Iron Mountain

Phase A — Understand the business

Lens 1 · Company Overview

Iron Mountain is a $6.9B-revenue, S&P-500 REIT that began in 1951 storing business records in a depleted iron-ore mine near Hudson, NY, and is now a global information-management company trusted by 240,000+ customers in 61 countries, including ~95% of the Fortune 1000. It has operated as a REIT since its 2014 tax-year conversion.

The business is four engines bolted to one chain-of-custody reputation:

1. Records & Information Management (RIM) — the legacy cash cow. Physical document storage (740M+ cubic feet of volume), data-tape vaulting, secure shredding, media/archive services, and consumer storage. Recurring storage-rental contracts run 1–5 years. This is the "melting ice cube" the bears fixate on: paper volumes are mature and activity (retrievals/rotations) declines as data digitizes, but revenue management (price) has more than offset volume softness.
2. Global Data Center — the growth story and the reason the stock re-rated. 31 data centers across 21 global markets; 488 MW leasable (~97% leased) + 852 MW under construction or held for development = 1,340 MW total potential. Five of the largest global hyperscalers are customers; the DC business is >20 years old. WALE 10.3 years.
3. Asset Lifecycle Management (ALM) — decommissioning, data-erasure, and resale/recycling of enterprise + hyperscale IT hardware. Revenue $633M in 2025 (+63% reported / +40% organic) against a TAM management sizes at $35B ($10B enterprise + ~$25B data-center decommissioning); Fortune-1000 ALM customers rose to 360 from 270. Reported in Corporate & Other.
4. Digital Solutions — scanning/imaging, the Insight Digital Experience Platform (DXP, v2.0 launched Oct 2025 with "AI agents"), records digitization. The bridge that keeps the physical archive relevant.

Contract structure: majority-recurring; no single customer >~3% of revenue — genuinely diversified, unlike a hyperscaler-concentrated peer. ~29,400 employees (11,700 US / 17,700 ex-US). Reporting currency USD; ~33% of revenue is non-USD (GBP, EUR, CAD, AUD) so FX is a real swing factor (FY25 booked a $105.6M FX transaction loss).

Verdict (Lens 1): A genuinely diversified, sticky, mission-critical services franchise wrapped in a REIT chassis — using a low-growth-but-durable records annuity to bankroll a high-growth, capital-hungry data-center build. The model only works if the records cash actually covers the dividend and the DC capex; that is the whole debate (Lens 10–13).

Lens 2 · Supply Chain

Iron Mountain is asset-heavy and largely vertically integrated on its core, so the "supply chain" is really a capital + power + real-estate chain rather than a goods chain:

• Upstream (RIM): real estate (1,343 facilities, 98.6M sq ft — 1,111 leased / 232 owned), racking, secure-transport fleet (courier trucks, mobile shred units), and labor. Inputs are commoditized; the moat is the custody process, not the inputs.
• Upstream (Data Center): this is where named chokepoints live —
• Power & grid interconnect — the binding constraint for the entire sector. IM matches 100% of DC energy with clean energy (since 2017) via its Green Power Pass; founding signatory to the UN 24/7 Carbon-Free Energy Compact. Power availability + pass-through power pricing is explicitly the swing variable on DC margin.
• Construction / development capital — DC growth capex was $1,747.0M in FY25, funded by dedicated, single-asset project debt (Virginia 3/4/5/6/7 term loans, Arizona JVs) plus corporate debt.
• JV partners — AGC Equity Partners (20% equity-method JV) and a December-2025 Arizona 3 JV (NCI ~$74.8M) co-fund the build; GIP/Global Industrial Partners appears on the short side, not the build side.
• Customers (the demand chain): five of the largest global hyperscalers (DC); 95% of the Fortune 1000 (RIM); ALM sells back into the same hyperscalers (decommissioning their retired servers) — a neat closed loop where IM both houses and later recycles hyperscaler hardware.
• Downstream (ALM): secondary market for refurbished components / remarketed IT assets — component pricing is a revenue swing factor (called out as a tailwind in 2025).

