Iron Mountain Deep Dive: A Hybrid REIT with High Retention, High Leverage, and High Capital Spending

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

• QHow large is its market ceiling? Is it expanding an existing pie, or creating an entirely new market?5/10

  A

  The ceiling is large enough, but IRM is mainly taking share in an existing pie that is rapidly expanding, rather than creating a brand-new market from scratch. Measured by Baillie Gifford's LTGG lens, IRM has a much stronger story than traditional box storage. Even so, its growth runway comes from riding the AI/hyperscale computing wave opened up by others. It is not the pioneer defining the category.

  Start with two completely different ceilings. Traditional physical records and information management outsourcing (Global RIM) is a mature stock market with clear boundaries: customers outsource the physical records required for compliance, audit, and litigation holds. Demand is stable and migration is extremely slow, but the market itself will not expand rapidly. The report itself also acknowledges that paper and tape are in a slow decline as customers shift toward alternative technologies that take up less physical space. This is the cash cow, not the ceiling story. The real growth imagination comes from data centers. According to Iron Mountain's 2025 full-year results release, its data center revenue grew about 30% year over year in 2025, with an adjusted EBITDA margin of about 52%, 63MW of new leasing for the year, an operating portfolio of 488MW, and total developable capacity of about 1.3GW. This is indeed a pie expanding at high speed: a third-party industry source cited by the report, JLL, expects global data center capacity could approach 200GW by 2030, nearly doubling from 2025, driven by hyperscale cloud and AI.

  But a growing pie does not mean IRM is creating a new market. The data center REIT category was defined and occupied long ago by Equinix and Digital Realty. IRM is a fast follower extending laterally into it with a global facilities network and compliance-based customer trust, not the category creator. Its differentiation does not lie in inventing new demand, but in cross-selling its 240,000 customer relationships, network across 61 countries, and compliance credibility into digital and compute infrastructure.

  The honest conclusion is therefore this: the ceiling is set by a compute wave opened by others, and it is high enough. The share IRM can capture depends on whether it can deliver on power access, site permitting, and pre-leasing execution. This is about expanding its own share of an existing pie, not the kind of Baillie Gifford favorite where a company defines an entirely new category and becomes the ceiling itself. This item deserves “large, but not disruptive creation.”

  Jun 10, 2026
• QCan its revenue at least double over the next five years? Will growth mainly be driven by volume, price, or new businesses?5/10

  A

  A revenue doubling over the next five years is possible but not the base case, and growth will come from three forces at once: volume, price, and new businesses. Data centers, through both volume and price, are the key variable that will decide whether the company can actually double. Against Baillie Gifford's hard hurdle for a five-year double, roughly a 15% CAGR, IRM's current pace is slightly below the line. It needs data centers to keep beating expectations to get there.

  Start with the base pace. According to Iron Mountain's 2025 full-year results release, 2025 revenue was about $6.9 billion, up 12.2% year over year; the midpoint of the company's 2026 guidance is about $7.7 billion, implying roughly 12% year-over-year growth. Extending a 12% pace linearly for five years would take revenue to about $12.2 billion, or 1.76 times the current level, close to a double but still short. To truly double, the CAGR needs to reach about 15%, which means the growth businesses must keep outgrowing the company as a whole. The good news is that this acceleration does exist: Iron Mountain's 2026 first-quarter results showed quarterly revenue of $1.9 billion, up 21.6% year over year; data center revenue of $255 million, up 47%; and combined revenue from data centers, digital, and ALM up more than 50% year over year.

  Break the three drivers down:

  The first is price. This is the most firmly validated component. The report cites about 10% organic storage rental revenue growth in 2025, consistent with the official phrasing of “organic storage rental revenue growth 10%,” continued increases in storage revenue per square foot, and 2025 Q4 data center renewal pricing up 9% on a cash basis and 12% on a GAAP basis year over year. Achieving double-digit organic pricing in a traditional business widely assumed to have no pricing power is meaningful.

  The second is volume. New data center MW is the main engine of incremental revenue. In 2025, the company signed 63MW of new leases and had about 1.3GW of developable capacity; the report mentions a pipeline built around roughly 400MW that can be delivered over the next 24 months. Whether this volume can continue to come online depends on power access and development permits, the largest source of uncertainty.

