Countdown to the burst of the AI bubble? The trading logic surrounding AI is moving from "pure grand narrative" to "real returns."

Understanding that from Nvidia to Oracle and Broadcom announced their performance last week, even if their core performance indicators exceeded expectations, they failed to strengthen the so-called "AI bull market narrative" and did not boost global investors' "AI faith". Finally, last Friday, this skepticism exploded, causing popular technology stocks such as Oracle, Nvidia, and Broadcom, which have benefited from the AI investment craze since the end of 2022, to collectively plummet. Top Wall Street investment firms are increasingly questioning the sustainability of the unprecedented global AI investment craze led by the seven tech giants (the Magnificent Seven) in the US stock market. They believe that the so-called "AI bubble" is forming with a killing power comparable to the "Internet bubble" of that year. More importantly, these top institutions are beginning to bet on which major events or important moments could completely burst the bubble. Three years have passed since OpenAI's globally popular ChatGPT was released, sparking a global AI investment frenzy and a super-hot trend of global enterprises deploying AI. Although funds are still flowing towards leaders in the AI computing industry such as Nvidia, Broadcom, Google, Oracle, and Micron, who have benefited from the AI investment frenzy, doubts about the sustainability of this stock market AI boom are deepening, with some investors believing that the "best time" of this AI investment craze has come to an end. From the "AI chip super hegemon" Nvidia's announcement of incredibly strong performance failing to stop the recent decline in stock prices, to Oracle's stock plummeting after revealing continuously rising AI computing expenses, and Oracle's bonds dropping to yields close to "junk grade" with CDS hitting the highest levels since the 2009 global financial crisis, to the market's turning cold on a large number of technology companies related to OpenAI that are still burning money and with no profit in sight, the market's skepticism is continuously increasing. Looking ahead to 2026, some investors are debating whether to reduce their exposure to AI investments before a potential AI bubble bursts, or to increase their bets on this disruptive and transformative technology. The decline in Nvidia and Oracle stock prices may just be a trailer? Market winds are shifting: Wall Street sees the arrival of an artificial intelligence bubble and is no longer stubbornly all-in on the "Magnificent Seven" of the tech giants. Jim Morrow, CEO of Callodine Capital Management, says we are at a "critical moment." "Even though these stories are well told and the logic is still sound, it is now time to verify whether the investment returns are good enough." As depicted in the chart above, the Nasdaq 100 Index, known as the "global tech stock barometer," has risen more than double since the introduction of ChatGPT, indicating that the market's patience for massive AI infrastructure investments is reaching its limit, with investors eagerly awaiting substantial returns on investments in AI GPU/AI ASIC and other AI hardware infrastructure. Anxiety surrounding popular trades related to AI focuses on whether the applications or intelligent agents will eventually scale up on a large scale, the tremendous development and training costs involved, and whether global enterprises and consumers will continue to pay for these transformative and efficient services. The answers to these questions will have a significant impact on the future of the global stock market. The S&P 500 Index has accumulated approximately $30 trillion in gains over the past three years in a "super bull market," largely driven by the world's largest tech giants (such as the Magnificent Seven of the US tech giants) and powered by companies pushing massive investments into AI computing infrastructure (such as Micron, TSMC, and Broadcom) and power system suppliers (such as Constellation Energy). If they stop rising, stock indexes will be severely dragged down in an instant. The so-called "Magnificent Seven" tech giants, which account for about 35% of the S&P 500 Index, including Apple, Microsoft, Google, Tesla, Nvidia, Amazon, and Meta Platforms, have been the core driving force behind the index's continuous record highs. They are also seen by top Wall Street investment firms as the most capable combination to bring huge returns to investors under the backdrop of the largest technological revolution since the Internet age. Sameer Bhasin, Head of Investments at Value Point Capital, points out that stocks benefiting from the AI boom tend to see significant pullbacks not when growth rates decline, but when the growth rates stop accelerating. Therefore, the market is beginning to doubt the formation and abrupt bursting of this irrational bubble. Goldman Sachs recently released a