"Big short" is losing its temper? Berli angrily criticized Nvidia's $500 billion financing as "financial engineering of 2008," yet the market supports the demand for AI computing power with real money.

On Wednesday, at the beginning of the U.S. stock market, Michael Burry, known as the "Big Short," launched a fierce critique of the $500 billion artificial intelligence infrastructure financing arrangement led by "AI chip superpower" NVIDIA Corporation (NVDA.US). He likened the complex financing network composed of private credit and asset-backed securities to the systemic financial risks accumulated on the eve of the 2008 financial crisis, calling it a "Wall Street exclusive scheme reminiscent of the tricks of 2008." Burry has been posting apocalyptic comments about the impending doom on his Substack subscription platform, and while global capital continues to flock to the theme of AI computing infrastructure, he has significantly shorted popular AI tech stocks. The protagonist of the film "The Big Short" has recently intensified his shorting activity, focusing more clearly on crowded trades related to AI computing infrastructure and the capital expenditure cycle for AI semiconductors. For instance, Burry expressed a continued bearish position on NVIDIA Corporation, Tesla, Inc., Micron, Applied Materials, Palantir Technologies, and the iShares Semiconductor ETF through put options or short positions. In fact, he has upgraded his short portfolio into a systematic bet against "AI capital expenditures + AI valuations." However, the recent strong rebound in AI semiconductors and a broader resurgence in AI computing themesespecially driven by the moderate CPI growth data released in the U.S. on Wednesdayare clearly putting significant pressure on this short AI trading mechanism. Notably, the current exhaustion of CoreWeave's capacity, the extension of the A100s lifespan to 2029, Nebiuss price increase and $40 billion customer commitments, along with the decline in NVIDIA Corporation's credit spreads, have all tilted the balance significantly toward Jensen Huang/Morgan Stanley's framework of "financing assets for AI factories." Some investors have commented that Burry's recent harsh criticism of NVIDIA Corporation's financing activities, labeling the $500 billion financing as a Wall Street financial engineering akin to 2008, seems to indicate that the short seller is becoming desperate. Meanwhile, CoreWeave has raised its 2026 capital expenditure forecast to $3539 billion, with its stock soaring post-earnings, and Nebius reported a staggering year-on-year revenue increase in Q2, stating that under current conditions, it could even sell out its planned capacity for 2027. Clearly, the real-world trading and the supply chain aspects are strengthening the bullish narrative that AI computing demand is still genuine and scarce. In contrast, Burry, who is shorting Nebius, may now be facing pressure to cover. Burry's Scion Asset Management, which he previously led, will be deregistered by November 2025, so what we see more of now are Burry's personal trades disclosed actively via his paid subscription platformSubstacks "Cassandra Unchained"rather than traditional 13F holdings snapshots. Recent public information indicates that he still maintains short exposure to NVIDIA Corporation, Palantir, the SOXX Semiconductor ETF (Philadelphia Semiconductor ETF), and QQQ (Nasdaq-100 Index ETF), extending some NVIDIA Corporation/QQQ put options to 2027, while also disclosing a direct short on Nebius at around $212 on August 7, stating that the position size is relatively larger. However, after the strong earnings reports from Neocloud's CoreWeave and Nebius on Wednesday, the stock prices of these two new cloud companies soared nearly 20%. Palantir, following its explosive earnings report on August 4, saw its stock rise approximately 29.5% in a single day, resulting in a record paper loss of around $3 billion for the entire Palantir short camp; however, Burry had recently re-established out-of-the-money put options on Palantir due to expire in 2027, prior to this strong earnings announcement. Thus, his short portfolio is clearly under new Mark-to-Market pressurewhen market prices rise, short sellers (the bears) incur unrealized losses because the borrowed assets they sold have increased in price, forcing brokers to continuously demand margin replenishments. The $500 billion AI financing is under attack from the Big Short! Burry questions the leveraged cycle dominated by NVIDIA Corporation. One of the seven major tech giants in the U.S., NVIDIA Corporation