  The third is new businesses. Growth businesses, meaning data centers plus digital plus ALM, have risen from 15% of revenue in 2021 to 28% in 2025. This is structural, not a marginal experiment.

  To be honest, a doubling is not impossible, but it depends heavily on data centers continuing to deliver at a 30%+ pace. If compute demand cools or power becomes a bottleneck, the whole company falls back toward a roughly 12% pace, or about a 1.7 times increase. Baillie Gifford would classify this as “real growth, but a doubling requires the optimistic case to hold.”

  Jun 10, 2026
• QAfter five years, what will take over as the next growth engine? Does this “second curve” exist today?5/10

  A

  The “second curve” already exists today, and the handoff is already under way. It is data centers, supplemented by two smaller auxiliary curves in ALM (asset lifecycle management) and digital services. This is exactly where IRM differs from most transformation stories that have a nice PPT but only a promised second curve. Its next growth engine is not a concept. It is a real business that already contributes nearly 30% of revenue.

  Using Baillie Gifford's framing, “what takes over after five years, and does that curve exist today,” IRM's answer is quite clear. According to Iron Mountain's 2025 full-year results release, the three growth businesses, data centers, digital, and ALM, together accounted for about 28% of total revenue in 2025, compared with only 15% in 2021. Data center revenue grew about 30% year over year in 2025, with an adjusted EBITDA margin of about 52%, far above the traditional business. In other words, the first curve, traditional RIM records management, still generated cash steadily and still represented about 76.7% of revenue in 2025, while the second curve has already stood up and accelerated: in 2026 Q1, data center revenue grew 47% year over year and the three growth businesses combined grew more than 50%.

  The key is this second curve's renewability and runway length. Data centers are tied to the long-cycle capex of AI and hyperscale cloud. The industry forecast cited by the report expects global capacity to nearly double by 2030. IRM has about 1.3GW of developable capacity, the report mentions a pipeline of roughly 400MW under construction or deliverable, and a meaningful share of the capacity under construction has already been pre-leased. The report's figure is 68.3% pre-leased under construction, which provides revenue visibility over the next 24 months. This means the second curve is not a one- or two-year pulse, but a multi-year extension with structural reserves.

  Two honest reservations still matter. First, this second curve is essentially about riding someone else's wave rather than opening its own. In data centers, IRM is a follower, not the purest or strongest leader; EQIX/DLR are purer. Second, it is highly capital intensive. The faster the second curve grows, the more growth capex it consumes, as discussed in the unit economics and cash flow questions. The more precise conclusion is therefore that the second curve clearly exists and is taking over, but it changes the company from an asset-light cash cow into a high-capex growth platform. The price of the handoff is an ongoing test of cash flow and leverage. As for a third curve, whether ALM/digital can independently grow into large businesses remains to be seen because their scale is still small. Overall, this item is a rare strength for IRM: the second curve is not a promise, it is already running.

  Jun 10, 2026
• QWhat is its core competitive advantage? Will this moat widen or narrow over the next three to five years?6/10

  A

  The core moat is a combination of compliance trust, high switching costs, network density, and already pre-leased power/site resources. Over the next three to five years, it is likely to move from “stable” to “slightly wider.” But the source of widening, data centers, sits in a highly competitive red ocean, so a widening moat does not automatically mean higher shareholder returns. Through a Baillie Gifford lens, this is a real and measurable moat, but one with a ceiling on its width. It is not the kind of self-accelerating, winner-takes-most network-effect moat.

  Start with the traditional records management side. Its moat is real and can be validated by data. According to Iron Mountain's 2025 full-year results release, organic storage rental revenue grew about 10% in 2025, storage revenue per square foot continued to rise, and the report disclosed an overall customer retention rate of about 98% and that about 95% of the Fortune 1000 are customers. Consistently achieving double-digit organic pricing in a mature business widely assumed to have no pricing power is itself evidence of pricing power and switching costs. Moving a company's huge body of compliance-related physical records creates direct relocation costs, chain-of-custody and compliance risks, operational disruption, and audit risk. This switching cost is the thickest part of IRM's moat.

