IBM’s Historic Stock Crash Revives a Playbook From 1993

IBM shares crashed 25.21% on July 14, 2026, the worst single trading day in the company’s 115 year history, wiping out roughly $67 billion in market value in one session. The drop beat Black Monday in 1987. Chairman and Chief Executive Arvind Krishna told investors plainly that the company had not moved fast enough.

IBM has faced this exact question before. Louis Gerstner pulled the company back from bankruptcy odds in 1993 with a similarly blunt admission of failure, then rebuilt it around services instead of hardware. Krishna is now betting Red Hat, quantum computing and a security business called Lightwell can do something comparable, even as the mainframe franchise that funded IBM for a century keeps shrinking.

IBM’s Worst Trading Day in 115 Years

The numbers from July 14 are stark. IBM’s preliminary second quarter revenue came in at $17.2 billion, up 1% year over year but short of the roughly $17.9 billion Wall Street had modeled. Operating earnings of $2.93 a share missed the $3.02 analysts expected. Infrastructure revenue, the mainframe and related software business, fell 7%.

Krishna’s own letter to investors laid out the cause. Clients shifted capital spending in the final weeks of June toward servers, storage and memory to secure supply-constrained AI hardware ahead of expected price hikes, leaving less for IBM’s Z mainframes and Transaction Processing software. Large deals slipped into later quarters. Shares opened down 22% and closed at $217.07, a 25.21% single-session loss that topped the 22.96% drop of Black Monday in 1987 for the worst day in IBM’s recorded history.

Peers fell in sympathy. ServiceNow and Salesforce dropped 7% and 5%, while Accenture and Cognizant, both exposed to the same enterprise consulting model, fell 8% and 7%. Microsoft slipped 3%.

These conditions require our teams to execute perfectly, and this quarter we faltered. We did not adapt and move quickly enough, and numerous large deals failed to close on the timelines we expected.

Krishna wrote that in the letter IBM posted the morning of the crash, an unusually candid admission for a sitting chief executive of a Dow component.

The Gerstner Playbook From 1993

IBM has stared down worse. In 1993, more than 90% of its profit still came from mainframe sales that were collapsing as personal computers spread. The company posted an $8 billion loss that year, and shares that had traded near $43 in 1987 could be bought for about $12, according to Wharton’s business school. The board had forced out CEO John Akers and was preparing to split IBM into a dozen or more independent units.

Louis Gerstner, an outsider from RJR Nabisco with no computing background, took over as chairman and chief executive on April 1, 1993. He kept the company together, killed IBM’s habit of selling only bundled, IBM-only products, and pivoted the business toward services, middleware and what became an early bet on the internet. He fired 35,000 of IBM’s 300,000 employees. By the time he stepped down in January 2002, IBM shares had risen ninefold while the S&P 500 gained 154%, Fortune reported in Gerstner’s obituary.

Gerstner died at 83 in December 2025, seven months before IBM posted the worst trading day of its history. Krishna has run IBM since 2020 on a strategy he calls hybrid cloud and AI, built on the 2019 purchase of Red Hat. July 14 was the sharpest test yet of whether that strategy can do for IBM in the 2020s what Gerstner’s services pivot did in the 1990s.

Two Different Flavors of AI Panic

The February COBOL Scare

This was not IBM’s first AI related shock of 2026. On February 23, shares fell about 13% after Anthropic said its Claude Code tool could automate the most labor-intensive parts of modernizing COBOL, the Common Business-Oriented Language, decades-old software that still handles an estimated 95% of ATM transactions in the United States. The steepest single-day loss since October 2000 wiped out roughly $31 billion in market value and dragged the stock down 27% for the month, IBM’s worst monthly slide since at least 1968.

The fear in February was competitive: that AI could directly replace the consulting work IBM sells around legacy code. It looked overdone almost immediately. A month earlier, IBM had reported its highest mainframe revenue in 20 years, and Krishna had credited part of that strength to IBM’s own AI code-conversion tools, first released in 2023.

The July Capital Shift

The July crash worked differently. It was not about AI replacing IBM’s product. It was about AI capturing the marginal capital expenditure dollar instead. Enterprise clients had budget to spend, and they redirected it toward the servers, storage and memory racks that AI data centers needed, rather than toward IBM’s mainframes and transaction software. Krishna had waved off similar supply chain worries just three months earlier, telling investors in April that any impact would be manageable given IBM’s history of running global supply chains. By July, the magnitude had outrun that confidence.

Where the Turnaround Is Already Showing Up

The same earnings letter that detailed the miss also showed a business reallocating itself, in real time, toward the categories IBM is betting will matter next.

