In the previous letter we talked about IBM stock, and what happened to it in trading last week — it dropped 25% in a single day, meaning a quarter of the company's value simply vanished. In absolute numbers, $67 billion of market value. I've known this market for 30 years from personal experience (and further back from reading), and when I see something like this it's usually one of two things: either a fraud has surfaced at the company, some kind of factual failure, or the market heard news it believes points to a real possibility that the company will collapse, or at least shrink dramatically.
Naturally I was on alert, and I went looking for the data in real time. I started at the beginning — what are we even talking about? The company announced a significant drop in sales (of a particular product). Okay, got it. What's the effect on the company's profit? In terms of earnings per share (that is, how much the company earned, divided by the number of its outstanding shares) they announced they'd earn $2.93 per share in the second quarter. The number on its own is meaningless, unless you compare it to the company's profit in the equivalent quarter (in this case, the second quarter of 2025). And how much did the company earn then? $3 per share. In other words, roughly the same.
It's important to understand — that same profitability justified a share price that had traded, right up to that day, at about $300 a share. So for that you lose $80 a share?
Moving on. Maybe the company announced it won't earn more in the future. I checked that too. The company said it expects growth of 5% for the year ahead (updated to 4-5% when the official reports came out). Well, do you drop 25% instantly over that? That's not it either. The angle is apparently a different one, and it's also broader than the company itself. Keep in mind that ever since AI models entered the workforce en masse, investors are extra-sensitive, afraid that every company is at risk of extinction, that every product will be replaced by internal code written at minimal cost. That's probably the story here, and IBM "took a bullet," as they say on the street, for a fear about a great many old-economy (supposedly) technology companies. Let's get into what IBM's CEO said in the press release:
Large deals simply didn't close in the final weeks of the quarter, because customers took the money earmarked for software and consulting and redirected it to hardware — servers, storage and memory — to grab AI infrastructure before prices rise. Stop and think about it for a second: that same AI wave everyone is excited about, the one that's supposed to be IBM's big promise, is exactly what swallowed its money right now. The wave that is the future is also the problem of the present. This is no longer one company's excuse, it's a macro signal: in the short term, the rush to AI hardware is draining budgets out of the rest of the technology world.
Underneath all of this sits the real question, and we already got close to it in the Nvidia letter (#15): is IBM a growth stock or a value stock? The difference is simple. A growth stock is a company whose business grows fast on its own, and that is what justifies paying a high price for it. A value stock is a stable, profitable company that hands its profit to shareholders but doesn't really grow. IBM sells the market a growth-stock story, "a software and AI company." But look at the number: its software grew 5% in the quarter, which sounds nice, until you discover that almost all of that growth came from a company IBM bought recently for about $11 billion. Strip out the acquisition and look only at the existing business, and software growth was close to zero. That's the difference between organic growth and bought growth: organic growth is when the business you already own gets bigger; bought growth is when you buy another business and glue its revenue onto yours. On the table the revenue looks identical, but the two are not the same thing. One says the engine is strong, the other says you bought a new engine.
Alongside this, and this is where much of the dream lives, IBM keeps investing in quantum computing — a new kind of computer that promises to solve problems no ordinary computer can (we covered this in Letter #11). Just this week, on July 23, it announced the purchase of a quantum research lab called HRL from Boeing and General Motors. It sounds big, and in the narrative it really is. But in actual money it's almost nothing: all of IBM's quantum orders from 2017 through 2024 add up to roughly $1 billion, against revenue of more than $60 billion a year. Anyone buying IBM for the quantum is buying a lottery ticket, not a cash flow. Quantum is a nice option for the next decade, not an engine that pays this year.
So let's lay it out. Here is the picture in the most honest form I can give it: first the pros, then the cons, then one opinion of mine in a single sentence.
IBM's advantages, by the facts
1. A real cash machine. IBM generated about $2.5 billion of "free cash flow" this quarter — the money left in hand after the company paid for everything it needs to operate and develop. That's spare cash, and it's what makes the dividend possible.
2. A steady dividend. The company kept its dividend — the slice of profit returned to shareholders in cash every quarter — even in a bad quarter. A company that keeps paying its shareholders signals confidence in its cash flow.
3. A predictable revenue base. About 80% of IBM's software revenue is "recurring," meaning it comes from contracts and subscriptions that renew on their own. That's revenue that doesn't vanish from one quarter to the next, and it's what keeps the company steady even in a bad week.
4. The AI demand is real. About half of the new deals signed this quarter included an AI component, and the order book in that area is large and still growing. The demand exists, even if the quarter itself was weak.
5. The stumble looks like timing. According to management, roughly a third of the large deals that slipped had already closed in the first weeks of the current quarter. If that holds, the money didn't run away, it was just late.
The disadvantages, in the same spirit
1. Almost all the growth is bought. As we saw, nearly all the software growth came from an acquisition, not from the existing business. A company that has to buy its growth isn't really a growth company.
2. Guidance was cut. IBM itself lowered expectations from "above 5%" to 4-5%, a sign that management too sees the road ahead as less good than it had promised.
3. The growth engine depends on acquisitions. To grow above the market, IBM has to buy one company after another, and that costs money and piles on debt — about $62 billion as of today.
4. Quantum doesn't pay this year. The investment is enormous in the headlines but tiny in actual money; anyone buying for the quantum is paying for a dream, not a near-term profit. This isn't really a disadvantage, it's simply about setting expectations to reality.
5. Management's credibility cracked. A sudden profit warning, after years of promises that IBM is already "a software and AI company," damages trust, and trust is hard to win back.
And EcoMan's opinion, in one sentence: IBM today is a stable value company with a nice dividend dressed up as a growth stock, and whoever buys it for the cash and the dividend may well be satisfied. The quantum is a tremendous value option for the coming years, and at the company's new price, you get it practically for free. And remember — growth achieved through acquisitions is still growth. In the end, money is money, and if IBM knows how to earn it through growth by acquisition, something that sits in its strategic plans going forward, that's good for shareholders.
What does this mean for you? Most of you hold IBM without knowing it. It sits in the Dow index and in the index of the 500 largest U.S. companies, so any fund that tracks those indexes holds a slice of it, and so does your portfolio. Anyone looking for the old world of technology, with an option on the new world of AI and quantum, can find an interesting opportunity here — not a startup, definitely a strategy challenge, but an attractive valuation in the middle of all the current AI frenzy.
The question that will follow this stock from here is whether the deals that slipped really do come back next quarter, as management promises, or whether they're simply gone. The next report, in October, will give the answer. If the deals return, the "stumble" turns out to be timing. If they don't, we'll find the problem ran deeper.
We shall see.
See you in next week's letter, or during the week — if I can't hold myself back until then...