I have been thinking about this gap a lot lately, and about a song.

The Song

The Division Bell came out in March 1994. Pink Floyd, or what remained of Pink Floyd after Roger Waters walked out the door in 1985 (before the album he announced he was leaving, then spent the better part of a decade trying to prevent the others from using the name) put out what turned out to be their final studio album.

It did what Pink Floyd albums do: it sold in staggering numbers, it toured arenas to sold-out crowds, and it divided the faithful roughly down the middle about whether it was a masterpiece or an elaborate exercise in nostalgia.

The debate, frankly, continues.

But there is one track on that album that has settled the argument, at least for a certain kind of listener. Track eight. "Coming Back to Life."

The song is credited solely to David Gilmour, which is notable because most of the Division Bell material was co-written. This one was his entirely. He has since said it was written for and about his wife, Polly Samson, who co-wrote many of the album's other lyrics.

It is a love song, ostensibly, a song about someone who was emotionally absent for a long period and then, through some combination of grace and good fortune, found their way back.

But here's the thing about great songs: they refuse to stay within their original intentions.

The song opens with a guitar introduction that takes roughly two full minutes before Gilmour sings a word. Two minutes! In a commercial pop landscape where the hook is supposed to arrive in the first eight seconds, this is practically an act of aggression.

What Gilmour is doing in those two minutes is not warming up. He is demonstrating something. He is showing you what it sounds like to have nowhere to be. To not be in a hurry. To let the light gather before you speak.

The melody he plays in those two minutes is not particularly complicated by the standards of a guitarist of his caliber. What it is, is certain. Every note lands where it intended to land. There is no searching in it. What you are hearing is a man who has been lost for a very long time and has, finally, found his way back to shore and the finding has made him unhurried in a way that only people who once had nothing can be.

Then he sings.

Where were you when I was burned and broken? While the days slipped by from my window watching. While I was staring straight into the shining sun.

The song proceeds through verses that describe a kind of emotional paralysis, someone who stood still while the world kept moving, who was lost in thought and lost in time. And then, at its pivot point, the line that earns everything: I knew the moment had arrived / for killing the past and coming back to life.

Killing the past. Not releasing it, not letting it go, not moving on in the soft therapeutic sense that phrase has acquired. Killing it. There is intention in that word. There is violence in it, and necessity. Some things don’t simply fade; they require deliberate termination.

Then, after that line, Gilmour does something extraordinary. He plays a guitar solo which, in my entirely subjective estimation, is among the finest three minutes he has ever recorded.

It does not show off. It does not sprint. It rises slowly, the way a tide comes in, you don't notice the water at your ankles, until it's at your knees. By the time you realize you are in the middle of something transcendent, you have been in it for a while already.

This is a song about long stretches of darkness and the slow, patient work of return. It is about the period between being lost and being found, which (and this is the important bit) is never a short period. It is never an afternoon. It is seasons and years and a kind of silence that looks, from the outside, like nothing is happening at all.

IBM investors should relate deeply to this song.

The Stock

Let us talk about patience in its least romantic form: the kind that involves watching a stock you own slide for years, reach levels you didn’t think it would reach, stabilize, plateau, confound a generation of analysts, generate a wave of think pieces about its inevitable irrelevance, and then with the quiet certainty of a Gilmour guitar note hit an all-time high.

On November 12, 2025, International Business Machines Corporation, ticker IBM, closed at $313.19. This was the highest closing price in the company's history. Shortly before that, the stock had touched an intraday high of $324.90, a number that, if you were an IBM shareholder in 1999, would have seemed like a hallucination. A hallucination almost 26 years in the making.

Here is why that matters.

IBM's high watermark during the dot-com bubble was approximately $246 in 1999, before a two-for-one stock split in May of that year. Split-adjusted, that bubble peak was around $123. The stock, having participated enthusiastically in the great speculative frenzy of the late 1990s, a period when the market's collective valuation of technology companies had achieved a kind of feverish self-reference that was more theology than finance and then began its long purgatory.

When the bubble burst, IBM actually held up better than most. While the NASDAQ fell nearly 60% in the year following its March 2000 peak, IBM dropped only 5%. People took this as evidence that the company had been reasonably valued all along. They were, as events would prove, prematurely congratulatory.

IBM shares hit $54 in late 2002. The company had gone from peak to roughly half its peak-adjacent price while the technology indexes were staging what the financial press kept calling a "recovery."

