원문 보기
When the hedge fund manager who famously predicted the 2008 subprime collapse turns his sights on artificial intelligence, Wall Street pays attention. Yet, over the past 3 years, Michael Burry's crusade against the AI trade has yielded a string of costly financial missteps. In 2023, he allocated 1.6 billion dollars in notional put options against the S&P 500 and Nasdaq, accepting a brutal 40% realized loss as tech momentum surged.
Undeterred, he bought put options on the SOXX semiconductor ETF, watching helplessly as the index ripped 16% higher and vaporized his options premium. By late 2025, Burry escalated his bear campaign with over 1.1 billion dollars in short puts targeting Nvidia and Palantir. But as Nvidia smashed past a 5.5 trillion-dollar market capitalization, Burry found his bearish options buried deep underwater, serving as a stark reminder that even legendary contrarians can get crushed standing in front of a multi-trillion-dollar technological shift.
Rather than capitulating, Burry doubled down in his latest disclosures, expanding short stock and derivative positions to consume over 21% of his entire portfolio. His short book takes direct aim at five major pillars of the AI ecosystem: semiconductor giant Nvidia, enterprise assets leader Palantir, database heavyweight Oracle, neo cloud provider Nebius Group, and industrial manufacturer Caterpillar.
However, a deeper examination reveals an unprecedented mechanical defense. To protect his fund from catastrophic upside risk, Burry deployed roughly 4% of his capital to buy December Nvidia call options at strike prices in the mid to high $200. In a rare public disclosure, Burry admitted these call options were not a bullish trade, but essential upside insurance offset by his massive short exposure.
Acknowledging that post-earnings price reactions represent a coin flip, Burry finds himself trapped in a peculiar paradox, admitting Nvidia appears undervalued on paper at a forward PE of 23.5x, while insisting corporate hardware spending will inevitably collapse. At the center of Burry's thesis is Nvidia, where he accuses the company of relying on deceptive circular financing to engineer artificial GPU demand.
Burry argues that backstopping customer debt and providing residual value guarantees across $500 billion in infrastructure capital is reminiscent of vendor financing bubbles destined for liquidation. Yet, fundamental data exposes a flawed premise. Nvidia isn't manufacturing phantom sales. It operates like a sovereign AI nation-state, strategically subsidizing its computing ecosystem to lock in decades of CUDA platform dominance.
By partnering with BlackRock, Apollo, and Goldman Sachs to provide a 25% residual value safety net, Nvidia leverages asymmetric industry insights to secure long-term demand. The balance sheet tells a story of unmatched economic power, boasting a 62% net profit margin and a compressed PE ratio lower than the Nasdaq average, even after delivering 96.2 billion dollars in quarterly revenue and projecting 70% growth through fiscal 2028.
With enterprise cloud revenue skyrocketing 138% to over 40 billion dollars and supply commitments topping 279 billion dollars, Burry is shorting a monopoly that remains severely supply constrained rather than overbuilt. The second leg of Burry's trade focuses on Palantir Technologies, where he acquired March 2027 put options with strike prices in the low to mid 100-dollar range, anticipating a collapse from its peak above 200 dollars.
Burry appears to be banking on headline valuation multiples to trigger a sell-off, but betting against Palantir's operating leverage is proving to be a serious, mathematically dangerous venture. Over three quarters, Palantir's revenue surged from 1.18 billion dollars to 1.94 billion dollars, fueled by triple-digit expansion across its US commercial division.
Even more remarkable is the company's profitability engine, which saw GAAP net income margins explode from single digits to an astounding 55% driving GAAP operating margins to 47% and generating an elite rule of 40 score of 155%. As earnings scale faster than share price, Palantir's forward valuation multiples are compressing while the stock trades near record highs.