Single-source / chokepoint flags: (1) power is the industry-wide bottleneck and IM's DC growth is gated by energized capacity — management's "~400 MW energized and ready within 24 months" is the key supply metric; (2) capital markets access — the build is debt-funded, so a closed high-yield window or a downgrade is a genuine supply-chain risk (covenants have no rating trigger, which helps).

Verdict (Lens 2): Not a fragile goods chain — a power-and-capital chain. The vulnerabilities are grid interconnect timing and continued access to project + corporate debt, not a missing component.

Lens 3 · Competitive Advantages (moats)

RIM moat — strong and underappreciated:

• Switching costs / inertia — once millions of cubic feet of a customer's records sit in IM's custody, moving them is costly, risky, and pointless. This produces multi-decade retention and is why price increases stick even as volume is flat. The bears' "melting ice cube" framing understates how slowly the ice melts and how much pricing power sits on top of it.
• Scale + global footprint — 1,340 locations, 61 countries. No competitor matches the global single-vendor offering; rivals (Access Corp, Restore plc, Stericycle/Shred-it, Crown) are regional. IM acquired its biggest historical rival, Recall, in 2016 (DOJ-mandated divestitures of overlapping assets) — i.e., it literally bought the #2.
• Chain-of-custody / compliance brand — for "the world's most heavily regulated organizations," trust and certifications (ISO 27001/22301, SOC 2, HIPAA, FISMA HIGH) are the product. Reputational, hard to replicate.

Data Center moat — real but contested:

• Compliance + sustainability differentiation (Green Power Pass, FISMA HIGH) is a genuine niche for regulated/government workloads, but in the broad hyperscale market IM competes with far larger, better-capitalized operators (Equinix, Digital Realty, CoreWeave-style neoclouds, the hyperscalers' own self-build). On pure scale and cost of capital IM is a price-taker, not a price-setter here.
• The cross-sell ("we already store your records, now take our DC / ALM") is plausible but small — only ~5% of customers buy from >1 business unit, which is both the bull's runway and the bear's evidence the synergy is mostly theoretical so far.

Bargaining power: strong over RIM customers (captive archives); weak-to-neutral over hyperscale DC tenants (who have alternatives and negotiate hard, though 10-yr WALE locks them once signed); improving over ALM customers as the network scales.

Verdict (Lens 3): Two different moats stapled together. The RIM moat is wide and durable (switching costs + scale + brand). The DC moat is narrow (compliance niche) and the company is a follower on cost of capital. The equity thesis is essentially "can a wide-moat annuity fund a narrow-moat growth call without breaking the balance sheet?"

Lens 4 · Segments

Two reportable segments (Global RIM, Global Data Center) + Corporate & Other (ALM + Fine Arts). FY2025 vs FY2024, all ``:

Consolidated revenue split: Storage Rental $4,052.5M (+10.1%, organic +9.6%) / Service $2,849.2M (+15.5%, organic +11.2%).

Trend read:

• RIM is the ballast: ~77% of revenue, growing mid-single-digits almost entirely on price (revenue management), not volume. 44.7% segment margin. Decelerating-but-reliable.
• Data Center is the accelerant: ~12% of revenue but +29.6% and margin expanded 620bps to 51.8% as prior-period leases commenced and pricing/power-ramp kicked in. DC storage-rental specifically +31.5%. This is the line the multiple keys off.
• ALM (in Corp & Other) is the wildcard: +46.6% reported, but the segment still runs an Adj-EBITDA loss (−$205.9M, improving from −$269.3M) because corporate overhead is parked here — so ALM's true unit economics are obscured. Worth watching.

Q1-FY26 confirms the mix shift: DC revenue $254.7M (+47.1%, margin 52.1%), RIM $1,404.1M (+11.8%), and management/press note the growth trio (DC + digital + ALM) now >30% of total revenue. The business is visibly tilting from records toward digital infrastructure.

Geography: ~67% USD-reported; largest non-USD exposures GBP (6.8%), EUR (6.8%), CAD (4.4%), AUD (2.7%). No single-country concentration risk outside the US.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 FY2026, period end 2026-03-31)

All figures unless tagged.