report indicating that large tech companies that invest the most capital in hopes of "winning" the AI equipment race may face prolonged periods of highly uncertain investment returns. Goldman suggests that the AI story is now entering the "final chapter of the prologue," with the idea that most of the value created by AI will belong to AI language models, and the widespread bullish market sentiment around anything related to AI is likely coming to an end, as the market becomes more discerning of who the real beneficiaries will be. Multiple Wall Street major banks, including Goldman Sachs, Bank of America, Yardeni Research, and Morgan Stanley, have recently stated in their annual summaries and outlooks that market skepticism is growing over the high valuations of tech stocks and the enormous AI investments' ability to generate significant returns. This has shifted the market's focus towards industrial and energy sectors, traditionally cyclical sectors, rather than the high-valued "Magnificent Seven" tech giants such as Nvidia and Amazon, located at the center of the "AI bubble." Ed Yardeni, founder of Yardeni Research, even suggested for the first time since 2010 that investors should "reduce" the Magnificent Seven tech giants relative to the rest of the S&P 500 Index. However, reasons to remain optimistic still exist: the tech giants heavily investing in AI have strong resources and a solid fundamental background, promising to continue pouring money into AI in the coming years. Meanwhile, AI developers (such as Google and OpenAI) continue to advance the AI boom with new AI big models, leading to a more optimistic outlook. Therefore, the debate in the market regarding the "AI bubble" and whether the AI frenzy can sustain long-term continues, deepening with each passing day. Below are some key trends that Wall Street investment institutions believe are crucial to focus on when navigating through this "AI turbulence". Capital-raising ability of AI application developers such as OpenAI OpenAI alone, an AI developer led by Sam Altman and became the world's most valuable startup company in October, plans to spend a staggering $1.4 trillion in the coming years on new AI data center construction and expansion. However, the company's revenue is far below its operating costs, not to mention its heavily loss-generating balance sheet. Media reports in September suggested that OpenAI may not begin to generate positive cash flow until 2030. So far, OpenAI's fundraising has not been difficult, having raised $40 billion earlier this year from investors such as SoftBank Group led by Masayoshi Son. In September, Nvidia pledged to invest up to $100 billion, as part of a series of investment deals aimed at "infusing blood" to customers, sparking concerns in the market about "cyclical financing" in the AI tech industry similar to the dot-com bubble era. If institutional investors become hesitant to continue adding more capitalindicating a weakening capital-raising abilityOpenAI may face significant troubles, which could negatively impact companies within its ecosystem, such as Nvidia and AI cloud computing service provider CoreWeave. Eric Clark, Senior Portfolio Manager at Rational Dynamic Brands Fund, notes that when "trillions of dollars" crowd into a few themes and stocks, at the first sign of short-term funding difficulties in that theme or when valuations stretch to a level that it becomes difficult to continue growing at the same pace, funds may begin to withdraw simultaneously. Many companies besides OpenAI depend on external financing to chase their AI ambitions. Oracle's stock price once surged due to a significant increase in cloud computing orders, but building incredibly massive AI data centers requires massive funds, prompting the company to raise billions of dollars through bond issuances. The use of debt will continue to increase pressure, as debt holders need to receive interest cash payments on time; and equity investors typically profit more through stock price appreciation. On Thursday, Oracle's stock price tanked after announcing that its capital expenditures for the second quarter of the fiscal year significantly exceeded expectations, and the growth rate of cloud computing sales was below the average expectations of Wall Street analysts. Media reports on Friday stated that some AI data center projects developed by Oracle for OpenAI were severely delayed, leading to further decline in Oracle's stock price and dragging down other stocks related to AI computing infrastructure. Meanwhile, an index measuring Oracle's credit risk (i.e., CDS) rose to the highest level since 2009. However, an Oracle spokesperson stated in a release that the company remains confident in meeting debt obligations and future expansion plans. The chart above shows that Oracle's stock price has fallen by more than 40% from the September high pointhowever, the stock still yields positive investment returns for the year. Kim Forrest, Chief