Michael Burry: From 2008's $100M Triumph to 2026's Flawed AI Bubble Thesis

Michael Burry cemented his reputation as a legendary contrarian investor by generating extraordinary returns in 2008. His hedge fund, Scion Capital, delivered approximately $100 million in personal profits and $700 million for his investors—an achievement etched into popular culture through the Hollywood film “The Big Short,” where Christian Bale portrayed the prescient analyst. Today, over a decade later, Burry commands attention whenever he shares market commentary. However, his recent bearish calls on artificial intelligence warrant closer examination, particularly when scrutinized against hard data and industry fundamentals.

How Michael Burry Made His Fortune in 2008—And Why It Matters Today

The legendary 2008 bet that made Burry wealthy rested on identifying the subprime mortgage crisis before the market recognized the danger. By positioning his fund with credit default swaps, he capitalized on the subsequent stock market collapse. While most investors will never achieve such returns, Burry’s 2008 performance established him as a market contrarian worth listening to. Yet since that windfall, his track record tells a more complicated story. He has issued numerous bearish forecasts as U.S. equity markets surged higher, consistently arriving at pessimistic conclusions that the market subsequently invalidated. By late 2025, Burry shut down his hedge fund, citing fundamental misalignment with ongoing market dynamics—a telling admission about the difficulties of operating in a prolonged bull market.

The Consistency Problem: Why Burry’s Recent Market Calls Have Missed the Mark

Burry’s performance since 2008 illustrates an important reality: past success in identifying one market inflection point does not guarantee accuracy in predicting subsequent cycles. His recent string of incorrect, gloomy forecasts—made as markets reached new heights—suggests that his contrarian instinct, while potent in 2008, may no longer be reliably calibrated to current market conditions. The market’s persistent advance despite his warnings indicates either a fundamental shift in how markets operate or a limitation in applying a single winning thesis across different economic regimes.

Burry’s AI Bubble Thesis: Three Arguments, Multiple Flaws

Michael Burry’s current stance mirrors his cautionary approach to the broader market: he contends that artificial intelligence stocks are experiencing a 1999-style mania, and predicts that AI’s trajectory will mirror the catastrophic dot-com bust. His argument rests on three distinct pillars, each of which warrants scrutiny when measured against contemporary evidence.

Claim 1: Tech Giants Are Gaming Depreciation Schedules—But The Reality Is Different

Burry asserts that companies like Meta Platforms, Microsoft, and Alphabet are manipulating their accounting by employing overly aggressive depreciation schedules to artificially enhance reported earnings. He notes, for instance, that Alphabet depreciates its servers over a period of only four to six years.

However, this criticism misses critical nuances about AI infrastructure economics. While GPUs (the processors driving AI systems) do depreciate faster than traditional server hardware, most AI infrastructure actually maintains utility across 15 to 20 years of operational life. This extended productive lifespan offsets any accelerated depreciation of individual components. Additionally, older GPU models do not become valueless when newer chips arrive. Dated processors can continue operating for inference—running pre-trained models for end users rather than training new systems—generating ongoing revenue streams. The economic reality diverges substantially from Burry’s accounting critique.

Claim 2: AI Investments Will Drain Cash Flow—Instead, They’re Fueling Growth

Burry warns that unprecedented capital expenditures on AI infrastructure will strain corporate cash flows to unsustainable levels. This argument projects that companies will spend capital recklessly without securing corresponding returns.

The empirical evidence points to the opposite conclusion. Hyperscalers like Alphabet are not experiencing cash flow compression; they are witnessing dramatic cash flow acceleration. Alphabet’s operating cash flow (trailing twelve months) expanded from under $100 billion to $164 billion in 2026—a remarkable surge directly attributable to AI monetization. Operating margins across major tech firms are expanding sharply rather than contracting. Industry data shows that companies implementing AI at scale are reporting returns exceeding $3 for every $1 invested in infrastructure. Furthermore, the latest frontier—agentic AI systems that perform complex tasks autonomously—is delivering documented cost reductions of 25% or greater for early adopters. These figures suggest AI investments are yielding positive cash returns, not depleting reserves.

Claim 3: NVIDIA is the New Cisco—A Comparison That Doesn’t Hold Up

Burry frequently compares NVIDIA, the dominant AI infrastructure provider, to Cisco Systems, the networking equipment darling that peaked in March 2000 and required more than two decades to recover—using valuation multiples as his primary comparative metric.

This analogy fails under closer examination. When Cisco reached its 2000 peak, the company commanded a price-to-earnings multiple exceeding 200—an extreme valuation disconnected from fundamentals. By contrast, NVIDIA’s current P/E multiple stands at 47—a premium valuation reflecting genuine earnings power, but hardly the speculative extreme that characterized late-1990s networking euphoria. The valuation comparison simply does not support the structural equivalence Burry is attempting to establish. NVIDIA also benefits from secular tailwinds (AI adoption) that were not present in Cisco’s market, and the company generates substantially stronger cash flows and profitability metrics than Cisco did during its bubble phase.

H100 Prices and Options Bets Signal Strong Conviction in AI Infrastructure Demand

Market behavior provides additional context for evaluating Burry’s bearish stance. The NVIDIA H100—a powerful data center processor essential for accelerating AI workloads and training large language models—has experienced dramatic price appreciation. Since mid-December 2025, H100 rental prices surged approximately 17%, reflecting persistent GPU scarcity and sustained demand. This price action contradicts any narrative of declining AI infrastructure demand.

The spike aligns with explosive growth in agentic AI adoption, which carries bullish implications for GPU-dependent companies like Nebius Group, CoreWeave, and IREN. Demand for AI processors naturally extends benefits to infrastructure enablers such as Bloom Energy, whose power solutions address the energy bottleneck confronting hyperscalers—a constraint emerging as AI compute at scale encounters physical power limitations.

Options market participants have reinforced this bullish positioning through substantial derivative bets. Deep-pocketed institutional traders placed aggressive call option wagers on both Bloom Energy and NVIDIA in the run-up to earnings. On a recent trading day, Bloom saw 400 call option contracts purchased at maximum strike prices, representing a one-million-dollar directional bet. Meanwhile, a prominent options whale placed an approximately $9 million wager on March $205 calls on NVIDIA. These large institutional positions suggest confidence among sophisticated market participants that AI infrastructure demand will persist and expand.

Notably, Bloom Energy shares broke above technical resistance, exhibiting a constructive weekly chart pattern (bull flag formation) that typically precedes further upside movement. The options activity combined with improving technical positioning indicates market participants are positioning for sustained strength in AI infrastructure—a positioning that directly contradicts Burry’s thesis of an imminent AI crash.

The Bottom Line: Michael Burry’s Legacy Doesn’t Guarantee Accuracy in AI

Michael Burry’s 2008 track record remains extraordinary, and his contrarian credentials deserve respect. However, his current bearish narrative regarding artificial intelligence confronts substantial headwinds in the form of operational data, financial metrics, and real-time market signals. From surging H100 prices indicating robust GPU demand to exponential cash flow growth at major technology firms, from documented positive returns on AI capital deployment to improving cost structures via agentic AI adoption, the empirical landscape contradicts his AI bubble framework.

Being a legendary contrarian is a powerful historical credential, but it does not exempt Burry from the same analytical rigor applied to any market prognosticator. On the evidence, his current AI thesis appears to rest on selective interpretations of accounting practices and infrastructure costs while overlooking the substantial positive feedback loops now embedded in AI monetization. Whether this represents merely another premature bearish call or the beginning of vindication remains to be seen—but the burden of proof now rests with the skeptics, not the builders.