• Revenue $1,936.1M, +21.6% YoY (constant-currency +18.6%, organic +17.2%). Storage rental $1,094.8M (+15.4%, organic +12.4%); Service $841.4M (+30.6%, organic +24.3%).
• Operating income $395.2M, +55.4% (margin 20.4% vs 16.0%) — the jump is partly clean operating leverage and partly the disappearance of Restructuring (Project Matterhorn) costs → $0 vs $54.7M in Q1'25 (Matterhorn completed end-2025).
• Adjusted EBITDA $707.9M, +22.1%; margin 36.6% (+20bps).
• Net income $149.0M vs $16.2M — the optical 9x jump is flattered by (i) Matterhorn restructuring rolling off, (ii) Other income flipping to +$4.7M from a −$28.5M FX loss. Diluted GAAP EPS $0.48 vs $0.05.
• AFFO $426.1M / $1.43 per share, +22% — the metric the market and the bears both fight over. (Q1'25 AFFO was $348.4M / $1.17.)
• Segments: DC $254.7M (+47.1%, margin 52.1% — down 30bps on higher pass-through power); RIM $1,404.1M (+11.8%, margin 44.0%); ALM/Corp Adj-EBITDA loss narrowed to −$42.5M.

Beat & guidance: AFFO topped consensus and management raised FY2026 guidance. The raised guide: revenue $7.825–7.925B, Adjusted EBITDA $2.925–2.965B, AFFO $5.79–5.86/sh ($1.735–1.755B) — up from the February initial guide of revenue $7.625–7.775B / AFFO $1.705–1.735B. Tone shifted clearly more confident on DC.

Balance-sheet flags (the part that matters): OCF doesn't cover the build. FY25 operating cash flow $1,340.0M vs investing outflow $(2,574.2)M — the ~$1.2B+ gap (DC capex) is plugged with debt + equity. Cash on hand only $158.5M. Receivables $1,443.7M (growing with revenue). Deferred revenue $402.1M.

Market reaction: stock hit a 52-week high $134.09 on 2026-05-06 on the print/guidance raise; ~$132 late June. The print was the catalyst.

Verdict (Lens 5): A genuinely strong operating quarter — DC +47%, organic +17%, guidance raised. The cleanliness is the only question: net income is flattered by one-offs and AFFO leans on add-backs the bears dispute.

Lens 6 · Earnings Calls (sentiment trend) — `` (no transcripts on shelf)

From available call/coverage summaries (Q4-FY25 through Q1-FY26):

• Dominant theme: data center + AI demand. CEO Meaney's recurring lines: "32 megawatts through April," "meaningfully above the 100 MW guidance," "~400 MW of capacity energized and ready within the next 24 months". The narrative has been deliberately re-centered from "records storage" to "digital infrastructure / AI."
• Tone trajectory: increasingly confident. Q4-FY25 ("record results," 17% Q4 revenue growth) → Q1-FY26 (guidance raise). Management leans hard on "records-business resilience + DC/ALM acceleration."
• What they emphasize now vs before: more airtime to DC leasing MW, ALM remarketing, and "Project Matterhorn is done, the investment phase is behind us" (a margin-inflection pitch). Less airtime to physical-volume trends (the soft spot).
• What they've stopped saying: the heavy "transformation investment" framing of 2023–24 (Matterhorn) is gone now that the $574.4M program closed.

Sentiment caveat: management tone turned more bullish at the exact time Gotham (Nov 2025) attacked the credibility of the very metrics (AFFO, Adj EBITDA, leverage) the bullish narrative rests on. Treat the call optimism as management's case, not independent confirmation.

Verdict (Lens 6): Consistently and increasingly bullish, DC-centric, margin-inflection framed. The tone is a sell-side asset; it is also exactly what a short-seller would expect to hear. Get the actual transcripts onto the shelf next refresh to run real sentiment-delta.

Lens 7 · Comps

Peer set: data-center REITs (the multiple driver) + a records peer note. Multiples are `` with source/date or n/a; never fabricated.

