Iron Mountain Is Priced as a Data-Center Play While Data Centers Are 13% of Revenue

Iron Mountain spent most of its existence as one of the least glamorous businesses imaginable: a company that stores boxes of paper documents in warehouses and charges a small monthly fee to keep them safe. It was a slow, sleepy, high-dividend real estate trust — the kind of stock owned for yield, not growth. Then it discovered a new story. By building data centers and a fast-growing IT-asset-recycling business, Iron Mountain re-narrated itself as a participant in the artificial-intelligence and cloud boom, and the market re-rated it accordingly, from a boring storage REIT to a growth-and-AI name at a much richer valuation. Its first quarter of 2026 fed the narrative beautifully: revenue up 22% to $1.94 billion, organic growth of 17% — its highest in more than 25 years — data- center revenue up 47%, raised guidance, and improving leverage. But behind the data-center story sits a number the story does not emphasize. Data centers generated about $255 million of that $1.94 billion in revenue — roughly 13%. The other 87% is still the records storage, the document management, and the asset-recycling businesses that the AI narrative tends to forget. This is a piece about the gap between how Iron Mountain is now priced and what it actually still mostly is, about the capital-hungry, leveraged race it has entered to build the data centers that justify the re-rating, and about the legacy core that the market has decided to look past.

---

Begin with the genuine achievement, because Iron Mountain's transformation is real and impressively executed. A company that many had written off as a melting legacy business delivered 17% organic revenue growth, its best in more than a quarter century, and grew AFFO — the adjusted funds from operations that is the key cash-flow metric for a REIT — 22% to $1.43 per share, and lifted operating cash flow sharply. Its data-center business grew 47% with a 52% adjusted EBITDA margin, its asset-lifecycle-management recycling business grew 92%, it raised full-year revenue, EBITDA, and AFFO guidance across the board, and it improved its net lease-adjusted leverage to 4.8 times, the best since 2014. For a company once seen as a slow-declining paper-storage operation, this is a remarkable reinvention, and management deserves real credit for engineering it. Nothing here disputes that the quarter was strong or that the transformation is genuine.

The question is whether the valuation, which has re-rated to reflect a data-center and AI growth story, matches the business, which is still overwhelmingly something else. So this essay examines the actual revenue mix beneath the narrative, the secular position of the records-storage core, the capital-intensive and leveraged data-center build, the volatility of the recycling business, and what the re-rating assumes.

The data-center tail and the legacy dog

Start with the mix, because it reframes the whole story. The data-center business that drives Iron Mountain's re-rating generated roughly $255 million of revenue in the quarter — out of total revenue of $1.94 billion. That is about 13% of the company. The data center is growing fast, at 47%, and it is genuinely exciting, but it is the tail, not the dog. The other roughly 87% of Iron Mountain's revenue comes from its records-and-information- management business (storing physical documents), its digital and secure-shredding services, and its asset-lifecycle-management recycling operation. The company the market is now pricing as an AI data-center play is, by revenue, still predominantly a document-storage and services business.

This matters because the valuation has re-rated as though the data-center growth defines the company, when in fact it defines about an eighth of it. For the data center to move the whole enterprise's growth rate the way the narrative implies, it would have to become a much larger share of revenue, which takes years of capital-intensive building, and in the meantime the consolidated growth rate is the blend of a fast-growing small segment and a large, slow-growing-to-declining core. The 17% organic growth this quarter is genuinely impressive and reflects strength across segments, but a premium data-center-style multiple applied to a company that is 87% other things is pricing the small fast part as if it were the whole, and the large slow part as if it did not exist. The data center is a real and valuable business; it is also a 13% slice that the valuation has elevated to the entire thesis.

The core that is sticky but secularly challenged

The records-storage core deserves a careful look, because the bull and bear cases both rest on misreading it. The bear caricature is that physical document storage is a melting ice cube — digitization means companies create and keep fewer paper records, so the storage business inevitably shrinks toward zero. That caricature is wrong in the near term: Iron Mountain's storage business is extraordinarily sticky, because once a company puts boxes of records into Iron Mountain's warehouses, it rarely retrieves or destroys them, so the stored volume declines only very slowly while Iron Mountain raises storage prices steadily, producing durable, high-margin, low-churn revenue. The core is far more resilient than the melting-ice-cube story suggests, and it throws off the cash that funds the transformation.