Investment Officer at Bokeh Capital Partners, notes that credit market investors are "smarter" or at least more concerned about the "right things"they are eager to receive their money back on time. Capital expenditure processes of the tech giants Wall Street analysts' expected data shows that Alphabet, Microsoft, Amazon, and Meta are expected to spend over $400 billion in capital expenditures in the next 12 months, with the majority allocated to new AI data center construction and expansion. Despite these tech giants reaping profit growth related to AI from cloud computing and digital advertising businesses, it is still insufficient to cover the pace at which costs are rising. Michael O'Rourke, Chief Market Strategist at Jonestrading, warns that once growth expectations plateau or slow down to any extent, the market may begin to think "there are significant problems here." Bloomberg Intelligence (BI) statistical data shows that Wall Street analysts expect an 18% earnings growth rate for the "Magnificent Seven" tech giants in 2026, the slowest in four years and only slightly higher than the overall expected growth rate of S&P 500 index component companies. The depreciation expenses stemming from the "mad construction wave" of data centers are also a significant concern for the market. Alphabet, Microsoft, and Meta had a combined depreciation of about $10 billion in the fourth quarter of 2023; by the end of the quarter that ended in September, this number was close to $22 billion; and Wall Street expects it to be around $30 billion by this time next year. The chart above shows the rising depreciation expenses of the tech giants. These financial pressures may squeeze share buybacks and dividends (i.e., returning cash to shareholders). Research data from BI shows that Meta and Microsoft are expected to have negative free cash flow after accounting for shareholder returns in 2026, while Alphabet is roughly at breakeven. A deeper concern in the market is that this unprecedented AI spending reflects a strategic shift. For a long time, the valuation logic of large tech companies has been based on the logic of "achieving rapid and strong profit growth at low costs and accumulating massive free cash flow over time." However, the massive AI spending plans have somewhat overturned this traditional valuation model. O'Rourke of Jonestrading indicates that if these giants continue to "lever up" to build, hoping for strong future monetization, valuation multiples may significantly contract; O'Rourke adds that if the AI ambitions cannot be realized in the end, this shift could be a severe mistake. What the market needs is rational prosperity investment, not irrational excitement. While the tech giants are trading at high valuations, this is not yet at an extreme bubble bursting scenario as compared to previous market frenzy periods. The market often contrasts this period with the dot-com bubble of the 2000s, but the scale of the rise brought on by AI is not equivalent to the Internet era. For example, BI statistical data shows that the Nasdaq 100 index, mainly consisting of tech stocks, is currently trading at around 26 times estimated earnings; whereas during the peak of the dot-com bubble, this multiple exceeded an astonishing 80 times. The reason for the higher valuations during the Internet bubble era was not only due to exaggerated stock price increases but also because the tech companies of that time were "younger," had weaker profit capabilities, and sometimes had no profits at all. Tony DeSpirito, Global Chief Investment Officer and Portfolio Manager for Fundamental Equities at BlackRock, believes that the current valuations are not at "dot-com bubble levels." However, this seasoned asset management professional notes that this does not mean there are no sporadic speculative behaviors or irrational excitement localized in connection to AI-related stocks. In the current market, AI-related popular stocks that are considered "overpriced" include the AI application leaders focusing on "AI + data analytics" such as Palantir (with a PE ratio exceeding 180 times) and Snowflake (nearing 140 times). But Nvidia, Alphabet, and Microsoft are all below 30 times, and amid the AI investment frenzy surrounding them, these valuations are relatively moderate. These relatively positive factors are making investors dilemma: the risks are visible, but most companies' valuations have not reached a level of panic-inducing. The key question is where the AI trades are heading next, whether the funds will continue to favor leaders in the AI computing industry such as Nvidia and Broadcom, or be more inclined towards companies operating at the AI application layer or those utilizing AI for substantial cost reduction and productivity enhancement. Bhasin of Value Point believes that this kind of "groupthink" will begin to crack, not necessarily leading to a complete collapse like in 2000, but a significant "rotational change" will occur.