Howard Marks, co-founder of Oaktree Capital Management, says the surge in AI-linked equities does not yet meet his definition of a full market bubble, largely because traders are openly debating the possibility of one. In Marks’s view, the most dangerous phase of a bubble emerges when participants stop questioning valuations and become unaware of the risks attached to rising prices.
That skepticism remains visible across markets after a sharp reversal in major technology shares. The Nasdaq 100 reached a record earlier this month before falling 5.5% last week, reviving concerns that enthusiasm around artificial intelligence has carried a narrow group of stocks too far, too quickly.
Marks’s argument does not dismiss the risks in the market. It places the current rally in an uncomfortable middle ground: valuations are elevated, index concentration is extreme, and momentum has rewarded a small group of AI beneficiaries, but widespread caution may be preventing the kind of collective euphoria that typically defines a historic market peak.
Marks sees bubbles as a breakdown in judgment
Marks described a bubble as more than a period of expensive asset prices. His framework requires several conditions to appear together: valuations detach from rational analysis, risk-taking no longer reflects underlying conditions, and emotions reach a point where participants lose control of their decision-making.
He also pointed to the social pressure created by rapid gains. When a limited number of traders make outsized returns, others can feel compelled to chase the same assets rather than risk missing the rally. That behavior has appeared in previous speculative periods, especially when market leadership becomes concentrated in a fashionable theme.
The current debate over AI shares complicates that picture. Warnings about excessive valuations have come from market commentators, portfolio managers, and analysts, while volatility in recent weeks has shown that buyers remain willing to reconsider their positions. A market that is constantly questioning whether it is a bubble does not fit Marks’s description of widespread obliviousness.
That distinction does not offer a clean all-clear for technology stocks. It suggests the market may be vulnerable to corrections without necessarily having entered the final, indiscriminate stage associated with the dot-com boom or other major speculative episodes.
Valuations remain high, but below dot-com extremes
Marks compared today’s largest technology companies with the dot-com era to argue that current valuation levels are less extreme on one commonly used measure. The five largest technology stocks in the S&P 500 trade at roughly 23 times forward earnings, compared with about 65 times forward earnings for comparable market leaders one month before the 2000 peak, according to the figures he cited.
Forward price-to-earnings ratios measure a company’s share price against expected profits over the coming year. They can be useful in judging whether markets are pricing in unusually optimistic growth assumptions, though the measure also depends on forecasts that can change rapidly.
Other valuation measures paint a less reassuring picture. The cyclically adjusted price-to-earnings ratio, or CAPE, stands near 41, close to the 44 level recorded in 1999, according to the figures cited in the report. CAPE compares stock prices with inflation-adjusted average earnings over a 10-year period, reducing the effect of temporary swings in profits.
The gap between those measures illustrates the debate surrounding AI-related stocks. The largest companies generate substantial revenue and profits, unlike many dot-com-era firms that traded on distant promises of future growth. Yet the broader equity market is also valued at levels that leave less room for disappointment in earnings, interest rates, or economic growth.
A handful of companies now shape the index
Market concentration is one of the clearest sources of vulnerability. Nvidia, Apple, and Microsoft account for about 21% of the S&P 500 by index weight, according to the report. In October 1999, it took nearly 10 stocks to reach a similar portion of the benchmark.
That concentration means large moves in a few companies can determine the direction of the entire index, even if many other stocks are stable or falling. It also makes passive funds more exposed to the same names because index-tracking strategies automatically allocate more capital to companies whose market values grow the fastest.
The Nasdaq 100 has outperformed the S&P 500 by 250 percentage points over the past decade, the report said. That gap has made it difficult for portfolios to reduce their exposure to mega-cap technology companies without accepting the risk of lagging the main benchmarks during an extended period of AI-driven gains.
Signs of narrowing market breadth have added to the unease. Breadth measures how widely market gains are shared across individual stocks. When only a shrinking group of companies drives index returns, markets can become more sensitive to setbacks in those leaders.
Credit markets could test AI optimism
The recent pressure on CoreWeave has brought financing risk into sharper focus. Its shares nearly doubled between January and early May, according to the report, but last week’s sell-off left the stock negative for the year. Its bond yields also rose sharply, signaling that lenders were demanding greater compensation to hold the company’s debt.
Higher yields raise the cost of borrowing and can put pressure on companies whose valuations depend heavily on future growth. That is particularly relevant for AI-linked firms building data centers, acquiring advanced chips, or funding infrastructure with substantial capital requirements.
Inflation, midterm election uncertainty, and the Iran war have also contributed to periods of more defensive positioning in 2026, according to the report. Those factors can affect technology shares disproportionately when they push bond yields higher, since high-growth valuations are especially sensitive to changes in discount rates.
For cryptocurrency markets, the equity debate matters less as a claim that digital assets are a hedge and more as a measure of risk appetite. Bitcoin and other liquid tokens have frequently traded alongside technology and growth assets during periods of shifting rate expectations. A sharp reassessment of AI valuations could therefore influence crypto positioning through broader deleveraging and reduced demand for speculative exposure.
Marks’s view leaves room for both caution and nuance. Today’s AI trade carries several features associated with late-cycle markets—high valuations, concentrated leadership, crowded performance trends, and sensitivity to credit conditions—while retaining a degree of visible skepticism that has historically been absent at the most euphoric peaks.
Concerned about bubbles and AI-driven markets? Deepen your macro view with Toobit’s latest insights in Today AI Moves Shake Markets.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