Read:

• IRM's EV ~$58.7B on ~$2.9B FY26 Adj EBITDA → fwd EV/EBITDA ~19.8x, below EQIX (~25x) and broadly in line with DLR — i.e., the market already gives IRM a partial "data-center-REIT" multiple, well above where a pure records-storage company would trade (a Stericycle-type would be high-single/low-double-digit EV/EBITDA). The re-rating has largely happened.
• ttm EV/EBITDA ~25x is 62% above IRM's own 10-yr median of ~15.6x — so on its own history the stock is rich; the bull case requires the DC mix-shift to justify a permanently higher multiple.
• ROE is not a usable comp for IRM (negative equity, REIT) — the SKILL's "5-yr avg ROE" column is n/a here by construction, not by omission.
• Records-management peers are not cleanly listed (Recall absorbed by IRM in 2016; Access Corp private; Restore plc is UK-listed small-cap; Stericycle/Shred-it within a larger waste co), so there is no apples-to-apples public records comp — part of why the market prices IRM off the DC-REIT cohort.

Verdict (Lens 7): Already priced as a data-center REIT, not a records company — ~20x fwd EV/EBITDA, ~22–23x AFFO, above its decade median. Cheaper than EQIX, ~in line with DLR. The multiple leaves no margin for the AFFO/leverage credibility questions to be proven right.

Lens 8 · Stock-Price Catalysts (>5% moves, ~last 5 yrs) — mostly ``

What actually moves IRM:

• Up — guidance raises on DC leasing. The May-2026 +14.7% move after Q1-FY26 record results + raised guidance (52-wk high $134.09, 2026-05-06). The dominant positive catalyst type is DC leasing momentum + AFFO guidance, not the records business.
• Up — the AI/data-center secular bid. Through 2025–26 IRM rode the AI-datacenter tape; YTD it outran most traditional REITs and traded near record highs. It has become an "AI-adjacent" name.
• Down — the Gotham City short report (2026-11-21… i.e., Nov 21, 2025). Stock fell ~5–6% intraday on a forensic report attacking AFFO/EBITDA/leverage credibility (full detail in Lens 10 & 13).
• Down — leverage / rate sensitivity. As a ~$16.5B-debt, ~22%-floating-rate REIT, IRM sells off on higher-for-longer rate fears; ~$3.64B variable debt, +100bps ≈ −$41.9M net income.
• Institutional exit signal: Baron Capital's Real Estate Income Fund fully exited IRM, citing weaker DC bookings and the short report.

Pattern: the market reacts to (1) DC leasing/guidance, (2) the AI narrative, and (3) credibility/leverage shocks — not to the records business, which is treated as stable background. This makes IRM more of a "data-center growth + balance-sheet-trust" stock than a "storage annuity" stock, despite ~77% of revenue being records.

Verdict (Lens 8): The tape keys off DC leasing and balance-sheet trust. The Gotham report is the single most important negative catalyst on the board and is unresolved.

---

Phase C — Judge people & books

Lens 9 · Management — +

• CEO William L. Meaney — appointed Jan 2013, ~13.3-yr tenure. Prior: CEO of The Zuellig Group (~$12B B2B conglomerate, sales tripled on his watch). A long-tenured, credentialed operator.
• Track record (quantified): under Meaney, 5-yr cumulative TSR ~+329% vs ~+23% for the MSCI US REIT Index (to 12/31/2024) — a genuinely exceptional relative record. Revenue roughly doubled across his tenure (Matterhorn-era acceleration); FY24 record results (rev $6.15B, AEBITDA $2.24B, AFFO $4.54/sh), FY25 again record (rev $6.9B, AFFO $5.17/sh).
• Skin in the game: Meaney holds ~295,650 shares directly (~0.10% of shares) worth ~$37M — meaningful in dollars, small in %. Comp ~$18.4M (FY24), ~93% at-risk / performance-linked; PUs vest 0–~350% on revenue + relative-TSR vs the MSCI US REIT index, with a min-ROIC gate in year 3.
• Capital allocation: the central judgment call. Bull read — disciplined pivot of a low-growth annuity into a high-growth, high-margin DC platform that the market rewarded 3:1 vs peers. Bear read — funded almost entirely with debt + dilution against negative book equity, paying a dividend that exceeds GAAP earnings several-fold, with AFFO/EBITDA definitions Gotham says flatter the picture. ROE is unusable (negative equity); ROIC discipline is asserted via the PU gate but not independently verified here.
• Red flags: (i) comp is heavily tied to revenue and Adjusted metrics — exactly the metrics under SEC/short-seller scrutiny, a potential incentive to maximize the disputed numbers; (ii) the dividend-vs-cash-flow gap; (iii) related-party — December-2025 Arizona 3 JV and AGC JV are disclosed and look arm's-length, no obvious self-dealing in the 10-K.
• Archetype: professional manager (not founder), long-tenured, capital-markets-savvy, growth-and-TSR-oriented. Appropriate for a scaling-REIT stage — but the incentive design rewards the very aggressiveness the bears flag.

