International Business Machines (NYSE: IBM) just delivered its worst single-day drop on record — losing 25.2% on July 14 and erasing roughly $68 billion in market value. The selloff looks puzzling on the surface: global AI spending is booming, with Gartner forecasting a 47% surge to $2.59 trillion in 2026. So why did one of tech’s most recognizable names crater? The answer lies in where those dollars are actually flowing — and IBM was in the wrong spot at the wrong time.
IBM disclosed preliminary Q2 results eight days before its scheduled earnings report, and the numbers told a painful story. Revenue rose just 1% to $17.2 billion — about $660 million below Wall Street consensus. Adjusted earnings of $2.93 per share missed the $3.02 estimate. Software growth decelerated from 11% in Q1 to just 5%. Consulting was flat. Infrastructure fell 7%. CEO Arvind Krishna’s own investor letter used the word “faltered,” and clients had rushed in June to secure supply-constrained servers and memory ahead of price hikes — leaving IBM’s mainframe and deal pipeline short. Several large consulting contracts slipped to future quarters. Oppenheimer, which had modeled 12% software growth, responded by slashing its $350 price target and downgrading the stock entirely.
The real problem: IBM had been re-rated as an AI reinvention story, trading at roughly 26x forward earnings after more than doubling over five years. That premium required consistent software acceleration — and a single quarter of 5% growth cracked the thesis. Hedge fund sentiment was already mixed: 59 funds held IBM at Q1 end (down from 64), while short interest had risen 14.5% to 34.28 million shares. For investors on the sidelines, the smart move is patience. Red Hat grew 11%, generative AI bookings hit $12.5 billion, and the IBM Z mainframe cycle isn’t dead. But until IBM proves those slipped deals actually closed and software growth returns to double digits, the stock is a hold — not a contrarian buy. The 25% drop removed a premium the business hadn’t yet earned back.
