Data Center REITs 2026: Best AI Infrastructure Income Plays for Dividend Investors
By Editorial Team | Updated August 24, 2026 | 13 min read | Finance, ETF, REIT
Quick Answer
Data center REITs own and lease the physical buildings — power, cooling, and interconnection infrastructure — that hyperscalers like Microsoft, Amazon, Google, and Meta run AI workloads on. The three largest US-listed pure plays are Digital Realty (DLR), Equinix (EQIX), and Iron Mountain (IRM), with dividend yields generally in the 2%-3% range. As a group, data center REITs returned about 33% year-to-date through June 2026 — outpacing the broader REIT sector and the S&P 500 — driven by hyperscaler AI capex guided above $700 billion for 2026, about 75% of it earmarked for AI infrastructure.
Table of Contents
What Are Data Center REITs?
A real estate investment trust (REIT) is legally required to distribute at least 90% of its taxable income to shareholders as dividends, in exchange for favorable tax treatment at the corporate level. Data center REITs are a specialized subsector that owns and operates the physical facilities housing servers, networking equipment, and the power and cooling infrastructure those servers need — then leases that space and capacity to tenants.
Those tenants fall into three broad groups: hyperscale cloud providers (AWS, Microsoft Azure, Google Cloud) leasing enormous power blocks for their own infrastructure; enterprises housing their own IT equipment; and, increasingly, AI labs and model developers that need dense, power-hungry compute clusters for training and inference.
Within the sector, two business models dominate:
- Wholesale / hyperscale operators lease entire buildings or large power blocks directly to hyperscalers on long-term contracts. Digital Realty is the clearest example.
- Interconnection / colocation operators rent smaller footprints to many different customers who pay a premium to be physically close to each other's networks, creating a network effect. Equinix is the clearest example of this model.
Why 2026 Is the Inflection Point
Data center real estate has gone from a niche subsector to one of the clearest ways to own the physical infrastructure behind the AI boom, and the numbers back it up:
- Capex is enormous and still rising. Microsoft, Google, Meta, and Amazon are guiding combined 2026 capital expenditure above $700 billion, with roughly 75% of that earmarked for AI infrastructure.
- Power, not land, is now the constraint. Hyperscalers are locking in 15-year leases years before facilities even break ground, and vacancy in primary US data center markets remains near record lows.
- The sector has outperformed dramatically. Data center REITs returned roughly 33.2% year-to-date through June 30, 2026, according to Nareit — well ahead of the broader REIT sector (up roughly 10%+ YTD, as measured by the Vanguard Real Estate ETF) and the S&P 500 (up roughly 10.5% YTD over the same stretch).
- Company-level results confirm the demand. Equinix posted Q1 2026 revenue of $2.444 billion, up 10% year-over-year, with a record $378 million in quarterly bookings and a raised full-year guidance. Digital Realty posted Q1 2026 revenue of $1.6 billion, up 16% year-over-year, signed the largest hyperscale lease in company history, built a $1.8 billion backlog, and raised its full-year core FFO guidance to $8.00-$8.10 per share — roughly 9% growth at the midpoint.
The read-through: data center REITs are one of the few ways to get direct exposure to the AI infrastructure buildout without owning a chipmaker, and without taking on the binary technology risk of a pre-revenue AI startup.
Top US-Listed Data Center REITs
| REIT | Ticker | Model | Approx. yield* | 2026 signal |
|---|---|---|---|---|
| Digital Realty | NYSE: DLR | Wholesale / hyperscale | ~2.8% | Q1 2026 revenue +16% YoY; FY26 core FFO guidance raised to $8.00-$8.10/share |
| Equinix | NASDAQ: EQIX | Interconnection / colocation | ~2.4% | Q1 2026 revenue +10% YoY; record $378M in bookings; raised FY guidance |
| Iron Mountain | NYSE: IRM | Diversified (records storage transitioning into data centers) | ~3.1% | Highest yield of the large-cap US names; more business-model transition risk |
*Approximate figures as of mid-2026 company filings and third-party REIT data trackers; yields fluctuate daily with share price. Verify current data before investing.
Equinix vs. Digital Realty is the classic quality-versus-growth trade-off in this sector. Equinix's interconnection moat — 281 data centers spanning six continents with roughly 513,000 network connections — creates a network effect where the platform becomes more valuable as more companies join, which supports a premium valuation and lower yield. Digital Realty's broader wholesale footprint (over 300 data centers across more than 55 markets, serving customers including Microsoft Azure, AWS, Google Cloud, Nvidia, Oracle, and Dell) gives it more direct operating leverage to hyperscale AI capex, which shows up as faster FFO growth and a higher current yield.
Iron Mountain sits apart from the other two: its core business has historically been physical records storage, and its data center segment represents a strategic pivot rather than a legacy strength. That transition carries more execution risk, but it's also why IRM typically offers the highest yield of the three.
Data Center REIT ETFs for Diversified Exposure
If picking and weighting individual REITs isn't your preference, two ETFs offer a single-ticket way to hold the theme:
- Global X Data Center REITs & Digital Infrastructure ETF (DTCR) — its two largest holdings are Equinix and Digital Realty, together representing close to a quarter of the fund, rounded out by smaller data center operators, towers, and adjacent digital infrastructure names.
- Pacer Data & Infrastructure Real Estate ETF (SRVR) — a broader digital-infrastructure real estate fund spanning data centers, cell towers, and related property types rather than data centers alone.
An ETF trades convenience and diversification for a management fee and less control over which specific names — and which yield/growth trade-off — you own.