Verdict (Lens 9): A strong, long-tenured operator with a standout TSR record and high pay-at-risk — but compensated on the Adjusted metrics that are under SEC/Gotham scrutiny, which is the governance tension at the heart of the name.

Lens 10 · Forensic Red Flags

Income statement / cash-flow divergence (the core issue):

• AFFO ($1.54B FY25) is ~2.6x FFO (Nareit) ($584.2M) and ~1.4x FFO (Normalized) ($1,079.9M). Gotham's central claim — a typical REIT's AFFO runs 15–20% below FFO; IRM's runs above it, every year since REIT conversion — is directly corroborated by the filing's own reconciliation. The gap is built from large add-backs: real-estate depreciation ($421.6M), restructuring ($195.9M), SBC ($140.3M), "other expense" ($118.5M), plus recurring-capex treatment. None of these are prima facie improper for a REIT, but the magnitude and direction relative to peers is the anomaly.
• Net income $152.3M vs Adjusted EBITDA $2,574.0M — a ~17x gap, bridged mostly by D&A ($1,024.4M) and interest ($829.3M). Heavy-capex REIT math, but it means GAAP profitability is thin (net margin 2.2%) while the headline numbers are large.
• SBC $140.3M flatters every non-GAAP line (it's added back to Adj EBITDA, FFO-Normalized, AFFO and Adj-EPS).

Balance sheet:

• Negative book equity: total stockholders' deficit −$981.0M (widening from −$503.1M); total deficit −$709.3M; distributions-in-excess-of-earnings −$5,405.1M. The company has paid out cumulatively ~$5.4B more than it has earned — sustainable only if AFFO (not GAAP) is the right yardstick. If Gotham is right that AFFO overstates distributable cash, the dividend is being part-funded by capital, not earnings.
• Goodwill $5,285.8M on a −$0.98B-equity base — any impairment would deepen the deficit.
• Total debt $16,544.5M, cash $158.5M → net debt ~$16,273.5M.

Leverage — the disputed metric:

• Company reports net total lease-adjusted leverage 4.9x (FY25) / 4.8x (Q1-FY26) vs a 7.0x covenant max; fixed-charge coverage 2.5x vs 1.5x min; in compliance. Covenants use EBITDAR with adds for completed acquisitions, executed-but-not-commenced DC leases, and restructuring — i.e., credit for revenue not yet earned. Gotham calculates true leverage at ~9x and says Adjusted EBITDA is overstated 25–35%. Net debt / reported Adj EBITDA ≈ 16.27/2.57 = ~6.3x — already above the 4.9x lease-adjusted figure, lending arithmetic plausibility to the bears' "headline understates it" point, though 9x requires their EBITDA haircut to hold.
• WAIR 5.6%; ~22% ($3.64B) floating. Debt ladder well-termed (2027–2034); no maturity wall; covenants have no rating trigger (a real positive).

Receivables/inventory vs revenue: receivables $1,443.7M grew ~12% with revenue — in line, no obvious channel-stuffing flag.