  Now consider the data center side, which is what could make the moat wider. The barriers here are not technological; they are soft resources: power access, location, development permits, and already pre-leased capacity. The report shows that IRM has about 1.3GW of developable capacity, about 68.3% of capacity under construction already pre-leased, and pre-leased customers that are mostly hyperscale customers on 10–15 year long-term contracts. In today's power-constrained environment, these resources are genuinely scarce and hard to replicate quickly, creating a real barrier against followers. Growth businesses rising from 15% of revenue in 2021 to 28% in 2025 is a quantified sign of moat extension.

  But Baillie Gifford-style honesty requires deductions:

  First, network effects are weak. One additional storage customer does not make the service exponentially more valuable to existing customers. This is not a social platform or exchange-style moat that reinforces itself, so the width has a ceiling.

  Second, the battlefield where it is widening is not friendly. The data center industry is highly capital intensive and fiercely competitive. The report cites a North American vacancy rate of only 1.4% at the end of 2025 and supply growth of 36% year over year, meaning supply is increasing rapidly. IRM is not the purest or strongest leader in this track, and the purity of its moat is lower than EQIX's interconnection ecosystem.

  Third, moat widening can decouple from returns. The report repeatedly emphasizes that shareholder returns will be determined not by revenue growth, but by whether project-level ROIC on large-scale data center investment exceeds the cost of capital. That still needs time to be proven.

  Overall judgment: the direction of the moat is upward, slightly widening, and the base is solid, with trust plus switching costs validated by double-digit pricing. But this is a bounded moat, not a self-accelerating moat, and the price of widening is high capex and high leverage. This item deserves a neutral-to-positive score, but it is not the kind of company Baillie Gifford loves most, where the moat widens relentlessly and almost intimidatingly.

  Jun 10, 2026
• QIf its core business is disrupted, does it have the DNA to reinvent itself? How does it treat mistakes and bad news?5/10

  A

  IRM has already proved that it has this DNA through one textbook reinvention: it proactively transformed an old-line records management company facing paper decline into a hybrid platform of data centers, digital services, and ALM. That itself is an early response to the risk that its core business could be disrupted. But the reinvention has been a top-down, gradual strategic pivot, not a founder-led radical rebuild, and it has not yet gone through a true life-or-death stress test. On “how it treats mistakes and bad news,” public evidence suggests management is fairly candid and does not avoid core risks. That is a positive.

  Start with the implicit premise of “reinvention DNA.” IRM's core traditional business faces the kind of slow disruption Baillie Gifford is most wary of. The report cites the company's own acknowledgment that customer storage methods are changing, and that paper and tape are migrating toward alternative technologies that occupy less physical space, which will erode storage revenue. Faced with this gradual disruption, IRM did not simply wait for decline. Through Project Matterhorn, it pushed the company toward data centers and digital businesses. The results are quantifiable: according to Iron Mountain's 2025 full-year results release, growth businesses rose from 15% of revenue in 2021 to 28% in 2025, and data centers grew about 30% in 2025. Turning a “box storage” company into a second-curve business shows that the organization does have the ability to migrate customer trust and network assets into new forms. That is real and already delivered reinvention DNA, stronger than most old companies that talk about transformation but show nothing in the financials.

  But Baillie Gifford-style honest limits are needed:

  First, this was a proactive managerial pivot, not a desperate fightback after disruption. IRM began positioning ahead while the traditional business was still steadily generating cash and had not collapsed. It was a controlled turn, not yet the kind of true resilience test where a core business suddenly loses blood and the company must rebuild. The DNA has been validated once, but not under extreme stress.

  Second, it is not founder-driven reinvention. CEO Bill Meaney has led the transformation for years and provided strategic continuity, but the report clearly states that the company is “not founder-led,” and management ownership is not high. It lacks the founder-like drive of someone who treats the company as their own child and is willing to tear it down and rebuild it.