But the bull case can overcorrect into pretending the core is a growth business, which it is not. The secular direction of physical-records storage is, over the long run, down: the world is digitizing, new records are increasingly born digital and never printed, and while the existing stored volume declines slowly, the inflow of new boxes is structurally shrinking. Iron Mountain manages this beautifully through pricing power and through selling digital and other services to the same customers, so the core can grow modestly for years. But it is a mature, low-growth, slowly-secularly-challenged business, and it is the majority of the company. The honest read is that the core is neither melting fast nor growing — it is a durable, high-margin, slow-declining-to-flat annuity, which is a fine thing to own at a storage-REIT multiple but not the engine of a data-center-style growth valuation.

The capital-hungry race the re-rating requires

Here is the tension at the heart of the data-center story: building data centers is enormously capital-intensive, and Iron Mountain is funding that build on a balance sheet that already carries $17.1 billion of debt. To grow the data-center business into a large enough share of revenue to justify the re-rating, Iron Mountain must keep pouring capital into constructing and powering data centers — a business where the leaders, Equinix and Digital Realty, have decades of head start, deeper specialization, and lower costs of capital, and where the hyperscale cloud companies are building their own capacity at a scale Iron Mountain cannot match. Iron Mountain is a credible entrant with real customer relationships and a growing footprint, but it is a challenger in a capital-intensive arms race against far better-capitalized incumbents.

This creates a strategic bind that the strong quarter obscures. As a REIT, Iron Mountain must distribute most of its taxable income as dividends, leaving limited retained cash to self-fund growth; yet the data-center build demands enormous capital. The result is that Iron Mountain must lever its balance sheet or issue equity to fund the very growth that drives its valuation, and it is doing so against competitors who can build more cheaply and at greater scale. Its improved 4.8-times leverage is genuinely encouraging, but the data-center ambition implies years of heavy capital deployment ahead, which will pressure either the leverage or the dividend or both. The re-rating assumes Iron Mountain wins enough of the data-center build to make the segment large and profitable; the path to that runs through a capital-intensive race the company enters as the smaller, more leveraged player against the incumbents who define the category.

The recycling business is volatile, not secular

Part of the growth story is asset-lifecycle management — Iron Mountain's business of decommissioning, recycling, and reselling enterprise IT equipment, which grew 92% in the quarter. This is genuinely fast-growing and fits a real circular-economy trend, and as data centers refresh hardware and enterprises retire equipment, there is a durable long-term opportunity. But it is worth being clear-eyed about its character: asset-lifecycle management is a more cyclical, lower-visibility business than recurring storage rent. Its revenue depends on enterprise IT refresh cycles, on the volume of equipment being decommissioned, and on the resale value of used components and the commodity value of recovered materials — all of which fluctuate.

A 92% growth rate is spectacular, but it is the kind of growth that comes off a small base and that can swing sharply with the IT spending cycle and component prices, rather than the steady compounding of a recurring-revenue annuity. Counting on asset-lifecycle management as a durable, smooth growth engine risks extrapolating a cyclical surge, and a valuation that capitalizes the recent triple-digit growth as a permanent rate is exposed when the cycle turns. The business is a real and welcome addition to Iron Mountain's portfolio, and it diversifies the revenue base; but it is a volatile growth vector, not a stable one, and the premium should account for the difference between a recurring annuity and a cyclical recycling business riding a temporary surge.

The dividend that complicates the growth

There is a structural conflict at the center of Iron Mountain's story that the strong quarter papers over: it is a real estate investment trust, and REITs are built to distribute, not to reinvest. By law and by investor expectation, a REIT pays out the large majority of its taxable income as dividends, which is precisely why income investors owned Iron Mountain in the first place. But a data-center build-out is one of the most capital-intensive endeavors in real estate, requiring billions in construction, power infrastructure, and equipment, and a company that distributes most of its cash flow has little left over to self-fund such a build.

This forces an uncomfortable choice that the data-center ambition makes sharper every year. To fund the build, Iron Mountain must either borrow more — adding to the $17 billion debt load and pressuring the leverage it has worked to improve — or issue new equity, which dilutes existing shareholders, or constrain the dividend that its income- investor base prizes. None of those is free, and all of them are in tension with the growth narrative: the more aggressively Iron Mountain chases the data-center growth that justifies its re-rating, the more it must lean on debt and equity that work against the very REIT economics that defined it. The bull case implicitly assumes the company threads this needle — funding a massive capital program while maintaining its dividend and its improving leverage — but threading it is the central financial challenge of the strategy, and a valuation that prices the data-center success without weighing the capital cost of achieving it is pricing the destination without the toll.

The pricing lever and its limit

The records-storage core's resilience depends heavily on one lever: price. Iron Mountain has grown its storage revenue for years not by storing meaningfully more boxes but by charging more to store the boxes it already has, a disciplined and effective strategy that exploits the stickiness of stored records — customers will not move their archives to a competitor or destroy them to save a few percent on a storage bill they barely think about. That pricing power is genuine and has been a quiet engine of the core's durability.

