Real Estate · Infrastructure & Digital REITs
Infrastructure & Digital REITs Primer
Tower and data centre REIT primer with an Excel valuation model. Covers escalators, churn, bookings, power constraints, P/AFFO and rate sensitivity.
2026 Edition · data as of June 2026
- pages
- 39
- sections
- 15
- model sheets
- 10
Choose a format
PDF only
The full primer
£25 / ~$32
- 39 pages, 15 sections
- Worked examples and applied cases
- 2026 Edition, data as of June 2026
Excel model
Valuation template
£45 / ~$58
- 10 sheets (.xlsx)
- Runs the primer's worked valuations live
PDF + Excel model Best value
Both files together
£59 / ~$76
- The 39-page primer
- The 10-sheet Excel model
- £70 bought separately · £11 less
Complete Real Estate Library
All three Real Estate industries: three primers, three Excel models (Equity, Mortgage, and Infrastructure & Digital REITs).
Inside the primer
The 15-section contents, a worked valuation page, and the Excel dashboard.
Contents
- 01 How Infrastructure and Digital REITs Make Money
- 02 Listed Company Types
- 03 The Asset Lifecycle
- 04 Segments and Sub-Markets
- 05 Revenue Drivers: Generic Build
- 06 Cost Structure: AFFO, Churn and Leverage
- 07 Valuation Frameworks
- 08 Worked Example: Organic Growth Identity
- 09 Worked Example: Two-Stage AFFO Discount Model
- 10 Applied Cases: Tower Economics
- 11 Applied Cases: Data Centre Economics
- 12 The Companies in This Primer
- 13 Key Metrics and Screening
- 14 Risks, Benchmarks and Case Study
- 15 Glossary and Cheat Sheet
Worked examples and applied cases · 39 pages · 15 sections · 2 worked valuations
The Excel model
Educational material. This primer and its model are not investment advice or a recommendation to buy or sell any security, and they are not personalised. Worked valuations use illustrative calibrations, not fair-value estimates for any company.
Towers and data centres share a REIT wrapper but earn differently. Tower landlords compound tenant billings through contractual escalators and colocation on built steel; data-centre landlords convert signed bookings into energised megawatts after construction and grid interconnection. Both trade on AFFO per share, not property NAV, and both sit at the rate-sensitive end of the REIT spectrum because the cash flows are long-dated. Standard equity-REIT toolkit (cap rates, same-store NOI) misses the escalator identity, the churn definition traps, and the bookings-to-revenue lag that drive these names.
Tower and data centre REIT valuation benchmarks by business model
Both asset classes trade on price to AFFO, but data centres sit higher because their growth runs faster and needs more capital. These are the primer's bands, with the operating metric that places a company inside each one.
| Business model | Primary multiple | Typical range | What sets the position |
|---|---|---|---|
| Tower landlord, US or global | P/AFFO; EV/EBITDA as a cross-check | 15-22x AFFO; 18-25x EV/EBITDA | Organic tenant billings growth (4-6% band) and churn (1-2% domestic) |
| Leveraged tower compounder | P/AFFO plus net debt/EBITDA | 15-22x AFFO; leverage 5-6x normal, above 7x caution | How much refinancing lands in a rate rise |
| Retail colocation | P/AFFO | 22-28x AFFO; 20-28x EV/EBITDA | Interconnection density; monthly recurring revenue churn, about 2-3% a quarter |
| Hyperscale / wholesale data centre | P/Core FFO or P/AFFO, basis labelled | 22-28x | Signed bookings and energisation dates, with a 12-24 month lag to revenue |
The primer discounts AFFO at a 9.5% cost of equity: the 4.50% 10-year Treasury plus a 5.0% equity risk premium. Rates set the level for the whole group, because when the Treasury rises AFFO yields rise with it and multiples compress. Within a band, a tower company moves up on a wider gap between its roughly 3% escalator and its churn, and on colocation that fills the rest. A data-centre company moves up on bookings close to energisation.
Worked example: a two-stage AFFO valuation of a tower REIT
Take a hypothetical tower REIT with $10.00 of AFFO per share. Organic growth comes from the primer's identity: a 3.0% escalator, less 2.0% churn, plus 4.0% from colocation and amendments, gives 5.0%. AFFO compounds at that rate for seven years, then at 3.5% for good, all discounted at 9.5%.
| Step | Per share |
|---|---|
| AFFO today | $10.00 |
| Year-7 AFFO at 5.0% growth | $14.07 |
| PV of years 1-7 AFFO at 9.5% | $59.39 |
| Terminal value in year 7 ($14.07 x 1.035 / 6.0%) | $242.72 |
| PV of terminal value | $128.59 |
| Value per share | $187.98 |
| Implied P/AFFO (an output) | 18.8x |
| Same model with the 10-year Treasury 100 bp higher (9.5% becomes 10.5%) | $160.79 (16.1x) |
The implied 18.8x lands inside the 15-22x tower band, but the model produced it; nobody typed it in. About two-thirds of the value sits in the terminal, which is why these are the most rate-sensitive REITs. A 100 basis point rise in the Treasury takes value down 14.5% while AFFO still compounds at 5%. The primer also shows the shortcut, an assumed move from 18x to 15x that costs 16.7%, and why re-running the cost of equity is the method.