  On “how it treats mistakes and bad news,” IRM looks reasonably good. The report notes that management has not avoided core risks in disclosures, explicitly listing paper/tape migration, data center power and service interruptions, capex funding risk, and REIT distribution constraints. In 2025, the company also plainly disclosed about $105.6 million of foreign exchange transaction losses and about $195.9 million of Matterhorn-related costs, without trying to prettify GAAP profit. The only reservation is that IRM makes heavy use of non-GAAP measures such as Adjusted EBITDA/FFO/AFFO, which can objectively pull market attention away from GAAP cash capex pressure. That is not manipulation, but investors need to stay alert.

  Overall: the reinvention DNA has been proved by one successful transformation, but not by a life-or-death crisis. Its attitude toward bad news is candid. This is a neutral-to-positive side of IRM.

  Jun 10, 2026
• QDoes management, especially the founder, have a long-term view and interests deeply tied to the company? Is it willing to sacrifice current profits for results five to ten years out?4/10

  A

  Management is strategically long-term oriented, and the compensation structure does a decent job linking pay to shareholder interests, but this is not a founder-style, high-ownership team whose interests are deeply tied to the company. The degree of alignment is only moderate. As for whether it is willing to sacrifice current profit for results five to ten years out, the answer is that it is already doing so in practice: IRM is using huge capital spending that consumes GAAP free cash flow in exchange for future growth. But this sacrifice is a structural choice in a REIT transformation, not an expression of a founder's personal will.

  Start with alignment, the dimension Baillie Gifford cares about most. The evidence is sober. The report discloses that as of March 4, 2026, CEO Bill Meaney beneficially owned about 394,000 shares of common stock and about 2.62 million vested options or options vesting within 60 days, representing about 1.0% under the proxy statement methodology; directors and executive officers together owned about 1.7%. This means management is not without equity at risk, but it is far from the founder-heavy ownership Baillie Gifford loves, where ownership is often in the double-digit percentages and personal net worth is locked to the company. The report explicitly notes that IRM is “not founder-led.” So on “interests deeply tied to the company,” IRM is moderate, not strong.

  The compensation structure is a relative positive. The report cites the 2026 proxy statement: 93% of the CEO's target total compensation was at-risk, meaning linked to performance/share price, and the CEO's 2025 base salary remained unchanged at $1.2 million. A high at-risk share and no voluntary increase in base pay suggest the board is serious about making management accountable for long-term outcomes. This part is better aligned than the ownership piece.

  On “long-term view + willingness to sacrifice the present for the future,” IRM actually has the hardest evidence. According to Iron Mountain's 2025 full-year results release and the report's figures, 2025 operating cash flow was about $1.34 billion, but cash capex was about $2.272 billion, including growth capex of about $2.068 billion, resulting in clearly negative GAAP free cash flow. This is essentially sacrificing currently distributable cash to bet on future high-return projects. Management's publicly stated capital allocation priorities are consistent: grow the dividend, invest in high-return growth projects, and maintain the balance sheet. It has also raised the dividend for the fourth consecutive year and underwritten growth through pre-leased developments with 10–15 year long-term contracts. Strategy is continuous and not short-term oriented.

  But Baillie Gifford's honesty requires separating whether “sacrificing current profit” is a virtue or a risk. It depends on project returns. The report repeatedly emphasizes that the real test is not whether the framework sounds good, but whether this growth capex ultimately generates returns above the cost of capital. That still needs ongoing tracking. If a high-ownership founder team were spending this aggressively, the market would be more comfortable. With IRM, a professional management team with modest ownership, compensation linkage, and rational strategy is spending heavily, so investors must monitor project-level ROIC themselves.

  Overall: long-term view and capital discipline pass; sacrificing today for the future is already happening; but on Baillie Gifford's core preference for founder-level deep alignment, IRM is only moderate. This item is neutral and should not be elevated.

  Jun 10, 2026
• QIf it disappeared tomorrow, how much would customers miss it? Is its growth sustainable and not dependent on harming society or fighting regulators?6/10

  A

  If IRM disappeared tomorrow, its core storage customers would miss it considerably, because migrating and rebuilding compliance-related physical records could hardly be done overnight. Its growth model is healthy, because it sells compliance, security, and compute infrastructure, and does not depend on harming society or opposing regulators. IRM performs solidly on this item: indispensability is moderately strong, especially in the traditional business, and social/regulatory sustainability is good.