But price-led growth on a flat-to-declining volume base has the same ceiling it has everywhere: it cannot run forever, because at some point the cumulative increases prompt customers to digitize, consolidate, or finally destroy records they no longer need, accelerating the very volume decline the pricing was offsetting. Iron Mountain has managed this balance skillfully, and the core remains a high-margin annuity, but the lever has limits, and the core's slow secular decline means the company increasingly depends on the data center and the recycling business for growth precisely because the storage core, however well-managed, cannot be the growth engine. The pricing power is real; it is also a maturing tool on a maturing business, and leaning on it harder eventually hastens the decline it has so far held at bay.

What the bulls genuinely get right

In fairness, the bull case is strong and Iron Mountain's transformation is genuinely impressive — the debate is the valuation and the mix, not whether the company has reinvented itself. Iron Mountain delivered 17% organic growth, its best in over 25 years, which is an extraordinary figure for a company once dismissed as a legacy business, and it did so across multiple segments, not just one. Its data-center business is real, fast-growing, and high-margin, with a credible footprint and genuine customer demand from the AI and cloud build-out. Its records-storage core is far stickier and more durable than the bears assume, throwing off reliable, high-margin cash that funds the growth. Its asset-lifecycle-management business taps a real circular-economy trend. It raised guidance across the board and improved its leverage to the best level since 2014, showing financial discipline alongside the growth. And management has executed a genuinely difficult transformation with skill. For investors who believe the data center scales, the core holds, and the diversified growth engines compound, Iron Mountain is a legacy business successfully reinventing itself into a faster-growing, more valuable enterprise, and the re-rating reflects a real change in trajectory.

The honest synthesis is that Iron Mountain is executing a genuine, impressive transformation — and that it has been re-rated to a data-center-and-AI growth valuation while data centers are about 13% of revenue, the majority is a slow, mature, secularly-challenged core, the data-center build is a capital-intensive leveraged race against better- funded incumbents, and a key growth engine is a volatile recycling business. The bull is right that the 17% organic growth, the data-center momentum, the sticky core, and the disciplined execution are all genuine. The skeptic notes that the AI narrative is being applied to a company that is 87% other things, that the data-center ambition demands years of heavy capital against stronger rivals, and that the re-rating prices the transformation as more complete than it is.

What the re-rating assumes

Pull it to the valuation, because that is where the disagreement resolves. Iron Mountain's multiple has expanded to reflect a growth-and-AI story, and that re-rating embeds a set of assumptions: that the data center grows into a much larger share of revenue, that it does so profitably despite the capital intensity and the competition, that the records core holds up rather than declining faster, and that the balance sheet absorbs the build without straining the dividend. Each is plausible given the strong execution, but each is also a real question, and the multiple has priced the optimistic answer to all of them at once.

The asymmetry is the familiar one. If the data center scales as hoped and the core holds, Iron Mountain grows into its re-rated multiple and rewards holders; if the data-center growth slows, or the capital intensity strains the balance sheet, or the core declines faster than expected, the multiple compresses back toward the storage-REIT valuation it carried before the re-rating, and that gap is large. The transformation is genuine, but the valuation has run ahead of the mix, pricing a company that is 13% data center as though the data center were most of the story. The strong quarter supports the optimism; it does not, by itself, close the distance between a 13% segment and a whole-company growth valuation.

The kicker

Iron Mountain has pulled off something genuinely impressive: it took a sleepy paper-storage business and grafted onto it a fast-growing data-center operation and a booming recycling arm, delivered its best organic growth in 25 years, and convinced the market to re-rate it from a yield stock to a growth-and-AI name. The transformation is real and the execution has been excellent. But the re-rating has outrun the mix. Data centers are about an eighth of the company; the other seven-eighths is still the records storage, the services, and the cyclical recycling business the AI narrative overlooks. To grow the data center into the company the valuation imagines, Iron Mountain must win a capital-intensive race, on a leveraged balance sheet, against rivals who build cheaper and bigger. The boxes of paper still pay most of the bills. The market has decided, for now, to price the data center instead — and the distance between the 13% it is and the growth story it has become is exactly the risk in the re-rating.

A company that grew rich and dull storing other people's paper found a way to be exciting again, bolting a data-center business onto the warehouses and letting the market call it an artificial-intelligence stock; the growth is real and the reinvention is genuine, but the data centers are barely an eighth of what the company actually does, the rest is still the slow, sticky, secularly-fading business of keeping boxes safe, and the path to making the exciting part big enough to deserve the new price runs through a capital-hungry race against richer rivals — so the re-rating has run ahead of the revenue, pricing the small eighth that glitters as though it were the whole of the mountain.