Regulatory findings (required sub-section):

• SEC EDGAR EFTS (LR + AAER): zero Litigation Releases and zero AAERs naming Iron Mountain in 2021-06-29→2026-06-29.
• SEC comment-letter trail (material): Gotham states, and multiple outlets corroborate, that the SEC began questioning Iron Mountain's Adjusted EBITDA via correspondence (comment letters) earlier in 2025. This is a disclosure-review process, not an enforcement action — but it is directly on the disputed metric and is the single most important open item. (Not independently verified against EDGAR CORRESP in this run — flag to pull the actual UPLOAD/CORRESP letters next refresh.)
• 10-K Item 3 (Legal Proceedings): "In the opinion of management, no material legal proceedings are pending…". So as of the FY25 10-K, the SEC comment-letter process had not risen to disclosed material litigation.
• Non-SEC enforcement: no material FTC/DOJ/FDA/CFPB actions surfaced in web search. (Historical note: the 2016 Recall acquisition required DOJ-mandated divestitures — an antitrust remedy, long resolved, not an enforcement penalty.)
• Auditor: Deloitte & Touche LLP; ICFR assessed effective; clean 404(b) attestation.

Verdict (Lens 10): The accounting is aggressive-but-disclosed, not fraudulent-on-the-face. The anomalies Gotham names — AFFO > FFO, negative equity funding a >100%-of-GAAP dividend, lease-adjusted leverage that flatters a ~6.3x raw net-debt/EBITDA, SBC-heavy non-GAAP — are all visible in the filings, which is what makes the short credible rather than crankish. The unresolved SEC Adj-EBITDA comment-letter is the swing factor: a benign close is a clearing event; an escalation or a forced restatement of the Adjusted metrics would break the multiple.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (AFFO/sh — the REIT-relevant metric, not EPS)

For a REIT the scoreable number is AFFO per share, not EPS. Built bottom-up from FY25 actuals + guidance; every input labeled; output ``.

Anchors (/): FY25 AFFO $5.17/sh; FY25 revenue $6,901.7M; FY25 Adj EBITDA $2,573.95M (37.3% margin); shares ~297M; FY26 raised guide AFFO $5.79–5.86/sh, revenue $7.825–7.925B, Adj EBITDA $2.925–2.965B.

Drivers: RIM +~5–6% (price-led, volume ~flat); DC +~30–45% (488 MW→ stabilizing + ~400 MW energizing over 24 mo, 100 MW+ leasing/yr); ALM +20–40% (off a loss-making but improving base); offsets — interest expense ($830M+ and rising with the build), ~2% annual share dilution (SBC/equity funding), pass-through power compressing DC margin modestly.

Reading: the bull and base are guidance-consistent extrapolations of a double-digit-AFFO-grower. The bear path is not "slower growth" — it is "the AFFO definition changes." If an SEC outcome or accounting reset forces tighter recurring-capex / lower add-backs, reported AFFO/sh could fall even with revenue rising, which is the asymmetric risk the headline trajectory hides.

Forecast log (NOT executed — unattended watchlist; forecast.ts create is skipped per SKILL): the candidate Brier forecast to log on a genuine commit would be "IRM FY2026 AFFO/sh ≥ $5.79, p≈0.80, resolves 2026-12-31, tags iron-mountain,deep-dive" — i.e., lean toward management hitting the low end of the raised guide, with the tail risk being a metric-definition shock rather than an operational miss.

Verdict (Lens 11): Base AFFO ~$5.82 (FY26) → ~$7.00 (FY28) if the AFFO definition holds. The entire projection is hostage to whether "AFFO" survives SEC review intact.

Lens 12 · Bull vs Bear

Bull case. A wide-moat records annuity (240k customers, 95% of the Fortune 1000, switching-cost-locked, price-power-positive) throws off durable cash that funds a genuinely high-growth, 50%+-margin data-center platform levered to the strongest secular demand in tech (AI/cloud). DC +47% in Q1-FY26, leasing tracking "meaningfully above" 100 MW, ~400 MW energizing within 24 months, 1,340 MW total potential runway. ALM adds a second 30–60%-grower into a $35B TAM, closing a loop with the same hyperscalers. Management has a 5-yr +329% TSR vs +23% peer record and just raised guidance. At ~20x fwd EV/EBITDA (below EQIX) with a covered-on-AFFO ~2.7% dividend growing ~10%/yr, you're paying a reasonable price for a double-digit AFFO compounder with optionality on the AI buildout.

