAQ fell nearly 75% from its peak, wiping out trillions in market value. The parallel today? AI is similarly being priced as a world-changing platform, which it very well may be, but the speed and scale of valuation growth feel reminiscent of that era.

2. The Electric Vehicle (EV) Mania (2020–2021)

More recently, the EV sector witnessed euphoric investor enthusiasm. Dozens of startups went public at multi-billion-dollar valuations before proving unit economics or demand sustainability. As growth slowed and capital costs rose, most of those stocks corrected sharply. The takeaway: technological revolutions often start with legitimate promise but investors tend to overpay early on. 

Burry’s bet, therefore, isn’t necessarily against AI as a technology. It’s against the market’s current expectations that everything AI touches will turn to gold.

The Numbers Behind the AI Bubble Narrative

The data tells a mixed story.

So far, AI investment is generating infrastructure growth, not widespread profitability. That gap between spending and monetisation is precisely where Burry sees risk.

Why Investors Should Care

For everyday investors, Burry’s argument raises three key concerns:

1. Profit Overstatement: If depreciation is stretched, future profits will fall when equipment needs replacement sooner than expected.
2. Valuation Risk: Current price-to-earnings (PE) ratios assume sustained, rapid profit growth. If earnings drop, valuations could correct sharply.
3. Concentration Risk: A handful of AI-linked stocks now drive over 30% of the S&P 500’s total market value. A stumble in one could shake the broader market.

Counterpoint: Why AI Boom Could Still Be Different

Supporters of the AI boom argue that unlike past bubbles, this one is backed by real adoption. AI is already improving search, automation, design, and cloud computing.

Big Tech’s heavy spending isn’t speculative, it’s about securing long-term dominance in a trillion-dollar opportunity. Demand for GPUs remains strong, and early productivity gains are visible across industries.

In short, there’s genuine substance beneath the speculation, just perhaps too much optimism about how soon the profits will show up.

What Smart Investors Should Do

You don’t have to “short” AI to take Burry’s caution seriously. Instead:

• Look beyond hype: Read financial statements — are companies increasing depreciation schedules or capitalising expenses aggressively?
• Focus on real monetisation: Are AI products adding to revenue, or are they still in the “demo” stage?
• Diversify exposure: Avoid portfolios overly concentrated in a few tech names.
• Track refresh cycles: If hardware upgrades accelerate, costs will return faster than expected.

One must know that every major technological revolution has two phases, innovation and irrational exuberance. The internet changed the world, but dot-com investors who paid any price for “future potential” got burned first.

Michael Burry’s bet is a reminder that even genuine breakthroughs can host bubbles. The question isn’t whether AI will transform business, it clearly will. The question is whether investors are paying too much, too soon, for that future.

Disclaimer:

The content is meant for education and general information purposes only. Investments in the securities market are subject to market risks, read all the related documents carefully before investing. Past performance is not indicative of future returns. The securities quoted are exemplary and are not a recommendation. This in no way is to be construed as financial advice or a recommendation to invest in any specific stock or financial instrument. Readers are encouraged to verify the exact numbers and financial data from official sources such as company filings, earnings reports, and financial news platforms and to conduct their own research, and consult with a registered financial advisor before making any investment decisions. All disputes in relation to the content would not have access to an exchange investor redressal forum or arbitration mechanism. INDmoney Global (IFSC) Private Limited, Registered office address: Office No. 507, 5th Floor, Pragya II, Block 15-C1, Zone-1, Road No. 11, Processing Area, GIFT SEZ, GIFT City, Gandhinagar – 382355.