• Red Hat, the open-source software unit, grew revenue 11%, its fastest pace since recent acquisitions HashiCorp and Confluent were folded in.
• Distributed Infrastructure, covering Power servers and storage hardware, jumped 37%, its best quarter on record, exiting with a $500 million backlog.
• Lightwell, a $5 billion open-source security initiative backed by more than 20,000 engineers, went live July 8 with early adopters including Bank of America, Goldman Sachs, JPMorgan Chase and Visa.
• Quantum computing drew a commitment of more than $10 billion over five years, spanning research, manufacturing and acquisitions, with IBM targeting a large-scale fault-tolerant quantum computer by 2029.

The quantum push includes a letter of intent with the U.S. Department of Commerce to build Anderon, described as the world’s first pure-play quantum wafer foundry, split between $1 billion in CHIPS Act incentives and $1 billion of IBM’s own cash. That bet lands as Washington pushes its own timeline for the technology, the subject of a separate 2028 federal quantum target that has left quantum-linked stocks split on how fast the payoff arrives.

None of it offset the quarter. Z mainframe and Transaction Processing revenue still fell hard enough to erase the gains elsewhere, even though the z17 program itself was running at roughly 130% of its predecessor’s pace, by IBM’s own account the strongest mainframe launch in company history now simply cycling down.

Wall Street Can’t Agree on the Next 90 Days

Analysts split hard on what July 14 actually means for IBM’s longer-term footing.

• BofA Securities analyst Wamsi Mohan trimmed his price target to $280 from $330 while keeping a Buy rating, writing that he was surprised by the size of the revenue miss, and cut his 2026 revenue estimate to $69.6 billion from $71.4 billion.
• HSBC went further, downgrading IBM to Reduce from Hold with a $191 target, the most bearish call on the stock.
• Patrick Moorhead, chief analyst at Moor Insights and Strategy, argued IBM’s technology is strategic enough that demand will return, saying rising AI-related costs mean enterprises have to cut spending elsewhere to afford it.

The Street’s consensus still sits at Outperform, with an average price target near $300 spanning a range from HSBC’s $191 to a high of $375, a 96% gap between the most bearish and most bullish calls on a single mega-cap stock. Prediction market Polymarket was pricing just a 25.5% chance IBM beats expectations when full results land.

History offers a mixed guide. IBM has fallen 10% or more in a single session eight times since 1968, including this one. In the month following prior drops, the stock averaged a 6.75% gain with a 71.43% win rate, and the edge held out to three months. By six and twelve months, that edge disappeared, with average returns actually trailing IBM’s own unconditional baseline over those stretches.

IBM reports full second-quarter results on July 22, 2026, five weeks after Krishna’s letter first admitted the company had fallen behind.

Frequently Asked Questions

What is COBOL and why does it matter to IBM?

COBOL, short for Common Business-Oriented Language, is a programming language built in the late 1950s that still runs an estimated 95% of ATM transactions in the United States along with core banking, airline and government systems. IBM’s mainframes are the primary hardware running COBOL workloads today, which is why any AI tool aimed at COBOL modernization gets read by investors as a direct threat to IBM.

When does IBM report full second-quarter 2026 results?

IBM’s regular earnings call is scheduled for July 22, 2026, at 5 p.m. Eastern time, with the webcast posted through IBM’s investor relations site. The July 14 figures were explicitly preliminary, and IBM has said final numbers could differ slightly.

What is IBM’s Lightwell initiative?

Lightwell is a $5 billion commitment backed by more than 20,000 engineers, aimed at securing open-source software against vulnerabilities. It reached general availability on July 8, 2026, with early adopters including Bank of America, Goldman Sachs, JPMorgan Chase and Visa among its first enterprise clients.

What is the Anderon quantum foundry?

Anderon is a planned quantum wafer foundry that IBM intends to build under a letter of intent with the U.S. Department of Commerce, funded through $1 billion in CHIPS Act incentives and a matching $1 billion IBM cash contribution. It is part of a broader five-year, $10 billion-plus quantum computing commitment aimed at a fault-tolerant quantum computer by 2029.

Is IBM stock a buy after the crash?

Analyst opinion is split. BofA Securities kept a Buy rating with a trimmed $280 target, while HSBC downgraded shares to Reduce with a $191 target, and the Street’s broader consensus remains Outperform near a $300 average target. The decision likely hinges on whether IBM’s July 22 full report shows the mainframe slowdown was a one-quarter air pocket or something more structural.

Disclaimer: This article is for informational purposes only and is not investment advice. IBM shares carry ordinary market risk, analyst price targets are opinions and not guarantees, and all figures reflect preliminary company data available as of publication.