What happened
IBM lost close to a quarter of its market value in one day. The company issued a rare preliminary warning ahead of its full second quarter results, disclosing adjusted earnings of 2.93 dollars per share on revenue of 17.2 billion dollars. Analysts had expected 3.01 dollars per share on 17.86 billion. The shares fell 23 percent, pacing their worst day since 19 October 1987, the day of the Black Monday crash, when they fell 23.7 percent.
Chief executive Arvind Krishna pinned the shortfall on a sudden shift in how corporate customers were spending. In the final weeks of June, he said, clients redirected their technology budgets away from software and towards hardware — servers, storage and memory chips — rushing to lock in supply before expected price rises. IBM did not see it coming, and said so plainly.
The damage did not stop at IBM. Microsoft, ServiceNow, Salesforce and Intuit all fell between 3 and 5 percent as investors extrapolated the warning across the sector. If IBM customers are raiding the software budget to buy servers, the reasoning went, so are everyone elses.
IBM is scheduled to publish its full second quarter results on 22 July, which gives the market a week to speculate about how deep the problem runs.
Why it matters
This is the clearest evidence yet that the artificial intelligence boom is not lifting all technology companies equally — it is actively cannibalising some of them. Companies have finite technology budgets. Every pound diverted into GPUs, servers and data centre capacity is a pound not spent on the software licences that firms like IBM, Salesforce and ServiceNow depend on.
For years the assumption was that AI would be a rising tide for the whole technology sector. IBM has just demonstrated that it can be a redistribution instead: money flowing out of the software layer and into the silicon underneath it. That is a genuinely different investment thesis, and the sector wide selloff shows investors have noticed.
It also matters because IBM is a bellwether. It sells to large, conservative, slow moving corporate customers — banks, insurers, governments. When those buyers change their spending patterns abruptly, it usually reflects something structural rather than a one off blip.
Finally, a 23 percent fall in a company of this size is a real destruction of value in pension funds and index trackers around the world, most of which hold IBM automatically simply because it is a large US listed company.
Explained simply
Imagine a household that has always spent a fixed sum each month on streaming subscriptions. Suddenly the boiler breaks and the plumber warns that parts are about to get scarce. The family cancels Netflix and buys the boiler. IBM is Netflix in this story.
Technology spending inside a big company works much like a household budget. There is a pot of money, and it has to cover everything. Broadly, it splits into two buckets. One is infrastructure: the physical machines — servers, storage drives, memory chips — that do the computing. The other is software: the programs and licences that run on top of those machines, usually sold as an annual subscription.
For the past two years, the AI arms race has made the infrastructure bucket enormously urgent. Memory chips and servers are in short supply, prices are rising, and waiting lists are long. So corporate technology chiefs did the rational thing: they grabbed the hardware while they still could, and they funded it by squeezing the software bucket.
IBM sits mostly in the software bucket. It also sells hardware, but not the kind that is currently scarce and fashionable. So it got squeezed from both sides — it did not capture the hardware boom, and it lost software revenue to it.
The reason the whole sector fell is simple. If the squeeze is real, it is not an IBM problem, it is a software industry problem. Microsoft, Salesforce and the rest all sell into the same corporate budgets. Investors sold first and asked questions later.
What it means for you
Check what you actually own. If your pension sits in a default global equity fund or a US index tracker, you own IBM, Microsoft, Salesforce and ServiceNow whether you chose them or not. Technology is roughly 30 to 35 percent of a typical global tracker, so a bad day for software is a bad day for your fund, even if the number on your annual statement never mentions it.
If you hold a specialist technology fund — and plenty of UK investors piled into these during the AI enthusiasm of 2024 and 2025 — the concentration risk is much higher. A fund holding 10 percent in a handful of large software names could see a meaningful dent from a single day like this one.
Practically, this is not a reason to sell anything. Single day moves in individual shares are noise for anyone investing over decades. It is, however, a good prompt to check whether your portfolio is as diversified as you believe. If more than a third of your investments sit in US technology, you are running a concentrated bet on one industry, and you should know that you are doing it.
The one group who should pay closer attention are people within a few years of retirement holding heavily technology weighted funds. A 23 percent move in a major holding is exactly the kind of volatility that a portfolio nearing its drawdown date can least afford, and it is worth reviewing your asset mix with an adviser.
The bigger picture
IBM has spent a decade trying to reposition itself from a hardware dinosaur into a software and cloud business. This warning suggests the destination it chose may be less safe than the one it left. The irony is sharp: the AI revolution IBM has publicly championed is the very thing hollowing out its software revenue.
The number to watch is not IBM at all. It is whether Microsoft, Salesforce and the other large software firms confirm the same pattern when they report over the coming weeks. If they do, the market will have to reprice an entire category of businesses that were, until this morning, considered among the most reliable earners in the world.
If they do not — if this turns out to be an IBM specific execution failure dressed up as an industry trend — then today will be remembered as an overreaction and a buying opportunity. The full results on 22 July are the first real test.