Key Metrics to Evaluate Before You Buy
REITs don't screen well on ordinary stock metrics like P/E, because depreciation distorts GAAP net income for asset-heavy real estate businesses. Use these instead:
- FFO / AFFO per share and its growth rate — funds from operations is the REIT equivalent of earnings; look at the trend, not just the level.
- Price-to-FFO ratio — the REIT sector's equivalent of a P/E multiple, useful for comparing valuation across names.
- Dividend yield and payout ratio together — a high yield paired with an unsustainably high payout ratio is a warning sign, not a bargain.
- Occupancy rate and WALE (weighted average lease expiry) — longer WALE means more visible, contracted future revenue.
- Gearing / debt ratio — determines how exposed the REIT is to rising interest rates.
- Tenant concentration — how much revenue comes from the top one to three customers; heavy hyperscaler concentration cuts both ways.
- Signed-but-not-commenced backlog — increasingly the most important forward indicator in this specific sector, since hyperscalers are now signing leases years before a building is even finished.
Risks to Watch
- Interest rate sensitivity. REITs rely on debt financing; rising rates raise borrowing costs and can compress valuations across the sector, regardless of individual company execution.
- Power and grid constraints. Securing sufficient electricity for new capacity — not securing land — has become the industry's single biggest operational bottleneck, and it can delay revenue recognition on signed leases.
- AI capex cycle risk. Current lease demand is underwritten by hyperscaler capex guidance well above $700 billion for 2026. A deceleration in that spending, or a shift toward more compute-efficient chips that need less physical footprint, would soften demand.
- Tenant concentration. A handful of hyperscaler customers account for an outsized share of revenue at several of these REITs — a strength in a growth market, a vulnerability if any single relationship sours.
- Currency risk for cross-border holders. SGD- or MYR-based investors buying USD-denominated distributions (Digital Core REIT) or US-listed shares (DLR, EQIX, IRM) take on exchange-rate risk on top of the underlying investment.
- Valuation risk after a strong run. With the sector up roughly 33% year-to-date through mid-2026, the margin of safety is thinner than it was a year ago — growing backlogs support current valuations, but leave less room for disappointment.
Building a Data Center REIT Income Allocation
A common framework among REIT-focused analysts: treat data center REITs as an "AI infrastructure sleeve" within a broader real estate or dividend allocation — not as a standalone growth bet, and not as your only income holding, since large-cap US names yield less than traditional REIT categories. A conservative starting point is roughly 5%-10% of your real estate exposure, split between a hyperscale wholesale name (Digital Realty) and an interconnection name (Equinix), which diversifies both business model and customer concentration risk. Investors in Malaysia and Singapore can layer in Keppel DC REIT or Digital Core REIT for regional exposure and typically higher current yield, while accepting the added currency and market-specific risk that comes with it.
If you're modeling this alongside a broader portfolio, run the position sizing through a dividend or total-return lens rather than in isolation — our ETF picks and AI stock coverage cover the adjacent pieces of an AI infrastructure allocation — chips, hyperscalers, and now the real estate underneath them.
FAQ
What is the best data center REIT to buy in 2026?
There is no single "best" pick — it depends on what you're optimizing for. Equinix offers the strongest interconnection moat but a lower yield. Digital Realty offers broader hyperscale exposure with a higher yield and stronger recent FFO growth guidance. Iron Mountain generally pays the highest yield of the large-cap US names but carries more transition risk. Investors outside the US often combine one or two US names with a Singapore-listed REIT like Keppel DC REIT for regional diversification.
Are data center REITs a good investment for dividend income?
They can be, but they're not the highest-yielding REIT subsector. Large US data center REITs typically yield 2%-3%, below traditional REIT categories, because the market is pricing in AFFO growth rather than pure income. Singapore-listed names like Digital Core REIT and Keppel DC REIT have generally offered higher yields, roughly 4%-7%.
What's the difference between Equinix and Digital Realty?
Equinix runs an interconnection/colocation model built around a network effect and generally trades at a higher multiple with a lower yield. Digital Realty runs a broader wholesale/hyperscale model with more direct operating leverage to AI capex, typically offering a higher yield.
Can Singapore and Malaysia investors buy data center REITs?
Yes. Singapore-based investors can buy Keppel DC REIT, Digital Core REIT, and Mapletree Industrial Trust directly on the SGX. Malaysia-based investors typically access the same SGX names through brokerages offering Singapore market access, or the US names through a brokerage with US market access. Watch currency exposure — Digital Core REIT distributes in US dollars.
What are the biggest risks with data center REITs right now?
Interest rate sensitivity and power supply constraints are the two most-cited risks, followed by tenant concentration for REITs heavily dependent on a small number of hyperscaler customers.
Are data center REITs overvalued after the 2026 rally?
The sector returned roughly 33% year-to-date through June 30, 2026, pushing valuations higher and thinning the margin of safety if hyperscaler capex growth decelerates. Growing backlogs and signed leases partly justify the re-rating, but the risk of multiple compression is real.
What's the easiest way to get diversified data center REIT exposure?
A data center-focused ETF such as DTCR or SRVR spreads exposure across multiple REITs and adjacent digital infrastructure names in a single trade.
Personalize This With AI
You can paste this article into Claude, ChatGPT, Gemini, or Perplexity and ask it to apply the framework to your own situation. Useful prompts:
- "Given my existing portfolio is [X], how would a 5-10% data center REIT sleeve change my overall dividend yield and sector concentration?"
- "Compare DLR, EQIX, and IRM on price-to-FFO using the most current data you can find."
Data referenced in this article is drawn from company Q1/Q2 2026 filings and investor updates, SGX company disclosures, and third-party REIT data trackers including Nareit and Hoya Capital, as of the dates noted throughout. Market data changes daily; always verify current figures before making an investment decision. This article is not personalized financial advice.
.png)
Comments