By continually forcing Wall Street analysts to raise revenue targets after every quarterly report, Palantir operates less like an overhyped software bubble and more like an untouchable enterprise monopoly converting corporate AI adoption directly into bottom line profits. Burry's third short target takes aim at Oracle, where he contends that the software titan's aggressive data center expansion will crumble under an unserviceable debt load.
Burry views Oracle's $638 billion as an accounting illusion, warning that a $70 billion capital expenditure plan for fiscal 2027 will wreck the balance sheet once AI demand cools. However, operational evidence heavily contradicts his thesis. Oracle concluded its fiscal year with $67.4 billion in revenue and $32 billion in operating cash flow.
While cloud infrastructure accelerated 93% to $5.8 billion, led by a 119% surge in GPU compute capacity. As Wall Street turns its focus directly to Oracle's September 10th earnings release, the company enters the quarter with 98% of available AI capacity contracted and data center utilization running at 97.5% Crucially, balance sheet exposure is heavily mitigated by customer prepayments and bring your own hardware structures.
Trading at a discounted forward PE of 18x relative to expected 34% top line acceleration toward $90 in revenue, Oracle represents a hyper-growth infrastructure backbone rather than a vulnerable debt trap. In targeting Nevius Group, Burry is wagering against the financial viability of emerging neo cloud providers, interpreting multi-billion dollar convertible debt offerings as evidence of a speculative dilution trap.
Yet actual performance metrics reveal an explosive infrastructure engine operating with elite efficiency. In its latest disclosure, Nebius delivered a 454% year-over-year revenue surge to $582.3 million, pushing its annualized run rate to $3.0 billion while generating $236.2 million in adjusted EBITDA at a 49.7% margin.
Removing execution risk, Nvidia holds a direct 9.3% equity stake in Nebius following a $2 billion strategic investment guaranteeing access to flagship GPU hardware. With GPU rental spot prices holding firm at 52% of on-demand rates and data center builds achieving over $20 million in annual contract value per megawatt, Nebius is realizing a payback period under 22 months and returns on invested capital above 30%.
Combined with high-value non-core holdings, including a 28% stake in ClickHouse valued at $15 billion and an autonomous vehicle division valued at $6 billion, Burry's short bet overlooks a hyper-growth asset capable of generating massive multi-year equity returns. Finally, Burry's short position in Caterpillar represents a macro bet against the physical backbone of the digital economy.
Burry views Caterpillar as a cyclical industrial trading at an inflated 33.5 times forward PE multiple that will plummet the moment technology spending slows. However, this perspective completely misses Caterpillar's transformation into an essential AI infrastructure vendor. In its latest financial report, Caterpillar posted a 24% revenue increase alongside a 73% jump in diluted earnings per share to $8.17, while operating profits surged 50%.
The primary catalyst behind this acceleration is Caterpillar's power generation segment, which surged 72% year-over-year as technology conglomerates scramble to procure heavy industrial generators to keep data centers operational. Supported by a $72 billion order backlog and a 13% net profit margin that doubles industry peers, Caterpillar is utilizing cash flow to aggressively retire $1.5 billion in stock each quarter.
Rather than a cyclical trap, Caterpillar's dominant grip on global power equipment confirms that the physical infrastructure build-out supporting artificial intelligence is only in its opening innings. When analyzing Burry's portfolio as a unified thesis, a striking reality emerges. By placing over 21% of his funds capital into short bets spanning semiconductor manufacturing, enterprise software, cloud infrastructure, neo cloud hosting, and physical power generation, Burry is betting against every single layer of the global AI transformation.
Yet, each of his five target companies continues to report record-breaking quarterly profits, massive backlog visibility, and expanding operating margins while trading at justifiable valuation metrics. Burry's decision to purchase December Nvidia call options demonstrates that even he recognizes the catastrophic risk of getting short squeezed by unstoppable earnings While Michael Burry built his reputation identifying systemic fragility in 2008, betting against hyper growth monopolies backed by hundreds of billions in contracted cash flow may prove to be the most expensive miscalculation of his investment career.
Comments 0