Bear case (2–3 permanent-impairment risks):

1. The metrics break. If the SEC comment-letter process forces a restatement/redefinition of Adjusted EBITDA or AFFO, or validates Gotham's ~9x-vs-5x leverage gap, the stock loses its data-center-REIT multiple and its dividend-coverage story simultaneously — a permanent de-rating, not a dip.
2. The balance sheet. ~$16.5B debt, negative book equity, a dividend exceeding GAAP earnings, and a DC build that OCF can't self-fund (−$1.2B/yr gap). A closed financing window, a downgrade-driven cost-of-capital jump, or a rate shock (22% floating) turns the virtuous build into a forced-deleveraging spiral.
3. DC is a price-taker market. IRM competes on cost of capital against EQIX/DLR/hyperscaler self-build; an AI-capex air-pocket or DC oversupply (weaker bookings — already flagged by Baron's exit) compresses the exact segment the multiple capitalizes.

Pre-mortem (18 months out, thesis broke): Q3/Q4-2026 DC bookings disappointed (the "100 MW+" didn't convert to commenced leases on schedule); the SEC comment-letter escalated into a required restatement of Adjusted EBITDA; the dividend-coverage narrative cracked; the multiple compressed from ~20x to ~13–15x EV/EBITDA (toward its historical median) and the stock halved from ~$132 to ~$65–75, with the negative-equity balance sheet leaving no cushion.

Are multiples too high? On its own 10-yr history, yes — ttm EV/EBITDA ~25x vs ~15.6x median. Versus DC-REIT peers, fairly priced if you accept the Adjusted metrics. The multiple embeds metric-credibility you cannot yet confirm.

Contrarian view (what the market refuses to see): The market is treating IRM as a clean AI-data-center compounder and pricing through the Gotham report as noise (Wells Fargo defended; stock near highs). The contrarian read is the inverse on each side: bulls under-weight that an unresolved SEC inquiry into the headline metric is a live, binary de-rating risk; but bears under-weight that the records moat is wider and the DC pivot more real than "melting ice cube" allows — IRM genuinely has hyperscaler tenants and a 1,340 MW pipeline, so even a haircut version of the AFFO story supports a far-from-zero equity. The truth is probably "great franchise, aggressive accounting, fully-priced stock" — which argues WATCH, not chase.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case (the Gotham thesis, sourced):

• "The non-GAAP numbers are engineered." — Adjusted EBITDA overstated 25–35%; reported leverage manipulated to ~5x when true leverage is ~9x; AFFO is 2x FFO (backwards vs peers) and has exceeded FFO every year since REIT conversion; dividend payout has exceeded 200% of true cash flow over ~12 years despite the claimed ~60%. The SEC was already questioning Adj EBITDA via comment letters in 2025. Corroboration: the filing's own reconciliation confirms AFFO > FFO and a >100%-of-GAAP-earnings payout, and net-debt/reported-EBITDA is ~6.3x raw — so the bear's direction is filing-supported even if the precise 9x needs their EBITDA haircut.
• Where revenue is concentrated / what shifts it: not customer-concentrated (good), but growth is concentrated in DC + ALM, the two most capital-intensive, most competitive, most cyclical lines. Strip those and you have a low-growth records business decorating a $16.5B debt load.
• Why the moat is weaker than bulls think: the DC moat is a follower's moat — IRM has worse cost of capital than EQIX/DLR and far worse than hyperscaler self-build; the "cross-sell" is ~5% of customers after years of trying.
• Most dangerous competitor bulls underestimate: the hyperscalers themselves (self-build) and well-capitalized DC REITs that can out-bid IRM on power and price; in ALM, the customers who in-source recycling.
• Worst capital-allocation / incentive issues: paying a dividend out of capital (negative retained earnings −$5.4B); comp tied to the disputed Adjusted metrics; funding growth with dilution against negative equity.
• Assumptions that must hold for ~$132: (i) AFFO/Adj-EBITDA survive SEC review unchanged; (ii) DC leasing keeps beating; (iii) rates don't spike on the 22% floating book; (iv) the multiple stays ~2x its historical median.
• If growth disappoints 20–30%: DC is where the multiple lives — a 20–30% DC-growth shortfall (oversupply / AI-capex pause) likely takes EV/EBITDA toward the ~15.6x median → ~30–45% downside.
• Single permanent-impairment scenario & plausibility: an SEC-driven restatement/redefinition of Adjusted EBITDA or AFFO. Plausibility: non-trivial — the inquiry is real and on exactly that metric; outcome unknown. That single event would simultaneously break the multiple and the dividend-coverage story.

