As valuations of artificial intelligence (AI) companies soar, substantial investments pour into the sector, raising renewed fears of an AI bubble. The level of enthusiasm for AI investments is unprecedented. Reports indicate that the total combined valuation of ten unprofitable AI startups globally surged by nearly $1 trillion over the past year, marking the fastest wealth expansion in history. Concurrently, these firms attracted over $200 billion in venture capital funding this year alone, accounting for two-thirds of total U.S. VC investment for the year. However, nearly all of these companies are currently operating at a loss.

Concerns about an AI bubble have intensified, arguably more than ever before, as valuations of many companies in the AI field have significantly increased, investment in AI continues unabated, and the AI ecosystem becomes increasingly self-referential, with model companies, infrastructure providers, and large-scale enterprises entering into agreements that blur the lines between customers, suppliers, and capacity providers. In light of these developments and mounting concerns, the notable rise in major tech stocks may be obscuring signs of broader market weakness, making the debate on the legitimacy of AI bubble worries a focal point in the market.

Previously, major U.S. banks reported record quarterly earnings, spurred partly by a boom in AI. However, several Wall Street executives have cautioned that the AI sector might be entering a phase of excessive excitement. This month, both the Bank of England and IMF President Kristalina Georgieva expressed concern over the U.S. stock market boom driven by AI enthusiasm. Furthermore, a recent survey by Bank of America has revealed that the proportion of global fund managers who believe the AI sector is in a bubble has reached an all-time high, coinciding with a robust increase in AI concept stocks this year. In the October survey, about 54% of respondents indicated that current valuations for tech stocks are too high; just a month prior, nearly half of the respondents disagreed with that assertion. Concerns over overvaluation in global equity markets also peaked in this same survey.

So, are these concerns justifiable or excessive? After years of optimism and continual market highs, some investors are now expressing worries about a bubble in the U.S. stock market, drawing parallels with the internet bubble era's capital spending frenzy followed by the subsequent collapse. Goldman Sachs has noted that while there are several concerning factors, they overall believe that the U.S. tech sector is not currently in a bubble (at least not yet). However, they are more concerned about the significant gap between public market valuations and those in the (higher) private market.

Goldman Sachs equity strategists pointed out that although some characteristics of the current period are reminiscent of past bubbles and the cyclicality of trading warrants caution, public market valuations and levels of capital market activity remain below the peaks seen during the internet bubble. They also noted that the so-called "Magnificent Seven" companies continue to generate excess free cash flow, buy back shares, and pay dividends—practices that were rare during the internet bubble period. Consequently, they seem hesitant to label the current public market scenario as a bubble, alluding to the notion that 「AI may not have become a bubble yet.」

Byron Deeter, a partner at Bessemer Venture Partners, is more optimistic about the boom in AI capital expenditures. In contrast, Sequoia Capital analyst David Cahn has a differing view. He believes that only through General Artificial Intelligence (AGI) can the justification for large-scale data center construction by 2030 be validated while remaining optimistic about the numerous opportunities in private AI application firms. New York University professor Gary Marcus also retains skepticism about the technology itself, at least in its current form. Impactive Capital managing partner Lauren Taylor Wolfe has also pointed out that the AI industry is in a bubble that is bound to burst eventually, drawing comparisons to the late 1990s internet bubble. Her core concern lies in the severe disconnect between investment in the AI field and the returns it offers: 「Trillions of dollars are being planned for investment in AI, while the tech giants are only generating a few hundred billion in free cash flow. Who can prove that trillions in profits can be generated in the next five years? That is simply not achievable; it does not hold up mathematically.」

Nevertheless, Goldman Sachs maintains that the current uptrend—particularly the strong performance of tech stocks—has primarily been driven by fundamental growth rather than irrational speculation. So how should investors position themselves? Goldman Sachs concludes that opportunities in the tech sector remain, but diversification is wise. In the tech space, they believe stocks poised to benefit from potential AI disruption that are not fully recognized represent valuable growth stories, and they see opportunities across AI infrastructure, platforms, and applications at various levels; they also affirm that investment themes related to leading semiconductor firms remain solid. Sequoia Capital’s Cahn likewise sees significant opportunities in AI application companies, mainly found in today’s private markets: 「High prices, but high-quality business models.」

Deutsche Bank emphasizes that identifying bubbles is nearly impossible, as no consensus exists on the precise definition of "asset prices significantly above intrinsic value." Historical experience suggests that bubbles do not develop linearly and typically evolve through multiple rounds of ups and downs. In summary, Deutsche Bank notes that markets may remain "irrational" longer than investors maintain solvency. The bank advises investors to adopt long-term holding strategies to obtain the risk premium necessary to compensate for equity investment risks.