The primer runs the same model on a hypothetical data-centre landlord growing at 9% (implied 25.2x) and grids tower value against rates and growth; the Excel model does both with your own inputs.
Which tower and data centre REIT valuation method fits which company
| Company type | Primary method | Check with | Why |
|---|---|---|---|
| Global or US tower landlord | Two-stage AFFO model; implied P/AFFO against 15-22x | Organic tenant billings growth; churn on its stated basis | Maintenance capex on steel is low, so AFFO tracks dividend capacity |
| Leveraged tower compounder | The same model plus net debt/EBITDA | The covenant definition of leverage | Leverage amplifies AFFO growth and adds refinancing risk |
| Retail colocation | AFFO model; implied P/AFFO against 22-28x | Monthly recurring revenue churn | Interconnection density earns the premium |
| Hyperscale / wholesale data centre | P/Core FFO or P/AFFO, basis labelled | Bookings trend against energisation dates | Growth capex is heavy and backlog leads AFFO by 12-24 months |
| Land bank or build pipeline | Yield on cost | Signed but not yet energised backlog | Not yet AFFO, so an AFFO multiple does not apply |
What the full primer adds
The primer builds from how the two machines make money through company types, asset lifecycle, and segment economics, before opening the AFFO, churn, and leverage screens. Two worked examples carry the method: the organic growth identity for towers, and a two-stage AFFO discount model for valuation and rate stress. Applied cases walk tower and data-centre economics in turn; a peer set, screening thresholds, and a risks chapter close the framework.
Free guides on tower lease escalators, churn definitions, AFFO for towers, the power constraint, and colocation vs hyperscale feed the same vocabulary. Research profiles for AMT, CCI, SBAC, EQIX, and DLR run the screens on filed results.
The companion Excel model runs ten sheets, from the tower and data-centre AFFO models to a rate-sensitivity screen, with the rate assumptions set on one sheet. Build the tower or data-centre path, then read the valuation and rate-stress outputs on any name in the sector.
Sheets: Quick Start, Instructions, Assumptions, Organic Growth & Churn, Tower REIT, Data-Centre REIT, Valuation Summary, Rate Sensitivity, Implied Multiple, Dashboard.
Tower and data centre REIT valuation: free guides
Tower and data centre REIT valuation FAQ
- How do you value a tower REIT?
- With a two-stage AFFO discount model. Grow AFFO per share at the organic rate for seven years, add a Gordon growth terminal value, and discount at a cost of equity built from the 10-year Treasury plus an equity risk premium. AFFO is already after interest, so it discounts straight to value per share. The implied P/AFFO is the output, checked against the tower band, not an input.
- What is a normal P/AFFO for tower and data centre REITs?
- The primer's bands are 15-22x AFFO for towers and 22-28x for data centres, with EV/EBITDA at 18-25x and 20-28x as the asset-level cross-check. Hyperscale landlords often lead with Core FFO rather than AFFO, so label the earnings basis before ranking a data-centre multiple against the band.
- Why are tower REITs so sensitive to interest rates?
- Most of the value sits far out. In the primer's tower example about two-thirds of value is the terminal, so a 100 basis point rise in the 10-year Treasury cuts value 14.5% even while AFFO keeps compounding at 5%. Leverage adds a second channel, because refinancing lifts interest cost on the same EBITDA.
- How do you calculate tower organic growth?
- Contractual escalator, less churn, plus colocation and amendments. At a 3.0% escalator and 2.0% churn the spread is only 1.0%, so a 4.0% colocation and amendment term is what lifts organic tenant billings growth to 5.0%. Stress churn first: the escalator is contractual, and the colocation term is the number you have to justify.
- Why do data centre REITs trade at higher multiples than towers?
- Faster capital-light growth. In the primer's model, the same 9.5% cost of equity gives 25.2x AFFO at 9% growth and a 4% terminal rate, against 18.8x at 5% and 3.5%, with no change of discount rate. The offset is heavier growth capex and a 12-24 month lag between a signed booking and energised revenue, so backlog is not yet AFFO.
See this methodology applied to a real company:
American Tower (AMT) →