Then something interesting happened: IBM recovered, then peaked again, then fell again, then recovered again. Each cycle told a slightly different story about what IBM was and what it was worth.

The pre-financial crisis high was around $130 in late 2008. During the 2009 sell-off, the stock dropped to $69. By late 2010, it was back above $140, making new highs. It reached its then-all-time high of roughly $213 in early 2013.

And then the real decline began.

This one lasted longer and felt more conclusive than the ones before it. IBM’s legacy business the hardware, the services contracts, the mainframe descendants, the whole enormous enterprise that had made it one of the most important companies in American corporate history was being slowly, visibly eaten by the future.

Cloud computing arrived and IBM was not Amazon Web Services. Mobile arrived and IBM was not Apple. The social layer arrived and IBM was, bafflingly, still IBM.

The stock drifted from $213 down toward $100. By 2018, it had touched $105. By 2020, it was in the same neighborhood. A company that had been founded in 1911, that had survived the Depression and two world wars and the transistor revolution and the personal computer revolution, was now being described in financial press headlines with the word "dinosaur" so frequently that it had practically become the company's unofficial descriptor.1

In 2020, IBM spun off its managed infrastructure services business into a separate company called Kyndryl. This was widely described as either a bold strategic pivot or an admission that the business was unsalvageable, depending on who you asked and what mood they were in. The stock continued to drift.

If you had bought IBM at its pre-split peak in 1999, you spent the next roughly 25 years in the financial equivalent of Gilmour's lost years, staring out the window while the days slipped by, watching other positions grow, hearing your friends talk about their Amazon shares and their Nvidia shares and your position in Big Blue sitting there like a large, vaguely stubborn piece of furniture you can't quite figure out where to put.

The Valuation Problem

Here is the thing that makes this story more interesting than it would be if IBM had simply failed: IBM, for most of the years it was “struggling,” was not actually a terrible business.

It was generating significant free cash flow. It was paying a reliable and growing dividend. It was doing what the financial press would call, with barely concealed contempt, “defensive.”

The problem was not that IBM was losing money. The problem was that the market had decided IBM was yesterday’s story, and the market is extraordinarily good at making that judgment feel permanent when it makes it.

There is a concept in value investing called the "price to intrinsic value gap." The idea is simple: a company's intrinsic value (what a rational acquirer with perfect information would pay for the business) can diverge from its market price for long stretches of time.

Sometimes the price is significantly above intrinsic value, in which case you have a bubble. Sometimes it is below, in which case you theoretically have an opportunity.

The uncomfortable reality is that this gap can persist for so long that the distinction becomes almost academic.

An investor who identified IBM as "cheap" in 2013 at $213 was not wrong about the cash flows. They were wrong about the timeline.

The gap between being right about a business and being rewarded for being right can be measured in decades, and there are very few investors who have the temperament, the mandate, or frankly the lifespan to sit with that comfortably.

Warren Buffett famously held IBM shares, accumulating a position in 2011 and watching it underperform substantially before selling in 2017. He later described the investment as a mistake, not in the sense that IBM was a bad company, but in the sense that he had misjudged how quickly the competitive landscape was shifting around it. If Warren Buffett can be wrong about the timeline on IBM, the rest of us are allowed to feel reasonably humble about our own projections.

As the famous market quote goes, markets can remain irrational longer than investors can remain solvent or, in a less dramatic formulation, longer than investors can remain patient. This is Keynes's observation and it has never become less true. The ghost of every failed value investor haunts every otherwise rational thesis about undervalued companies.

IBM was not a ghost story. But it spent a long time looking like one.

The Return

What changed?

The honest answer is: a lot of things, slowly, and then several things at once.

The seeds of the return were planted years before the harvest was visible. In 2019, IBM made its largest acquisition in history: Red Hat, the open-source enterprise software company, for $34 billion. The market's reaction ranged from cautious to skeptical. Enterprise hybrid cloud was a phrase that made sense as a strategy but required imagination to see as a stock-price catalyst. The imagination took a while to arrive.

The company that was, for decades, synonymous with proprietary systems and vendor lock-in reinvented itself as the company that bets on open-source infrastructure.

This is a bit like if a former oil major became a credible renewable energy company, theoretically possible, deeply counterintuitive, and the kind of transformation that the market prices in slowly and skeptically and then, all at once, believes completely.