Counter to the short: Gotham/GIP are talking their book; the company is covenant-compliant with a no-rating-trigger, well-laddered debt stack; Deloitte signs a clean ICFR opinion; "aggressive but disclosed" REIT accounting is not fraud; and a 240k-customer, 1,340-MW franchise has substantial value even on haircut metrics. The short is a credibility/de-rating thesis, not (yet) a going-concern one.

Verdict (Lens 13): The strongest, best-sourced bear case on this name is not "no-growth records company" — it's "the cash flows aren't as clean as the headline AFFO and the SEC is asking." It is filing-corroborated in direction, unresolved, and binary. That is precisely why this is a WATCH with a hard catalyst (SEC resolution) rather than a buy or a short.

Lens 14 · Management Questions (ordered by information value)

1. The SEC has corresponded with you on Adjusted EBITDA — what specifically did they question, what was your response, and is the matter closed? (The single highest-VOI question; resolves the central thesis risk.)
2. Walk us through why your AFFO exceeds your FFO every year since REIT conversion, when the typical REIT's AFFO runs 15–20% below FFO — what in your recurring-capex and add-back treatment drives that, and would a stricter recurring-capex definition lower it?
3. Reconcile your reported 4.8x lease-adjusted leverage with a ~6.3x raw net-debt/Adjusted-EBITDA — how much of the gap is the EBITDAR adjustment for executed-but-not-commenced DC leases, and what is leverage on trailing in-place EBITDA only?
4. Your dividend has exceeded GAAP earnings for years and you carry −$5.4B distributions-in-excess-of-earnings — on a pure cash basis (OCF − recurring capex − cash interest), is the current dividend covered without new financing?
5. DC capex was $1.75B in 2025 while DC generated $803M revenue — what is the stabilized unlevered yield-on-cost you underwrite, and how has it moved as construction and power costs rose?
6. Of the "~400 MW energizing within 24 months," how much is already leased vs speculative, and what is the lease-up assumption for the speculative portion?
7. ALM grew 40% organic but Corporate-&-Other still runs an Adjusted-EBITDA loss — what is ALM's standalone segment margin, and when does it turn profitable on its own?
8. With ~22% of debt floating and WAIR 5.6%, what is your plan if rates stay elevated through the 2027–2031 maturity ladder — and at what cost of capital does the DC build stop being accretive?
9. Records physical volume is mature — at what point does price-led RIM growth exhaust, and what is your assumed annual revenue-management (pricing) headroom?
10. The DC business competes with EQIX, DLR, and hyperscaler self-build that have lower cost of capital — what is your durable edge beyond the compliance/sustainability niche, and is it defensible at hyperscale?
11. Only ~5% of customers buy from more than one unit after years of cross-sell investment — what changes that, and what is the realistic ceiling?
12. CEO/exec comp is weighted to revenue and Adjusted metrics now under scrutiny — would the board consider adding a cash-flow-coverage or GAAP-based gate to the incentive plan?
13. Goodwill is $5.3B against negative book equity — what would trigger an impairment test, and how sensitive is it to a DC-multiple compression?
14. Baron Capital exited citing weaker DC bookings — what did they see in your booking trend that you'd push back on, and what's the actual booked-but-not-commenced backlog in MW and dollars?
15. If you could only fund either the dividend or the data-center build for the next three years, which wins, and what does that say about the capital structure?

---