  Start with the implicit premise of indispensability. For traditional records management customers, IRM's irreplaceability is high. According to Iron Mountain's 2025 full-year results release and the report's figures, the company has an overall customer retention rate of about 98%, about 95% of the Fortune 1000 are customers, and organic storage rental revenue still grew about 10% in 2025. A retention rate of about 98% is the most direct evidence that customers cannot easily leave. Companies entrust IRM with huge volumes of physical records required for compliance, audits, and litigation holds. If IRM disappeared, customers would face relocation, rebuilding compliance chain-of-custody, operational disruption, and audit risk. This switching pain is the root of the high retention. For data center customers, indispensability is also not weak: hyperscale customers sign 10–15 year long-term contracts and depend on the power and site resources IRM has already secured, making short-term relocation difficult.

  Still, the layers need to be separated honestly: the traditional business would be missed more than the data center business. In data centers, stronger competitors such as EQIX and DLR exist, and an individual data center customer theoretically has more alternatives, although long contracts and locked-in power provide stickiness. In compliance-related physical records, IRM's network density and compliance reputation have almost no publicly comparable counterpart. So the customers who would miss it most tomorrow are first the 240,000 storage customers, especially regulated industries, government, and healthcare customers.

  On whether the growth model is sustainable and harms society or regulation, this is one of IRM's cleaner sides, sharply different from business models that grow through data abuse, regulatory arbitrage, or user harm. IRM's growth comes from three positive activities: helping companies meet compliance and information-security requirements, which regulation requires and society needs; providing compute infrastructure for the digital economy; and using ALM for asset recycling and data destruction, which itself has environmental/data-security attributes. It is not expanding by inducing addiction, evading regulation, or harming user privacy. The regulatory constraint mentioned in the report, the REIT distribution requirement, is a neutral compliance framework, not a red-line risk. The power/supply-chain constraints facing data centers are operating bottlenecks, not a social-harm controversy.

  The one neutral risk worth mentioning is that data centers are major electricity users and may face social and regulatory scrutiny over energy consumption and grid load over the long term. But this is an issue shared by the whole industry, not a unique “harms society” label for IRM, and the company itself already treats power cost as a core risk to manage.

  Overall: indispensability is moderately strong, especially in the traditional business; social/regulatory sustainability is good; and the growth model is healthy. This is a defensible side of IRM and deserves a neutral-to-positive mark.

  Jun 10, 2026
• QWhat are the unit economics of this business, including gross margin and incremental returns? Do they improve or deteriorate with scale? Where does the money earned get spent?5/10

  A

  The unit economics are a tale of two extremes: at the business-unit level, gross margin/EBITDA margins are very attractive and improve with scale and business mix upgrade; at the company-wide level, most of the money earned must immediately be reinvested into data center growth capex, making GAAP free cash flow negative. So on “incremental returns,” the question Baillie Gifford cares about most, the current answer is: excellent segment margins, but company-wide cash conversion has not yet been delivered. This item must be answered honestly in two layers, without looking only at the glossy side.

  Start with the good layer, business-unit economics. According to Iron Mountain's 2025 full-year results release, the company's overall adjusted EBITDA was about $2.574 billion in 2025, with a 37.3% margin, up about 15.1% year over year. The data center adjusted EBITDA margin was about 52%, and Global RIM about 45%. Both numbers are strong: a 45% EBITDA margin in traditional records management shows a solid cash-cow base, while a 52% data center margin is above the traditional business, implying that as the mix shifts toward data centers, the blended margin has upside. On a segment basis, the idea that unit economics improve with scale is valid. Add continued increases in storage revenue per square foot and about 10% organic storage rental revenue growth in 2025, and the pricing leverage is real.

  Now look at where the money earned gets spent. This is the key point. The report discloses total cash-paid capex of about $2.272 billion in 2025, of which recurring maintenance capex was only about $147.3 million and growth capex was as high as about $2.068 billion. In other words, the old business needs very little capital to sustain itself, with recurring capex at about 2% of revenue, while the overwhelming majority of capital is going into new data center construction. This is exactly the “where does incremental capital go” question that Baillie Gifford would scrutinize closely: the money is not being used for buybacks or large distributions. It is almost entirely being bet on future high-return data center projects.