Then came the AI wave. Not the first AI wave, or the second — IBM has been claiming AI leadership since before most current market participants were born, and its Watson marketing campaign of the mid-2010s is now cited in business schools as a case study in how to overpromise a technology.

But the AI wave that arrived in 2022 and built through 2024 and 2025 was different in character.

It was infrastructure led. Enterprises didn't just want AI features,

• they needed the plumbing.

• They needed the hybrid cloud architecture to run workloads across multiple environments.

• They needed automation.

• They needed security.

• They needed, in short, exactly the portfolio that IBM had spent five years quietly assembling.

In February 2025, IBM closed its $6.4 billion acquisition of HashiCorp, whose Terraform tool had become the industry standard for infrastructure-as-code. The acquisition gave IBM something it had not had in years: a genuinely dominant position in a growing market segment, a product that infrastructure engineers actually chose to use without being told to by a procurement committee.

The free cash flow story also finally reasserted itself in a way the market could not dismiss. IBM generated $12.7 billion in free cash flow in 2024 and was projecting $13.5 billion for 2025. Software revenue was growing at double digits. The company was, in the words of one analyst who had spent a decade skeptical of it, "becoming the boring infrastructure company that the enterprise market has always needed and never wanted to admit it needed."

And in November 2025, the stock hit an all-time high.

Twenty five years after the dotcom peak. Nearly a decade after the $213 high that was followed by the long slide. Multiple cycles of hope and disappointment and dismissal and revised thesis. The moment arrived, as Gilmour puts it, for killing the past and coming back to life.

The IBM story does not prove that patience always wins. Let me be very clear about this, because the lesson people want to take from a story like this is almost never the right lesson. The right lesson is not "buy beaten down legacy tech companies and wait." Japan's Nikkei index peaked in 1989 and spent the better part of three decades below that peak. Plenty of companies that looked like IBM in 2015 are now genuinely defunct. The narrative of eventual return is survivor bias with extra steps.

What the IBM story actually illustrates is something more uncomfortable: that the gap between a company's value and its price can be real, substantial, and extremely long-lasting, and that even patient investors will test the limits of their patience before the resolution arrives. The people who held IBM from the $213 high in 2013 to the all-time high in 2025 held for twelve years. That is not an afternoon thesis. That is a commitment.

Most of us will not make that commitment for a stock that isn’t working. We have performance pressures and benchmark comparisons and clients who ask why we own something that hasn’t done anything. The institutional structure of modern investing is almost perfectly designed to punish the exact kind of patience that long cycle value investing requires.

This is also why, when a company like IBM finally delivers, the emotional response from long-suffering holders feels disproportionate to what is, technically, just a stock making a new high. It is not just a new high. It is the vindication of a worldview that has been mocked and tested and questioned. It is the moment the song has been building toward. It is the guitar solo after the silence.

The Division Bell and the Long Cycle

I want to return to the song for a moment, because there is something in its structure that maps onto this almost too precisely.

The Division Bell was not a critical success in the way The Dark Side of the Moon was a critical success.

It sold 15 million copies globally and was commercially enormous. But it was released in the post-Nirvana, grunge era, and its measured, atmospheric sound was somewhat out of step with the times. People who were twenty years old in 1994 were often not listening to Pink Floyd. They were listening to other things. The cultural moment had moved.

And yet. The album has undergone a slow, decades long reappraisal that is still happening. “Coming Back to Life” in particular has accrued a kind of reputation that it did not have in 1994, it is now recognized as one of Gilmour’s finest recorded performances, a song that rewards the kind of patient, unhurried listening that the streaming era actively discourages.

Gilmour still plays it live. He played it on his 2024 Luck and Strange tour. Every time he plays it, the audience responds as though hearing it for the first time, because the emotional content of the song does not diminish with repetition. If anything, it deepens.

This is also what happens to good businesses, eventually. The noise around them changes, the narrative shifts, the adjacent stocks get more exciting. But the underlying cash generation, the institutional knowledge, the client relationships, the quiet compounding of reinvested dividends, these do not go away. They are still there, doing their work, below the surface of the quotation screen, in the silence between the price updates.

IBM shareholders collected dividends through every year of the long drought. The company raised its dividend consistently. This is the financial equivalent of the long guitar introduction the part where nothing dramatic is happening, where the casual listener may have already changed the station, but where, if you stay with it, you can hear that something is being built.