has signed a much-anticipated memorandum of understanding with Apollo Global Management (APO.US), Blackstone Inc. (BX.US), BlackRock, Inc. (BLK.US), Brookfield (BN.US), Goldman Sachs Group, Inc. (GS.US), and KKR & Co. (KKR.US) to establish the first such AI computing infrastructure financing platforms globally, with an initial financing scale reaching as high as $500 billion. As a platform facilitating deals, NVIDIA Corporation plans to support these financing transactions with a "residual value mechanism," providing up to 25% guarantee for individual projects. If a project encounters difficulties, the chip giant will seek to minimize losses by heavily selling its exclusive AI chip products in high demand or find new corporate tenants for the related AI computing capacity. However, Burry has accused on his subscription platform that there exists a highly leveraged and cyclical funding chain behind these transactions. According to an infographic he created, annuity premiums paid by American retirees to Athene, a subsidiary of Apollo (APO.US), are transferred through an offshore reinsurance company based in Bermuda and further leveraged by Apollo into asset-backed debt. It is alleged that Apollo issued $3.5 billion in debt to a special purpose entity called Valor while NVIDIA Corporation injected $1.9 billion in equity capital directly. Subsequently, VCI purchased $5.4 billion worth of NVIDIA Corporation's GB200 GPU, which were then leased to xAI for its Grok supercomputer cluster. For skeptics like Burry of this AI boom, such complex financing structures artificially elevate revenue figures while shifting depreciation risks onto unsuspecting retirees. I think I know how this will ultimately turn out, Burry wrote in a post on X platform. Recently, he has further increased his direct short bets against leading players in the AI computing industry, such as NVIDIA Corporation, through put options. Meet the new boss. Theyre really no different from the old bosses of 2008, Burry sarcastically remarked. Is the $500 billion really a "financial trick" or financing for an AI industrial revolution? With credit spreads declining and Neocloud surging, capital seems to temporarily choose to believe Huang Jen-hsun. After Nebius announced its promising earnings report on Wednesday, its stock surged nearly 20% in early U.S. trading, coincidentally getting close to Burry's short cost of approximately $212. Some Wall Street analysts even believe that short covering could become a strong amplifier of the company's stock price increase. Undoubtedly, Burry is experiencing severe short price pressure, but there is currently no reliable evidence that he has faced forced liquidation or must cover. Recent stunning results and strong outlooks from ASML Holding NV ADR and Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR, the two most important capacity and supply forces in the upstream of the AI computing supply chain, alongside the latest strong performances and outlooks from CoreWeave and Foxconn with Nebius, are, in fact, sending an important signal to the global stock market: the AI computing supply chain is gradually transitioning from the "super cycle of AI capital expenditures to train AI large models" to a new phase of "exponential expansion of AI inference computing demand driven by large-scale applications of intelligent agents." These latest signals strongly counter the pessimistic rhetoric of "computing surplus" that has caused significant declines in themes related to AI computing, especially in the AI semiconductor sectors. CoreWeave, closest to the end demand for AI computing, has just raised its forecast for capital expenditures in 2026 to an astonishing $3539 billion, with Q2 revenue significantly exceeding expectations and claiming to continue signing cloud computing contracts under favorable terms with limited supply. Meanwhile, SK Hynix has also reported record quarterly profits and disclosed that it has signed long-term supply agreements with about ten core customers, reflecting that demand for high-performance storage like HBM has shifted from short-cycle orders to multi-year locked volumes. Different segments across the supply chain are all sending positive signals that AI computing demand is still accelerating. At this moment, does the market believe Burry or believe Huang Jen-hsun and Morgan Stanley? The answer is quite clear: marginal capital is currently clearly on the latter's side, but Burry's long-term risk argument has not been declared invalid. The most critical evidence is not stocks but the credit market: when