IBM Stock Is Down 16% This Year — But the Insiders Buying the Dip Know Something Wall Street Doesn’t

Observing IBM navigate 2026 has an almost poetic quality. In an era shaped by businesses that didn’t exist when IBM was already decades old, a company that has been around for more than a century and once defined what enterprise technology meant for an entire generation of businesses is now fighting for relevance. The stock closed at $246.74 on April 6, down almost 17% year to date and more than $75 below its 52-week high of $324.90 from November of last year. For many investors, that discrepancy reveals everything. However, the gap may be more nuanced than the headline figure indicates.

Because it established the tone for everything that came after, the February moment is worth revisiting. Following Anthropic’s claim that its Claude Code AI could update COBOL, the programming language that still powers IBM mainframes in major banks, insurance companies, and government agencies, shares fell 13.2% in a single session—the biggest one-day decline since 2000. The market responded appropriately because it was the kind of headline that hits IBM like a punch to the stomach. The question of whether the response was appropriate is entirely different. At the time, Brent Thill, a Jefferies analyst, contended that IBM had already begun utilizing generative AI to translate COBOL into Java and that automation and AI were far more important to the larger software revival narrative than mainframe loyalty. The market moved first, followed by inquiries. That usually occurs.

Since then, IBM has been working methodically to reconstruct the story. The business announced in late March that it would collaborate with Arm to create dual-architecture hardware for workloads involving AI and a lot of data. It had previously signed a 10-year R&D contract with ETH Zurich, the Swiss university that produced Einstein, among others, with the goal of fusing the development of quantum algorithms with artificial intelligence. These actions do not indicate a company giving up. There’s real substance behind them, not just press release optimism, but it’s still unclear if they’ll be sufficient to change sentiment in the near future.

With the January release of its most recent earnings report, IBM had a solid starting point. Revenue for the fourth quarter was $19.69 billion, above analyst estimates of $19.23 billion and up 12.2% year over year. EPS exceeded the consensus of $4.33 at $4.52. Infrastructure saw a 21% increase, while software revenue increased by 14%. According to CEO Arvind Krishna, the business is expected to grow its revenue by more than 5% in constant currency in 2026. For a business the size and age of IBM, those figures are respectable. Even so, the stock sold off, which may indicate more about the current sensitivity than the fundamentals.

The insider buying activity that is going on in the background is difficult to ignore. In late February, director Michael Miebach paid $233.33 for 434 shares. In January, director David N. Farr purchased 1,000 shares for $304. Over the previous ninety days, insiders purchased 1,484 shares totaling about $417,000. People purchasing shares with their own funds at these prices are usually not doing so out of anxiety, even though these are not huge transactions in comparison to the company’s size. They most likely have some knowledge, or at the very least, some belief, about what will happen on April 22.

That date is very important. The most obvious test to date to determine whether the software and AI momentum from Q4 was a one-quarter phenomenon or the start of something more sustainable is IBM’s first-quarter earnings release. Enterprise clients have been adopting the Watsonx platform; most recently, Tata Play Fiber in India is utilizing it to combine 25 different data sources into a single AI-ready environment. A crucial obstacle for U.S. government work has been removed with IBM’s FedRAMP authorization for 11 AI and automation products, obtained through its collaboration with Amazon Web Services. These are actual, tangible victories. On April 22, analysts will want to know if these kinds of victories are regularly translating into revenue or if they are still too dispersed to affect the overall figures.
There seems to be an unresolved tension in the way the Street is currently interpreting IBM. Evercore has an Outperform rating and a price target of $345. The current value of Royal Bank of Canada is $361.

Wedbush has an Outperform rating at $340. Then, in March, JPMorgan declared the stock neutral and lowered its target from $317 to $283. UBS changed its rating from Sell to Neutral, but it only set a target of $236, which is basically where the stock is currently trading. With an average target of about $314, the consensus comes to a Moderate Buy, suggesting significant upside. However, significant upside is meaningless if the driving force behind it is not yet apparent.

Fair points are brought up by some critics. Over the last five years, IBM’s revenue has grown at an average annual rate of just 4.1%, which is slow for a technology company during a time when the industry has typically moved quickly. Over the same time period, EPS growth averaged 5.9%, falling short of peer group norms once more. By 2026, the Confluent integration is anticipated to generate about $600 million in dilution, and the debt-to-equity ratio stands at 187%. These are genuine worries, not made-up ones. The consulting industry has been underperforming, despite the fact that it should theoretically benefit from businesses looking for assistance in implementing AI. The pinnacle of IBM’s 2019 acquisition strategy, Red Hat, saw growth cool to 10% in the fourth quarter, which is still respectable but less than the kind of figures that initially supported a $34 billion price tag.

IBM is not experiencing a crisis. It’s a company in transition once more, attempting to persuade a doubtful market that this change is real and long-lasting. That is not the first time that story has been told. The distinction now is that AI is producing actual enterprise spending, quantum computing is no longer theoretical, and IBM benefits from collaborations, government approvals, and research agreements that smaller, quicker rivals are just unable to match on a large scale. April 22 is expected to begin answering whether all of that results in a stock that is worth $314 or $236. The wait is still ongoing in Armonk.