Executive Summary
The year 2026 marks a critical point for data center Real Estate Investment Trusts (REITs). The industry is shifting from a cloud-dominated landscape to one shaped by the widespread integration of generative artificial intelligence (AI). This period features an unprecedented infrastructure investment supercycle. Global capacity is projected to double to 200 gigawatts (GW) by 2030. This growth will require $3 trillion in total investment across real estate and IT equipment. In the public REIT sector, market dynamics in 2026 are defined by a persistent supply-demand imbalance. Vacancy rates remain at historic lows, and pricing power favors operators.
Growth in 2026 is driven by the “second act” of AI, specifically the massive scaling of inference workloads. While 2025 focused on training large language models (LLMs), industry data shows that inference, the execution of these models, will become the main driver of demand by 2027. This trend requires a more geographically distributed and lower-latency network of facilities. As a result, capital is flowing to Tier 2 markets and power-advantaged regions. Traditional hubs, such as Northern Virginia, face insurmountable grid constraints.
However, the sector faces significant structural risks.
The “power expectation gap” between developers and utilities has widened to nearly two years. The electrical grid is struggling to support an IT load capacity projected to reach 150 GW in the United States by 2028. Local opposition is resurging, and the removal of “by-right” zoning in mature markets has extended permitting timelines and increased delivery risks. Despite these challenges, sector leaders like Digital Realty Trust, Equinix, and Iron Mountain maintain robust financial performance. Their success is supported by record preleasing rates in the mid-70% range and strong pricing reversions.
Large-scale capital market activity is also reshaping the strategic landscape. Brookfield’s $1.2 billion acquisition of Peakstone Realty Trust illustrates how institutional investors are securing industrial and warehouse assets to reposition them into the digital infrastructure ecosystem. Over the next decade, competitive advantage will hinge on onsite power solutions, modular construction techniques, and quantum-resilient security. This report provides a detailed analysis of these forces, offering a balanced perspective on both the growth potential and the challenges facing data center REITs in 2026 and beyond.
What Are Data Center REITs?
Data center REITs are investment vehicles that own, operate, and lease physical facilities for servers, storage devices, and networking equipment. These assets power the global digital economy. In 2026, the traditional view of data centers as simple “shells” for IT hardware has evolved. Investors now see them as high-density, power-intensive infrastructure platforms. The primary value driver is no longer square footage but available power capacity, measured in megawatts (MW) or kilowatts (kW).
The Functional Architecture of Digital Infrastructure
The core operations of a data center REIT focus on providing a controlled environment with redundant power sources, advanced cooling systems, and strong physical and cybersecurity measures. These facilities fall into several categories:
Wholesale Data Centers
These are large-scale facilities leased to a single tenant or a few hyperscale cloud providers (e.g., Microsoft, Google, AWS). Leases typically last 10–15 years and demand massive power.
Colocation Facilities
These are multi-tenant spaces where enterprises rent smaller amounts of space and power. Companies like Equinix provide high-density interconnection services that let tenants exchange data directly.
Edge Data Centers
These smaller, distributed facilities sit closer to end-users to deliver ultra-low latency for real-time applications, including autonomous vehicles and AI inference.
In 2026, these categories are increasingly blurred. Hyperscalers use smaller distributed nodes, and colocation providers build large AI-optimized campuses. The REIT structure remains appealing because it requires distributing at least 90% of taxable income to shareholders. This setup offers growth from the tech sector and income from real estate.
The Economic Model: From Rent to Power-Based Income
Income models for data center REITs now focus on power consumption. Investors and analysts effectively “underwrite megawatts,” quoting lease rates in dollars per kilowatt per month ($/kW/month). Power availability is the primary supply constraint. As grid limits tighten, a REIT’s ability to secure and deliver power becomes its most valuable asset, strengthening pricing.
Data Center REIT Market Profile (2026)
- Typical Lease Term: 5–15 years, often with annual rent escalators.
- Occupancy Levels: Historical highs, with vacancy rates near zero in Tier 1 hubs.
- Valuation Metric: Investors focus on Price to Adjusted Funds From Operations (P/AFFO) and premiums to Net Asset Value (NAV).
- Operating Margin: Highly efficient, with EBITDA margins often above 50%.
The shift toward AI infrastructure has increased sector capital intensity. Modern facilities require liquid cooling and high-density racks exceeding 50 kW per rack, leading to high upfront costs. Despite this, the sector remains popular with institutional investors. REITs trade at high premiums to NAV, averaging 18.9% in early 2026, while industrial and office sectors trade at double-digit discounts.
Growth Drivers in 2026
Data center REIT expansion in 2026 is driven by a mix of technological shifts and capital market behavior. The initial wave of cloud migration provided a steady foundation, but the current surge is fueled by the rapid industrialization of artificial intelligence and the restructuring of global data traffic patterns.
The Inference Pivot: AI’s Second Act
The key growth catalyst in 2026 is the shift from AI model training to model inference. In the early 2020s, demand was driven by “training,” where hyperscalers and AI labs such as OpenAI or Anthropic processed massive amounts of data to build foundational models. Training is power-intensive but relatively latency-insensitive, so it can occur in remote areas with cheap energy.
By 2026, the focus has shifted to inference, the real-world use of these models. Inference occurs whenever a user prompts an AI tool for a response. Unlike training, it requires low latency for a seamless experience. This shift has major implications for REITs:
- Widespread Demand: Industry surveys show that 98% of insiders expect inference to drive future demand.
- Geographic Reallocation: Inference workloads are moving compute closer to end-users, boosting demand in metropolitan markets and Tier 2 cities.