the $500 billion plan was first announced, concerns arose about NVIDIA Corporation bearing substantial off-balance sheet risks; Huang Jen-hsun later clarified that this $500 billion is intended to gradually mobilize third-party capital, not NVIDIA Corporation's income, nor funds already committed to a clientApollo, BlackRock, Inc., Blackstone, Brookfield, Goldman Sachs Group, Inc., and KKR will independently underwrite each project, with NVIDIA Corporation's support capped at approximately 25% of the collateral value of each project (Residual-value Backstop). Subsequently, NVIDIA Corporation's 2056 bond spread relative to U.S. Treasuries narrowed by about 2 basis points to 113 basis points, and the 5-year CDS narrowed by as much as 5 basis points to 72.11 basis pointsthis is a very clear message from the bond market: investors adjusted NVIDIA Corporation's credit risk pricing after Huang Jen-hsun clarified the structure. The assessments from Morgan Stanley and Bank of America are also broadly aligned: third-party professional capital leading due diligence and decision-making significantly weakens the "NVIDIA Corporation spends its own moneycustomers use that money to buy NVIDIA Corporation GPUsNVIDIA Corporation inflates revenues" cycle logic of pure speculation; at the same time, a usage-based linkage to actual revenues may allow NVIDIA Corporation to share cash flows throughout the AI factory lifecycle beyond one-time GPU sales. In other words, the market currently does not view this $500 billion as having "no credit risk," but rather perceives the risks as more akin to underwritable, priceable infrastructure financing risks, rather than the distorted asset demand seen before 2008. Burry bets on a repeat of 2008 Wall Street financial engineering, while CoreWeave and Nebius deliver evidence of soaring demand. Burry focuses on leverage, GPU depreciation, and the potential tail credit risk of asset securitization; meanwhile, bullish capital is trading based on computing supply not meeting demand, long-term contracts, pricing power, and sustainable cash flows. This platform from NVIDIA Corporation is primarily designed to mobilize third-party capital, with NVIDIA Corporation providing up to about 25% in residual value support per project; simultaneously, CoreWeave has disclosed over $100 billion in order backlog, while Nebius has reported over $40 billion in customer commitments. The strongest challenge to Burry's "2008-style financial engineering" argument comes from the two Neocloud leadersCoreWeave and Nebiusboth of which today provided evidence of physical demand, pricing, and residual values. CoreWeave's Q2 revenue reached approximately $2.6 billion, a year-on-year increase of 112%, and its order backlog surged to $104.2 billion, not including new commitments exceeding $25 billion signed at the start of the quarter. The company clearly stated that near-term capacity is "effectively sold out," so it is signing new contracts at increasingly favorable prices, with JPMorgan analyst team even attributing this price increase to the improving demand environment and the accelerated ROI for clients in cloud computing and AI applications. More importantly, CoreWeave directly provided a powerful counter-example to Burrys argument that GPUs only have a 23 year economic life: the company recently signed a new contract extending through 2029 for its A100 launched in 2020, and the pricing remains attractive; management also clearly stated that the old generation of GPU clusters is still essentially sold out, and re-leasing after the initial contract ends constitutes incremental revenue beyond already returned investments, while observing longer usage cycles and higher prices. Nebius's evidence is even more aggressive: Q2 revenue reached $582.3 million, exceeding the consensus expectation of $572.75 million, with AI Cloud (AI-related cloud computing business) revenue nearing a sixfold year-on-year increase, having signed four contracts with an average value of over $1 billion in a single quarter. The value of new customer contracts grew more than ninefold sequentially, with commitments exceeding $40 billion from customers and expected customer prepayments exceeding $9 billion this year; the company even stated that under current conditions, it could sell out all planned capacity for 2027. This latest set of operational data strongly supports Huang Jen-hsun's assertion that demand is real, and produced AI computing is a scarce productive asset as of August 2026, rather than supporting the notion that financing created false GPU demand.