Why IBM just had its worst day in 115 years

In 1997, IBM built a supercomputer that defeated the greatest chess player alive. It was the defining moment of artificial intelligence. Nearly thirty years later, the same company just suffered the worst single-day stock crash in its 115-year history. The cause? Artificial intelligence. The irony writes itself, but the lesson underneath it is worth sitting with.

But before we get to the full story, let's take a quick look at the markets and what matters…

3 Movers in 3 Minutes

1. Intel restructures, market applauds. INTC rose 3.6% on Monday after announcing a new round of layoffs across its data center group, cutting over 5,000 U.S. jobs as part of CEO Lip-Bu Tan's push to become "a more focused and efficient company." The market's read? Discipline ahead of Thursday's Q2 earnings. Intel is up around 150% year to date, but down 30% over the last one month alone.
   
   
2. Oil touched $90, then pulled back. Brent crude briefly breached $91 intraday as the U.S. completed its 10th consecutive night of strikes on Iran. It settled lower after Iran's Foreign Ministry signalled willingness to pursue negotiations through intermediaries. U.S. gasoline prices are back at $4 a gallon nationally.
   
   
3. Teradyne (TER) led the S&P 500. The chip testing equipment maker surged 8.4% on Monday, extending its year-to-date gain to around 60%. Even as software stocks crumble under AI budget shifts, the companies that physically test and measure semiconductors keep winning. It is a telling divergence.

And with that out of the way, let's get to today's big story: the company that invented AI and is now being devoured by it.

The Sip

The day the machine played like a god

On May 11, 1997, Garry Kasparov leaned over a chessboard in New York and resigned. The game had lasted less than an hour. His opponent was not a grandmaster. It was a 1.4-ton IBM supercomputer called Deep Blue, capable of evaluating 200 million chess positions per second. Newsweek called it "The Brain's Last Stand."

It was the first time a reigning world champion had ever lost a match to a computer under tournament conditions.

IBM had proven something extraordinary. Machines could beat humans at the game long considered the ultimate test of intelligence. The company then followed up with Watson, which defeated Jeopardy champions in 2011 using natural language processing. IBM was not merely participating in the AI race. It was leading it.

Fast forward to 2026. And the company that started it all just got crushed by its own creation.

Two crashes, one cause

On February 23 this year, IBM shares plunged 13% in a single session. The trigger was a blog post. Anthropic, the AI startup behind Claude, had announced a new tool that could modernize COBOL, the decades-old programming language that powers banking systems, airline reservations, and government infrastructure worldwide. COBOL runs on IBM mainframes. IBM charges handsomely to maintain and update those systems. An AI tool that automates the work? That was an existential headline.

Then came July 14.

CEO Arvind Krishna issued an unscheduled letter to investors. Preliminary Q2 revenue was $17.2 billion, missing the consensus estimate of $17.86 billion by roughly $700 million. Adjusted earnings came in at $2.93 a share versus expectations of $3.01.

The stock dropped 25% in a single day. It was the worst session in IBM's history, surpassing even Black Monday in 1987, when it fell 23.7%.

Here is the part that matters. The miss was not enormous. Revenue fell roughly 3.7% short of consensus. That is not usually a 25% stock crash. But the explanation was devastating.

The budget cannibals

Krishna told investors that in the final weeks of the quarter, enterprise clients redirected their capital expenditure away from IBM's software and infrastructure products and toward AI hardware purchases: servers, storage, memory chips.

The AI boom has a budget. And that budget is being taken from IBM.

This was not the February crash, where the threat was hypothetical. This time, clients had already moved. They were not buying less technology. They were buying different technology. Money that used to flow into IBM's software division was flowing into the physical infrastructure of the AI buildout instead.

And the damage spread across the sector. Three in five software stocks fell on the same day. Atlassian (TEAM), ServiceNow (NOW), and Adobe (ADBE) all dropped. Meanwhile, cybersecurity stocks rallied. The pattern was clear: clients who deferred software projects still spent on security, because security is harder to postpone.

The correlation between software and semiconductor stocks is now just 0.17. They are decoupling. The AI era is not lifting all tech boats equally. It is lifting the ones that make the inputs to AI and sinking the ones that sell to the same clients footing the bill.

What tomorrow must answer

IBM's full Q2 earnings call lands Wednesday at 5 PM ET. The preliminary numbers are already known. What investors need is context.

How many of the delayed deals have since closed? Is the full-year guidance being revised or reaffirmed? What do Red Hat and watsonx revenue look like underneath the headline miss? Can IBM confirm a free cash flow trajectory that sustains its 3%+ dividend yield?

There is also the matter of a securities fraud investigation that has opened in the wake of the crash, probing whether IBM's pipeline disclosures were misleading.

The stock sits at roughly $212, down 28% for the year. One quarter ago, IBM had beaten estimates, grown software revenue 11%, and looked like a company successfully pivoting toward AI and hybrid cloud. That narrative evaporated in a single afternoon.

The Long Angle

IBM's story is not really about IBM. It is about what happens when the technology you helped pioneer grows large enough to rearrange the economy around you.

Deep Blue proved machines could rival human intelligence. Watson proved they could understand language. But neither prepared IBM for a world in which the commercial explosion of AI would redirect enterprise budgets away from its own products.

There is a pattern here. The company that builds the early version of a transformational technology is rarely the company that profits most from its maturity. Xerox invented the graphical user interface. Nokia made smartphones mainstream. Yahoo built the first great internet portal. And IBM? IBM showed the world what artificial intelligence could do.

The companies capturing the value of AI today are not the ones who won the chess match. They are the ones building the hardware, the models, and the infrastructure that make the next move possible. IBM is learning what every pioneer eventually learns: the revolution does not owe royalties to the company that started it.

The MarketSips Takeaway

What is the biggest lesson you take from IBM's story? Hit reply and tell us.

Until then, sip slowly!

The Market Sip Desk