  But whether incremental returns are actually good cannot yet be settled, and this is the hard weakness that must be acknowledged. According to Iron Mountain's 2025 10-K, 2025 operating cash flow was about $1.34 billion, while capex was about $2.272 billion. On an “OCF minus total capex” basis, GAAP free cash flow was about -$900 million to -$1 billion. The report identifies the core issue: some investors look only at AFFO and say cash flow is strong; others look only at GAAP FCF and say the company does not make money. The more accurate view is that the old business cash cow is stable, while the new business consumes a lot of capital. The project-level ROIC and post-commissioning cash payback curve that will ultimately determine the quality of unit economics are explicitly listed by the report as open issues not yet fully disclosed and requiring ongoing tracking.

  One more point matters: IRM has negative book equity, about -$938 million in 2026 Q1 under the report's figures, which makes ROE/P/B and other traditional return-on-capital metrics completely distorted and unhelpful. This further shows that IRM must be assessed by how much incremental EBITDA/AFFO each new dollar of capital produces, not by book returns.

  Overall: segment gross margins and EBITDA margins are excellent and improve with business mix, while pricing power is real. But company-wide incremental returns are still in the investment phase, cash conversion is negative, and project-level returns remain to be validated. This is IRM's classic combination of quality and cash consumption, so the item is neutral and should not be elevated just because segment margins look attractive.

  Jun 10, 2026
• QWhat conditions would need to hold simultaneously for it to rise fivefold over ten years? Are those conditions realistic? What expectations are embedded in today's share price?3/10

  A

  For IRM to rise fivefold over ten years from the current roughly $124–126 to about $620–630, a string of quite optimistic conditions would need to hold at the same time. At today's price, the market has already priced in the success of the data center transformation in advance, leaving new buyers with a thin margin of safety. This is IRM's weakest area. The issue is not that the business is poor, but that the price already discounts most of the good outcomes.

  First anchor what today's price implies. According to stockanalysis data, IRM's share price is about $124, market cap about $36.9 billion, enterprise value about $56.5 billion, EV/EBITDA about 23, EV/Revenue about 7.8, trailing P/E as high as about 137, and dividend yield about 2.79%. Against actual 2025 AFFO per share of $5.17, the current price implies about 24 times AFFO. What does that mean? GAAP net income of only about $152 million is supporting a market cap of about $36.9 billion and a PE near 137, meaning market pricing is built entirely on expectations that future AFFO and EBITDA will keep growing quickly and that high-return data center investment will be delivered, rather than on current earnings. In other words, today's share price already embeds the assumption that the transformation will probably succeed. The good news has been bought.

  Now list the conditions required for a fivefold gain over ten years, which is the implicit premise:

  First, data centers must sustain 30%+ growth for a long time and deliver high project returns. About 1.3GW of developable capacity must come online on schedule, renewal pricing must continue, with the 2025 Q4 renewal increase of 9%/12% sustained for years, and project-level ROIC must truly exceed the cost of capital. The report clearly notes that the last point has not yet been fully disclosed and is the largest open question.

  Second, the slow decline in the traditional business must not accelerate. About 98% retention and double-digit organic pricing must hold, and paper substitution cannot suddenly speed up.

  Third, leverage must be able to decline while the company still grows. Current net leverage is about 4.8–4.9x, ratings are BB-/Ba3, and interest coverage is only about 3.2x. Over ten years, the company must sustain huge growth capex, with 2025 capex of about $2.272 billion and negative GAAP FCF, while keeping leverage within the 4.5–5.5x corridor and avoiding deterioration in funding costs.

  Fourth, valuation multiples must not compress materially. Even if all of the above happens, if the market compresses the current roughly 23 times EV/EBITDA back toward a more ordinary REIT multiple, the stock could still go nowhere or fall.

  Each condition is not absurd in isolation, but the probability that all of them hold for ten years without a stumble is not high. This is exactly the kind of stacked upside required for Baillie Gifford blue-sky imagination, but IRM's upside is heavily compressed by a high starting valuation. The report's three-scenario DCF intrinsic value range, conservative about 64, neutral about 100, and optimistic about 145 dollars, shows that the current price of about $124–126 is already above the neutral case and close to the lower end of the optimistic case. The report's reasonable buy range is 85–100 dollars, and the current price is well above that. The report therefore judges that the margin of safety is not obvious, and that above 125 dollars the stock looks more like a high-expectation price than a value price.

