IBM has been forced to reinvent itself many times in the past. After its biggest wipeout on record this past week, it will have to do so again.
Big Blue had been riding high. Yes, there were problems in consulting, as signaled by Accenture's woes, and in software, tipped off by weakness in ServiceNow and its sector peers. But the stock was trading at an all-time high as recently as June 2 as investors looked at the company's near-monopoly in mainframe computing, its quantum computing effort, and its prospects as an artificial-intelligence winner.
They were wrong. IBM stock tumbled 25% this past Tuesday, its worst single-day drop on record, following a rare pre-announcement of its quarterly results. Such a move is highly unusual for the company, which is traditionally disciplined when it comes to financial reporting. The last time IBM pre-announced earnings was in October 2008, in an effort to reassure investors it was on track to meet targets during the global financial crisis.
Investors faced a different reality this time around, as IBM posted second-quarter earnings and revenue that missed Wall Street targets. While there were plenty of problems -- slowing software and consulting sales, a massive reallocation of technology spending by its customers to chips, servers, and other AI needs -- the biggest drag on the company's performance was its infrastructure business. That includes its legacy mainframes -- the massive computers enterprises like banks and credit-card networks rely on to process billions of calculations and transactions in real time.
Big Blue is undoubtedly the dominant force in this space. A 2022 study by Celent, commissioned by IBM, found its Z Mainframe Servers line processed more than half of the world's transactions by value. But that didn't help the division's performance during the second quarter. Infrastructure revenue fell, as expected, but the 7% decline was significantly faster and harder than IBM had anticipated. Not only did fewer companies buy the actual mainframe hardware, they also bought less of the high-margin software required for tasks like banking and credit-card payments.
CEO Arvind Krishna attributed the results to poor execution. "We did not adapt and move quickly enough, and numerous large deals failed to close on the timelines we expected, driving the majority of our shortfall," he wrote in a letter to shareholders.
The 25% drop was massive -- and partly justified, even as it erased nearly $70 billion in market capitalization. "The stock had become a crowded AI infrastructure winner and was trading near all-time highs, so any sign of execution issues was going to get punished," says Dan O'Regan, managing director of equity trading at Mizuho Securities. "That said, a move of this magnitude suggests the market is now pricing in a much more prolonged slowdown than what management has implied."
It didn't help that analysts had set high expectations heading into the print. Morgan Stanley, for one, had predicted upside in infrastructure and software that was already priced into the stock. Oppenheimer, which downgraded the stock on Wednesday, had anticipated "no surprises in business trajectory," making the sudden pre-announcement a true blindside.
Analysts were quick to move to the sidelines following IBM's earnings miss, asserting that Big Blue would have to lean on major acquisitions or close deals that slipped past the quarter's deadline to recover lost ground. Now Oppenheimer is questioning the company's ability to achieve double-digit software revenue growth through 2027. The 5% growth in the latest quarter was sharply below the firm's 12% estimate.
The bigger issue might be whether IBM's infrastructure business itself is being disrupted. Even before Tuesday's plunge, IBM stock had been lagging behind the broader market after stumbling earlier in the year as fears of AI disruption began to take hold. One of the most significant drops occurred in February, when AI start-up Anthropic unveiled a COBOL modernization playbook for its Claude Code tool, claiming it could dramatically streamline updates to the outdated programming language that runs on IBM mainframes. Historically, the immense complexity and cost of migrating off these systems protected IBM's highly profitable mainframe business -- a protective moat AI now threatens to dissolve.
IBM stock closed on Wednesday at 16.54 times 12-month forward earnings, its lowest price/earnings ratio since June 2024. But that says less about where IBM is now than where it was before. As recently as June 2, the stock was trading for more than 25 times, above the S&P 500's 21.52, a premium valuation that might not have been deserved.
"[The earnings warning] does serve as a reminder that IBM's business is large and complex, with profits and cash flows still largely driven by mainframe sales, and thus a premium valuation multiple may not be sustainable," BNP Paribas analyst Stefan Slowinski wrote on Tuesday.
As tempting as it may be to scoop shares up after such a tumble, IBM still has a lot of work to do. In the worst-case scenario, investors fear that IBM's enterprise clients -- massive businesses with sprawling IT setups -- are redirecting their budgets toward AI instead of Big Blue's traditional offerings. At best, the company was simply caught off guard by a sudden capital expenditure shift, as Krishna asserted, and can reclaim that lost ground in coming quarters. As Mizuho's O'Regan sees it, the setup from here depends less on the AI narrative and more on management proving it can consistently execute.
"The market wants proof that this is an execution stumble, not the beginning of a structural slowdown in demand," O'Regan explains. "As a stock, the days of getting the benefit of the doubt are probably over for now."
At least until the next metamorphosis begins to take shape.
