Why IBM Just Suffered Its Worst Day in Nearly 60 Years

Anatomy of a Crash: Why IBM’s Stock Plunged 25% in a Single Day

On July 14, 2025, IBM CEO Arvind Krishna warned investors that the company’s second-quarter results would be “worse than expected.” When full figures landed on July 22, the damage was stark: a collapse in sales of its iconic mainframe computers, combined with soaring costs from the AI arms race, had torn a hole in the numbers.

The stock plummeted 25% on the warning day — the biggest one-day percentage drop since the 1987 Black Monday crash, though this was a company-specific rout rather than a market meltdown. The culprit was a sudden, unanticipated shift in how large enterprise customers spend their IT budgets. Instead of refreshing their IBM mainframes — the workhorses that process 70% of the world’s transaction value — they diverted spending toward servers, storage and memory, the physical infrastructure that underpins the generative AI boom. IBM’s chief financial officer, James Kavanaugh, said the combined hit from mainframe hardware and associated software alone shaved more than five percentage points off quarterly growth, against an expectation of just one to two points.

The shock reshaped IBM’s growth outlook. Krishna cut the company’s full-year 2026 revenue growth forecast from above 5% to a range of 4%–5%, though he stressed that the deferred deals were “postponement, not destruction” and that a third had already closed early in the current quarter. The dividend was left untouched, to be paid in August, but the stock closed at $205.77, below even its worst-day close.

The Strategic Crossroads: Mainframes, AI and the Future of IBM

Why Mainframe Sales Collapsed – and Why It Surprised IBM

The drop was not merely cyclical; it reflected a structural reprioritization of IT spending. With AI projects demanding more powerful, distributed computing, corporations are buying fewer monolithic mainframes and instead building out clusters of x86 servers and Nvidia GPUs. IBM’s mainframe business, still a profit pillar, is ill-positioned for this shift. The magnitude of the miss suggests that IBM’s own sales and supply-chain intelligence failed to detect the change until the final weeks of the quarter, undermining internal forecasting credibility.

Where IBM Stands in the AI Infrastructure Race

While Nvidia and cloud giants capture the AI infrastructure spending flood, IBM’s AI play is deliberately different. Its watsonx platform does not compete on the size of large language models; instead, it orchestrates AI across multiple systems and vendors, including models from OpenAI and Anthropic. That is a sensible niche for regulated industries, but it meant IBM missed the immediate hardware spending wave. The company’s infrastructure portfolio is anchored by mainframes, not the graphics processors and memory that customers are chasing now.

The Software Lifeline: Red Hat and watsonx

The bright spot is software, which grew 5% and was led by Red Hat’s 11% surge. At an 82.8% gross margin, software is now the profit engine, generating roughly two-thirds of IBM’s profit on 45% of revenue. watsonx Orchestrate integrates with over 80 enterprise systems, and IBM has stopped pushing only its own models, embracing a multi-vendor approach that positions it as the trusted “layer of governance” for AI. The question is whether this high-margin segment can grow fast enough to offset the persistent drag from hardware.

Krishna’s Admission and the Credibility Gap

Krishna’s admission that IBM “vacillou” — it faltered — was highly unusual for a CEO who normally projects unshakeable confidence. The market interpreted it as a crack in execution credibility. Even after the earnings call, in which Krishna explained that deals were merely deferred, the stock failed to recover, signaling that investors now need proof, not promises.

What IBM Leadership and Investors Should Do Next

• IBM management must urgently demonstrate that the mainframe sales fall was a timing issue, not permanent demand destruction. Closing the deferred deals in Q3 and publicly disclosing that metric is essential to restore faith.
• Krishna should articulate a clearer AI revenue model that connects watsonx’s governance layer to top-line growth. Investors need to see how orchestrating third-party models translates into sustainable high-margin income, not just a defensive positioning.
• Investors should focus on the split between infrastructure and software revenue growth in the next quarter. A persistent mainframe decline without compensatory acceleration in Red Hat and watsonx would justify a lower valuation.
• The company must guard against further competitive loss in the server and AI hardware space. Even a non-participation strategy needs to ensure that its core mainframe clients do not shift workloads permanently to cloud-based AI systems.
• The maintained dividend is a significant signal of balance-sheet confidence. However, if the growth reset continues, a dividend cut could become a risk. All stakeholders should monitor the cash-flow coverage ratio next quarter.