The 2020s might enter the history books as an era where the market skipped the stairs and kept taking the elevator either way.
The recent surge since the beginning of Q2 is the best example. According to BullTheory, U.S. equities added nearly $10 trillion in market capitalization in just 39 trading days. Yet, the realized volatility has become highly asymmetric. On up days, it is running roughly seven times as much as on down days.
Historically, that kind of imbalance tends to appear near exhaustion points, not in healthy bull markets.
Per Evercore ISI, current momentum "feels like 1999," with thrust levels comparable only to the 1982 breakout and the final phase of the internet bubble. Meanwhile, the divergence between cap-weighted indexes like the State Street SPDR S&P 500 ETF (NYSE:SPY) and equal-weighted ones like Invesco S&P 500 Equal Weight ETF (NYSE:RSP) continues to widen. It is a sign that leadership is narrowing rather than broadening.
Lagging AI Deployment
The structural asymmetry inside the market is even more alarming. AI-related companies now account for roughly 40% of total market capitalization. Meanwhile, healthcare stands at 8.3%, its lowest weighting in three decades. Yet, despite catching attention from institutions, the sector has been a stubborn underperformer. The Vanguard Health Care Index Fund (NYSE:VHT) remains around 5.5% in the red year-to-date.
At the same time, the top 10 stocks command approximately 35% of the S&P 500's total index weight. The concentration is a structural fragility. When a handful of names drive returns, it doesn't take much to reverse the trend.
That catalyst might come from AI scalability – or, to be more precise, from lack of it.
Arend Kapteyn, Global Head of Economics & Strategy Research at UBS, recently reviewed a survey that showed the scale of disappointment. The data showed that, as of March 2026, only 19% of companies surveyed reported a successful deployment of AI at a scale, compared to about 10% by March 2024.
In contrast to leaps in underlying technology, successful AI deployment advances by a mere 3% annually.
The question becomes: if enterprises cannot quickly deploy AI at a scale, how quickly can they start paying for it? If monetization fails to keep pace with capital expenditure, how quickly does capex become a balance sheet time bomb?
Cerebras IPO Signal
That concern makes tomorrow's Cerebras Systems IPO especially important. Reuters reported that the deal was more than 20 times oversubscribed, forcing the company to raise its offering range to $150–$160 per share and 30 million shares, up from $115-$125 per share and 28 million shares.
Despite reporting net income in 2025, Cerebras still showed an underlying operating loss on an adjusted basis. Yet, its valuation swelled from $8.1 billion last September to potentially more than $50 billion now.
Such stories are exactly the kind of setup that turns a narrative into a trap. Huge up-front capex, uncertain monetization, and investors paying for growth that has not yet arrived. It's why the "scaling" story starts to look less like secular certainty and more like leverage in disguise.
In 1999, extreme upside volatility preceded violent mean reversion once liquidity conditions tightened and narrative momentum broke. Today's market, while far more informed, shows similarities.
Euphoric concentration, speculative IPO excess, and deteriorating breadth are all present, while sticky inflation limits how aggressively the Fed can respond.
Image: Shutterstock
Login to comment