In the high-stakes world of artificial intelligence, where trillions of dollars are pouring into data centers and chip technologies, one investor is sounding a dire alarm. Michael Burry, the hedge fund manager immortalized in ‘The Big Short’ for predicting the 2008 housing crash, has placed a staggering $1 billion bet against the AI boom. His target: what he sees as inflated profits at companies like Meta Platforms Inc. and Oracle Corp., driven by questionable accounting practices around asset depreciation.

Burry’s thesis, detailed in recent disclosures and public statements, centers on how these tech giants are extending the estimated useful life of their AI-related hardware—particularly servers and semiconductors—from the industry standard of three to four years to as long as eight years. This maneuver, he argues, artificially boosts reported earnings by spreading out depreciation costs over longer periods, masking the true economic reality of rapid obsolescence in AI tech.

According to a report by Fortune, Burry shut down his hedge fund, Scion Asset Management, shortly after revealing this massive short position. He believes this accounting sleight-of-hand could overstate industry profits by up to $176 billion between 2026 and 2028, setting the stage for a market correction that echoes past bubbles.

The Accounting Maneuver Under Scrutiny

At the heart of Burry’s critique is depreciation accounting for fixed assets like AI servers. Traditionally, tech hardware depreciates quickly due to rapid advancements; a chip that’s cutting-edge today could be outdated in mere years. Yet, companies like Meta and Oracle have shifted their policies. Oracle, for instance, extended the useful life of its servers to eight years in fiscal 2025, as noted in a Bloomberg analysis.

This change isn’t isolated. Microsoft Corp. has similarly adjusted its depreciation schedules, contributing to what Burry calls ‘one of the most common frauds in the modern era.’ In a post on X (formerly Twitter), Burry elaborated that by understating depreciation, these firms are ‘hiding the brutal truth’ about the unsustainable costs of AI infrastructure.

A New York Times article highlighted investor worries about Oracle’s cloud computing financials, where such accounting shifts have helped propel stock prices amid the AI rally. But skeptics, including Burry, point out that if hardware lifespans are overestimated, future write-downs could be massive when reality catches up.

Burry’s Track Record and Current Wager

Burry’s reputation was built on foresight. In 2008, he bet against subprime mortgages, netting billions for his investors. Now, he’s applying that contrarian lens to AI. As reported by Benzinga, Burry doubled down on his claims even as his short trade initially backfired amid a stock rally, insisting that Meta and Oracle are overstating earnings through understated depreciation.

The investor’s position involves shorting stocks tied to AI hyperscalers. Posts on X from users like @HedgieMarkets reflect growing sentiment: ‘The AI bubble evidence is now overwhelming,’ with references to OpenAI’s massive cash burn and unprofitable status despite sky-high valuations.

Bank of America has echoed these concerns, warning in a recent analysis cited by Market Realist that the AI boom may face a cash crunch as tech giants pivot to bonds and debt to fund ambitions. Meta, for example, issued $30 billion in financing for a Louisiana data center, per the report, while carrying $37 billion in debt against $60 billion in cash reserves.

Industry-Wide Implications of Extended Lifespans

Extending asset lifespans isn’t new, but in AI’s context, it’s amplified. A Morningstar piece revealed that Microsoft’s finance-lease liabilities surged 70% to $46.2 billion in fiscal 2025, while Oracle recorded $4 billion in similar liabilities—up from zero a year prior. This ties into Burry’s point: higher reported profits today, potential pain tomorrow.

Oracle’s recent earnings, as covered by TradingView News, missed estimates but boasted an astounding outlook, driving shares higher. Yet, Burry argues this optimism masks depreciation realities. X user @GaryMarcus tweeted that Oracle’s $300 billion market cap surge on a non-binding OpenAI deal signals ‘peak bubble,’ given OpenAI’s lack of funds.

Meta’s involvement is equally telling. The company has ramped up AI spending, with CEO Mark Zuckerberg touting investments in infrastructure. But as Axios reported, Big Tech’s turn to private debt and special purpose vehicles is a warning sign of bubble-like behavior.

Voices from the Bull and Bear Camps

Not everyone agrees with Burry. Optimists point to real revenue growth. An X post from @Umbisam hailed Oracle’s ‘monstrous revs growth guidance’ through 2029 as proof AI is ‘just getting started,’ dismissing bubble talk as ‘utterly stupid.’

Conversely, bearish voices proliferate. @InfraStory on X predicted an ‘AI Mega Bubble’s gonna burst so big that will create a Massive Financial Tsunami,’ comparing market caps like Nvidia’s $4.5 trillion to modest revenues.

A Yahoo Finance article quoted Burry warning that the accounting hides ‘the brutal truth,’ with potential for a crash as he braces by closing his fund.

Risks of a Broader Market Reckoning

If Burry is right, the fallout could ripple beyond Meta and Oracle. The AI sector’s total capex has ballooned, with Bloomberg estimating investors have poured unprecedented sums without clear payoffs, as in their October 2025 piece.

X sentiment, including from @ShayBoloor, suggests that while pullbacks may occur, AI’s infrastructure builds ‘hard infra that throws off cash flow,’ differentiating it from the dot-com bust.

Yet, Oracle’s mixed Q2 2025 results, per ERP Today, included a new AI deal with Meta but followed a failed pact with Elon Musk’s xAI, underscoring partnership risks.

Lessons from Past Bubbles

Historical parallels abound. The dot-com era saw similar overvaluations based on hype. Burry’s bet recalls his housing short, where accounting gimmicks masked risks.

A recent WealthAdvisor article questioned if ‘talk of an AI bubble’ has become its own bubble, noting Burry’s focus on underestimated depreciation costs totaling $176 billion.

As markets digest these warnings, investors must weigh Burry’s prescience against AI’s transformative potential. The debate rages on X, with @ETNOWlive summarizing: ‘Hyperscalers like Oracle and Meta are overstating earnings by extending chip lifespans.’

Navigating the Uncertainty Ahead

Tech executives defend their practices. Oracle’s leadership, in earnings calls, attributes extended lifespans to durable cloud infrastructure. Meta emphasizes long-term AI bets yielding returns.

Analysts like those at Bank of America, via BizToc, highlight debt-fueled spending as a red flag, with Oracle borrowing over $100 billion amid ties to unprofitable OpenAI.

X user @TomAlphaTrades called this ‘late-cycle behavior,’ warning of stranded assets if partners like OpenAI falter. As 2025 unfolds, Burry’s bet may prove prophetic—or another contrarian miss in a resilient market.