  Honest conclusion: a fivefold gain over ten years is not impossible, but it requires a long list of optimistic conditions to all hold, and the purchase price itself must not be overpaying. Today's price has already prepaid for transformation success. This item is clearly negative: a good business, but an expensive price, high embedded expectations, and a thin margin of safety.

  Jun 10, 2026
• QWhy has the market not realized all this yet? Is it because investors do not understand it, look down on it, or cannot look far enough ahead? What could become the narrative inflection point?3/10

  A

  The premise of this question needs correction: the market has in fact already recognized IRM's transformation story. The share price rising over the past year from a low of about $78 to about $124–126, with the one-year range reaching as high as $134, is the result of full, even excessive, market pricing, not a sign that the market has not understood it. So IRM is not the kind of company Baillie Gifford likes most: a severely misunderstood, overlooked stock waiting to be discovered. The real perception gap is not that the market has missed the upside, but that the market may be underestimating the risks from high capital spending and leverage.

  Start by correcting the premise with data. According to stockanalysis data, IRM's current price is about $124, trailing P/E about 137, and EV/EBITDA about 23. This is a valuation that fully incorporates data center growth. It is certainly not an obscure stock that investors do not understand or look down on. The report itself repeatedly emphasizes that the market has already awarded it a high transformation premium, and that buying now is more like paying a price that fully reflects transformation success. So for Baillie Gifford's classic triad, “do not understand / look down on / cannot look far enough,” the answer for IRM is: the market basically understands it, is willing to pay a premium, and may even be looking at it too optimistically.

  Is there still a perception gap? Yes, but its direction is the opposite of Baillie Gifford's usual “upside is underestimated” setup. Here, the more likely gap is that risks are underestimated:

  First, REIT accounting complexity masks cash pressure. The market is used to watching AFFO, $5.17 per share in 2025 and up about 14% year over year, and adjusted EBITDA, which according to Iron Mountain's 2025 full-year results release was about $2.574 billion with a 37.3% margin. It can easily overlook that GAAP free cash flow is negative because of about $2.272 billion of capex, and that GAAP net income was only about $152 million. Optimists see attractive non-GAAP growth; cautious investors see huge capital consumption. This difference in accounting lenses is the real perception gap.

  Second, project-level returns have been priced before being validated. The market assumes growth capex will produce high returns, but the report clearly notes that project-level ROIC and post-commissioning cash payback curves have not yet been fully disclosed. The market is assuming returns are good, not confirming that returns are good.

  As for what could become the narrative inflection point, the implicit premise, it can cut both ways:

  Upside inflection: data centers keep exceeding expectations. In 2026 Q1, data center revenue already grew 47% year over year, the three growth businesses combined grew more than 50%, and AFFO per share rose 22% to $1.43. If the roughly 400MW pipeline accelerates and the company discloses strong project-level ROIC for the first time, it would reinforce the “transformation success” narrative and push valuation higher. Continued upward revisions to 2026 guidance, revenue of about $7.7 billion and AFFO of about $172 million, both around 12% growth, would also be positive catalysts.

  Downside inflection: any crack appears. A decline in data center pre-leasing, commissioning delays, slower renewal pricing, faster decline in the traditional business, leverage above 5.5x or worsening interest coverage, or higher financing costs could all push the market to compress the current roughly 23 times EV/EBITDA back toward a more ordinary REIT multiple, triggering a narrative downgrade. The worst case warned by the report is valuation compression causing a 35%–50% share-price drawdown and permanent capital loss for new buyers.

  Honest conclusion: IRM is not an overlooked stock where the market has failed to recognize value. It is a stock where the market has already priced the story fully, upside is compressed by a high valuation, and risks may be underestimated. Baillie Gifford would judge this item as negative-leaning: the direction of the perception gap is unfavorable to buyers at today's price, and the probability of a downside narrative inflection cannot be ignored. This is also consistent with the report's Watch rating.

  Jun 10, 2026