A Note on Dividends and the Art of Not Watching

There is a financial research finding that tends to surprise people when they first encounter it. Over long periods of time, dividends account for a remarkably high proportion of total shareholder returns from equity investments historically, something like 40% of total returns from the S&P 500 over the past century has come from dividends, not capital appreciation.

The capital appreciation is what everyone watches. The dividends are what actually do much of the work.

IBM in its “lost decade” from 2013 to 2023 was generating substantial dividends. Shareholders who reinvested those dividends were building position size during the drought years. When the stock finally broke to new highs, those additional shares acquired at lower prices amplified the gain considerably. Benzinga estimated that $1,000 invested in IBM at the peak of the dot-com bubble would be worth roughly $1,804 assuming reinvested dividends, not a spectacular return over 20+ years, but positive, and more than the raw stock price comparison would suggest.

The dividend is the song playing in the background while you wait. It is not the thrilling part. It is, in a meaningful sense, the most important part.

The Dangerous But Irresistible Pastime

Gilmour has a line in the song that I keep returning to: while I pondered on this dangerous but irresistible pastime.

In its biographical context, the line refers to emotional introspection — the risky but compulsive habit of sitting with difficult feelings rather than suppressing them. But the financial analogy is almost too perfect. Investing in companies that the market has given up on is a dangerous but irresistible pastime. You can articulate the thesis. You can walk through the free cash flow math. You can explain why the pessimism is overdone and the dividend is well-covered and the new strategy is gaining traction. And then you watch the stock go nowhere for another year, and another, and you wonder whether you have made a mistake or simply a bet whose resolution is patient.

The danger is real. The irresistibility is also real. Anyone who has ever held a conviction position in an unloved company knows exactly what Gilmour is describing — the compulsive return to the thesis, the checking and rechecking of the original analysis, the strange intimacy that develops between an investor and a company they have stuck with through multiple disappointing quarters.

Most of the time, this pastime ends in disappointment. The company doesn’t come back; it just keeps declining until it is acquired for a modest premium or restructured into something unrecognizable. The value trap snaps shut.

But sometimes — and this is the part the markets price in very slowly, very reluctantly — the waiting ends. The moment arrives. The stock breaks out. The guitar solo starts.

Big Blue, Final Thoughts

IBM hit a new all-time high in late 2025. The company that spent years described as a dinosaur, as yesterday’s story, as the cautionary tale of a technology giant that missed every wave, is now valued at over $200 billion.

The AI infrastructure buildout that everyone assumed would benefit Amazon and Microsoft and Google has, it turns out, also benefited the company whose consulting arm has the deepest relationships with the enterprises actually buying the infrastructure. Who knew. (A few patient analysts knew. They were not popular at conferences.)

The story is not over. New all-time highs are beginnings as much as they are endings. IBM’s $11 billion acquisition of Confluent, closed in early 2026, represents the next chapter of what has become a genuine transformation story. Whether the stock continues to rise or enters another multi-year consolidation period is a question to which no honest person has the answer.

But the journey to this point, the 25-year arc from dot-com peak to new high, through two additional cycles of rise and fall and the long flat purgatory of lost relevance, that journey is worth sitting with. It tells you something true about markets that the quarterly earnings cycle and the momentum-chasing algorithms cannot tell you.

It tells you that time is the most underpriced variable in finance. That the gap between a company’s value and its price, can look, from the inside, like a death sentence, and be, from the outside, merely a phase. That the most important investment decisions are often not what you buy, but whether you have the patience and the judgment to distinguish between something that is genuinely broken and something that is merely slow.

And it tells you that sometimes, after the long silence — after the years of watching from the window while the days slipped by, after the disappointments and the doubters and the embarrassing conference moments where you tried to explain why you still owned it sometimes the moment arrives.

For killing the past and coming back to life.

These pieces are my own observations and should not be construed as investment advice. IBM's past performance does not guarantee its future results, and anyone who tells you otherwise is trying to sell you something. Past performance of music, however, tends to be a more reliable indicator, and The Division Bell is still excellent on a late evening with adequate volume.

If you found this useful, the subscribe button is below. I write about markets, music, and the uncomfortable overlap between the two which, it turns out, is larger than anyone admits

Subscribe now