Burry Takes Aim at AI Darlings Oracle and Nebius
Michael Burry, the investor famous for predicting the 2008 housing collapse, has publicly declared short positions in two high-profile AI beneficiaries: Oracle and Nebius. In a Substack post on Thursday, Burry said he shorted Oracle at about $145 per share and Nebius at around $212, explaining that his research ‘just wouldn’t let me continue without doing so.’ The disclosures mark his most pointed bearish commentary on the AI infrastructure buildout to date.
Burry, who now writes about his personal investments rather than managing outside capital, painted a bleak picture of AI companies. He likened them to ‘fish in a barrel’ that have ‘gorged themselves on off-balance sheet liabilities. Backstops. Uncommenced leases. Purchase commitments.’ In his view, the sector’s rapid expansion has left companies vulnerable to a sharp correction. Oracle, whose stock surged from below $130 to nearly $350 during the 2024 AI frenzy, has since fallen about 60% amid mounting concern over the debt it used to finance a vast data-center buildout. Nebius, an AI cloud infrastructure company, saw its stock dip 13% on the day of Burry’s post but has still roughly doubled this year and risen ninefold in two years.
Neither Oracle, Nebius, nor Burry immediately responded to requests for comment. The investor has been warning for months that Big Tech is overinvesting in chips and data centers that could quickly become obsolete. He previously held bearish put options on Oracle and said in early August he had closed that wager at a ‘substantial’ profit. His latest short comes as Oracle reported that remaining performance obligations—a gauge of future contract revenue—soared 363% to $638 billion for the year ended May 31, while Nebius’ first-quarter revenue jumped nearly sevenfold year-on-year to almost $400 million. Yet Burry argues that such growth is built on unsustainable commitments and hidden liabilities.
Why Burry Sees a 'Homeresque Own Goal' in the AI Data Center Rush
The Oracle Debt Hangover
Oracle’s AI-fueled rally quickly reversed when investors began scrutinizing its balance sheet. The company borrowed heavily to expand cloud capacity, pushing its debt load to levels many analysts consider aggressive for a business with maturing legacy software revenues. Burry’s short entry near $145 reflects a view that the market had not adequately priced in the leverage and execution risk tied to this expansion. Despite the $638 billion in remaining performance obligations, skeptics note that converting these commitments into actual cash flow requires flawless delivery and constant technological relevance—two variables that the fast-moving AI hardware cycle puts under pressure.
Nebius’s Execution Gamble
Nebius, a younger and smaller AI infrastructure pure-play, faces a different version of the same risk. Burry called out the company’s ‘ton of execution risk,’ driven by high fixed costs, long data-center leases, and the challenge of securing long-term customer contracts in a field where the hyperscalers dominate. While revenue nearly hit $400 million in the first quarter and deferred revenue jumped 150% to $686 million, Burry’s conviction is that these numbers obscure liabilities that will crystallize if demand plateaus or if customers consolidate spending on larger, more established providers. His somewhat larger short size in Nebius suggests he sees the stock’s ninefold surge as particularly disconnected from fundamentals.
Burry’s AI Macro Call
Beyond the individual companies, Burry’s positions articulate a broader thesis: the AI infrastructure gold rush is inflating a bubble in off-balance-sheet commitments. Purchase obligations, lease backstops, and committed but unspent capex are piling up across the industry, creating a contingent liability overhang that he believes the market has ignored. If AI workloads shift to more efficient models or if enterprise demand cools, the massive capacity being built today could turn into a stranded-cost crisis. Bulls point to soaring revenue and contract pipelines as proof the demand is real, but Burry is betting that the financing structure behind that growth is fundamentally fragile.
What the Shorts Signal for Investors Tracking AI Cloud Stocks
- Monitor Oracle’s leverage and cash conversion: Oracle’s net debt-to-EBITDA ratio and the pace at which it converts its $638 billion in remaining performance obligations into actual revenue are critical leading indicators. A slowdown in conversion would validate Burry’s debt concerns and likely pressure the stock further.
- Watch Nebius’s deferred revenue trajectory: The company’s deferred revenue of $686 million as of March 31 is a key metric. If it stagnates or declines in the next quarterly report while fixed costs remain high, the execution risk Burry identifies will become more visible to the market.
- Track the hyperscaler pricing war: Amazon Web Services, Microsoft Azure, and Google Cloud are all aggressively expanding AI capacity and could squeeze smaller players like Nebius. Any pricing compression in AI cloud services would directly hit Nebius’s margins and contract pipeline.
- Burry’s own disclosures as a sentiment gauge: Because Burry now writes publicly about his trades, his future Substack posts on these positions—whether he covers or adds—could move the stocks. Investors with exposure should follow his commentary closely, not as investment advice but as a factor in short-term sentiment swings.
Risk & Opportunity Assessment
| Commercial Risk | High | Oracle's aggressive debt-fueled data-center expansion risks stranding capacity if AI demand growth slows; Nebius faces high fixed costs and uncertain customer commitment levels that could pressure its cash flow. |
| Competitive Risk | High | Both companies compete with hyperscalers (AWS, Microsoft, Google) that have deeper balance sheets and can sustain a pricing war, potentially eroding Oracle's cloud gains and Nebius's niche position. |
| Regulatory Risk | Low | No immediate regulatory changes specific to these companies are evident, though broader data-sovereignty or energy-consumption rules could affect data-center economics in the future. |
| Reputation Risk | Medium | A prominent skeptic shorting these names publicly could influence customer confidence and supplier negotiations if the market begins to question their financial sustainability. |
| Technology Disruption | High | Rapid advances in AI model efficiency or a shift away from GPU-intensive architectures could render current infrastructure investments obsolete, validating Burry's core concern. |
| Commercial Opportunity | Medium | If AI demand proves durable and off-balance-sheet liabilities are manageable, Oracle and Nebius could re-rate sharply as their revenue pipelines convert, creating a squeeze opportunity for the shorts. |
Comments 0