IBM's worst crash exposes earnings bubble hidden beneath AI boom

On the afternoon of July 14, just one day after IBM suffered the worst single-day stock crash in its 115-year history, economist Steve Hanke—a professor of applied economics at Johns Hopkins who advises governments including the Treasury Department and White House—offered a diagnosis of what he saw as a fundamental market mispricing. IBM shares had fallen 25%, erasing roughly $40 billion（約6.4兆円） in market value, on preliminary second-quarter earnings that included revenue of $17.2 billion（約2.8兆円） (missing consensus of roughly $17.9 billion（約2.9兆円） by about 3.7%) and adjusted EPS of $2.93 (versus expected $3.02). The company signaled only 1% revenue growth instead of the 5% expected by the market.

What made this selloff remarkable was its intensity relative to the offense. IBM was still growing; the miss was modest by historical standards. Yet the market's reaction was steeper than Enron's collapse when the SEC opened its accounting inquiry. That same day, JPMorgan had posted net income of $21.2 billion（約3.4兆円）—the highest quarterly profit for any bank in U.S. history—while Goldman Sachs reported an 84% jump in net earnings attributable to common shareholders, to $6.4 billion（約1兆円）, with total revenues hitting $20.34 billion（約3.3兆円）, up 39%. The contrast pointed to what Hanke called a "dual bubble" forming in AI markets: one a classic valuation bubble of price versus earnings (exemplified by the CAPE Shiller index), but a more dangerous one lodged in the earnings themselves.

Unlike traditional bubbles where prices race ahead of reported profits—as in 2000, when P/E ratios looked obviously stretched—an earnings bubble inflates the profits themselves, making valuations appear deceptively reasonable even while the market is dangerously mispriced. BCA Research's Peter Berezin has argued for months that today's AI trade is "primarily an earnings bubble rather than a valuation bubble," noting such bubbles historically cluster in boom-bust industries: pre-2008 banks, pandemic-era work-from-home stocks, and cyclicals like natural resources, airlines, and semiconductors—the last now sitting at the center of the AI capex story. The danger of earnings bubbles is their detection problem. Analysts typically cut profit estimates only after stocks have already fallen, leaving little early warning. When they burst, they leave behind real excess capacity—data centers, chip fabs, server farms—rather than just erasing paper gains. In late May, Berezin noted Wall Street analysts are "not particularly good at predicting when earnings bubbles will burst" because stocks begin falling before profit estimates do.

IBM's own aftermath bore out that exact detection lag. BofA and UBS both trimmed estimates, but only after the stock had already cratered 25%, with BofA cutting its price target to $280 from $330 and UBS holding its target at $236 while still lowering 2026 EPS forecasts—reactive moves, not predictive ones. Yet even after the selloff, Street opinion split sharply. BofA kept a Buy rating, arguing IBM remained "well positioned" once execution issues cleared, while HSBC downgraded to Reduce and Goldman warned the results would "fully validate the software bear case scenario." IBM CEO Arvind Krishna had written an unusually candid letter acknowledging underperformance, noting that conditions required "our teams to execute perfectly" and "this quarter we faltered," offering "not excuses, but … realities." Financial commentators immediately questioned whether IBM was a "canary in the tech coal mine" (per the New York Times' DealBook) or a "warning to the IT sector" signaling the actual arrival of the "SaaSpocalypse" that had spooked markets earlier in the year—the theoretical displacement of traditional software by AI.

Hanke's core argument connected IBM's crash to the monetary mechanism underneath both bubbles. The bull case for AI stocks has rested on the observation that today's AI leaders—Nvidia, Alphabet—generate real cash flow, unlike profitless dot-com names, and that S&P 500 valuations near 22x forward earnings sit below the 25x-plus threshold usually associated with true bubbles. But this defense addresses only the valuation side; it says nothing about whether the earnings themselves—swelled by capex cycles, circular AI investment, and easy money from private banks—are sustainable. Hanke pointed to record bank profits not as evidence of suspicious activity but as revelation of the monetary mechanism most investors misunderstand: it is private banks, not the Federal Reserve, creating the money fueling both the asset-price inflation and the reported earnings that justify those prices. JPMorgan CEO Jamie Dimon himself seemed to echo this concern, calling earnings "close to as good as it gets" before expressing concern at too much "exuberance" in markets. Whether IBM's crash represents the first visible crack in the earnings story underlying AI valuations, or merely a single-stock anomaly, would be answered by the rest of earnings season.

The IBM crash on July 14 arrives at a moment of sharp contradiction in financial markets. While JPMorgan and Goldman Sachs reported record profits—JPMorgan's $21.2 billion（約3.4兆円） net income the highest in U.S. banking history, Goldman's net earnings up 84% to $6.4 billion（約1兆円）—IBM's relatively modest 3.7% revenue miss triggered a historic sell-off. This juxtaposition has led economist Steve Hanke to articulate a thesis gaining traction among analysts: markets are mispricing not stocks' valuations but the earnings that justify those valuations.

The traditional AI bull case rests on two pillars: first, that leading AI companies like Nvidia and Alphabet generate real cash flow unlike profitless dot-com names, and second, that S&P 500 valuations sit at 22x forward earnings, below the 25x-plus threshold typically associated with bubble territory. BCA Research's Peter Berezin has argued for months that today's AI trade is "primarily an earnings bubble rather than a valuation bubble." The danger of such bubbles is their invisibility. Valuation bubbles announce themselves through stretched P/E ratios; earnings bubbles hide within reported profit figures that may be inflated by cyclical capex surges or easy credit. Analysts historically cut profit estimates only after stock declines have already begun, leaving investors without early warning. When earnings bubbles burst, they leave behind real excess capacity—data centers, chip fabs—rather than merely erasing paper gains.

IBM's own aftermath illustrates this detection lag. BofA and UBS trimmed earnings estimates only after the stock had already crashed 25%, with BofA cutting its price target from $330 to $280 and UBS lowering 2026 EPS forecasts while holding its $236 target. The Street's split response—BofA maintaining a Buy rating while HSBC downgraded to Reduce—suggests genuine uncertainty about whether IBM represents a sector-wide crack or an isolated execution miss. The answer will emerge across the rest of earnings season.

No comments yet. Be the first to share your thoughts!