- Silicon Evolution: About 72% of respondents expect inference to run on newer, more efficient silicon, requiring data centers with higher rack densities and specialized cooling.
AI workloads are projected to account for half of all data center traffic by 2030. REITs that meet technical requirements for inference are seeing record preleasing activity, often reaching mid-70% rates before completion.
The Infrastructure Supercycle and Hyperscale Leasing
The global data center sector is in an infrastructure investment supercycle. Nearly 100 GW of new capacity is expected between 2026 and 2030, doubling the global footprint. This expansion requires around $1.2 trillion in real estate asset creation. Tenants will spend an additional $1-2 trillion on IT equipment, including Nvidia’s latest GPU clusters.
Hyperscale cloud providers, known as the “Magnificent 7,” drive this growth. In 2026, they pursue a dual strategy: building massive campuses and leasing record amounts from public REITs to meet delivery timelines. For example, Keppel DC REIT reported that hyperscalers contributed 69% of its rental income in early 2026. REITs offer a speed-to-market advantage, securing power and navigating local permitting faster than new entrants.
Regional Expansion and the Rise of Tier 2 Markets
Primary hubs like Northern Virginia, Silicon Valley, and Dublin are reaching capacity limits due to power shortages and regulatory pauses. Growth is shifting to Tier 2 markets and emerging international regions.
2026 Growth Dynamics by Market
- Emerging U.S. Markets: The Sun Belt (Alabama, Mississippi, Georgia) and regions with ample natural gas see high growth.
- Texas (ERCOT): Its independent grid and faster interconnection timelines make it a major beneficiary.
- Taiwan: Valued at $1.97 billion in 2025, projected to reach $5.72 billion by 2031 (19.44% CAGR), supported by subsea cable investments.
- APAC Region: Capacity expected to grow from 32 GW to 57 GW by 2030, a 12% CAGR.
- Middle East/EMEA: Digital transformation and demand for “Sovereign AI” clouds drive a 10% CAGR.
In the U.S., developers are trading network proximity for “speed to power,” prioritizing sites where 300 MW+ can be delivered in under 36 months.
Institutional Capital and Consolidation Activity
Capital markets in 2026 reflect strong conviction in the data center thesis. More than $900 billion in investment is required through 2029. Financing structures are evolving to support this demand. Asset-backed securities (ABS) backed by data centers are projected to grow to $30–40 billion in 2026, providing lower-cost funding for stabilized assets.
This capital abundance has sparked M&A activity. Around 70% of industry respondents expect M&A to become more attractive in 2026. A notable deal is Brookfield Asset Management’s $1.2 billion acquisition of Peakstone Realty Trust. The all-cash purchase secured a portfolio of industrial outdoor storage and warehouse assets critical for data center logistics and future redevelopment. These moves show the strategic importance of acquiring “power-ready” industrial land.
Structural Risks and Sector Challenges
Despite strong growth projections, the sector faces vulnerabilities that could affect timelines and profitability. Risks include physical infrastructure limits and a more complex regulatory and social environment.
The Power Boundary: Grid Fragility and the Expectation Gap
Power availability has become the main constraint on industry growth. In 2026, the grid struggles to meet AI-driven demand. Key issues include:
- The Interconnection Queue: Tying new projects to the grid can take years, often longer than construction itself.
- The Power Expectation Gap: Developers and utilities are misaligned. Utilities estimate “time-to-power” 1.5–2 years longer than operators expect, particularly in Northern Virginia and Atlanta.
- Skyrocketing Demand: U.S. IT load capacity is projected to nearly double by 2028, from 80 GW to 150 GW, stressing legacy transmission infrastructure.
Energy affordability has become a political issue, with data centers cited as major drivers of electricity use. For REITs, securing power commitments can be more valuable than owning land, shifting valuations to “dollars per kilowatt.”
Regulatory Shifts and the Death of “By-Right” Zoning
The regulatory environment has become restrictive. The elimination of “by-right” zoning now requires public hearings and board votes for data center applications.
Consequences include:
- Increased Permitting Costs: Longer approvals and legislative advocacy raise upfront costs.
- Local Moratoriums: Rural and emerging markets impose construction bans to preserve community character.
- Home Rule vs. Dillon’s Rule: Risks vary by state. Home Rule states like Georgia allow local legislation, increasing pushback. Dillon’s Rule states like Virginia follow state frameworks, providing more clarity.
Communities also express concern over noise from cooling equipment (capped at 55 decibels) and visual impacts of large substations.
Supply Chain Resilience and Construction Inflation
Global construction costs are rising at a 7% CAGR, with costs per MW averaging $11.3 million in 2026. Drivers include:
- Technical Requirements: AI-ready facilities require liquid cooling and high-density power distribution.
- Labor Shortages: An estimated 499,000 new construction workers are needed, raising costs and causing delays.
- Critical Component Lead Times: High-voltage transformers and backup generators face extended delivery times.
REITs are assuming more construction risk, providing completion guarantees and covering cost overruns previously borne by tenants.
Data Sovereignty and Geopolitical Risk
Data is a strategic national asset. Data sovereignty laws now influence geographic strategies.
Geopolitical Uncertainty: 96% of investors report that tensions affect investment plans, prompting REITs to diversify footprints to reduce regional risks.
Sovereign AI Clouds: Europe and the Middle East are adopting these clouds to meet privacy regulations.
Extraterritorial Reach: U.S. CLOUD Act concerns influence vendor selection.
Financial Performance and Market Metrics
The financial health of data center REITs in 2026 shows contrasts. Property fundamentals are among the strongest in real estate. However, stock prices have been volatile as markets factor in high interest rates and the concentration risk of the “Magnificent 7.”
Valuation and Capital Market Behavior
Data center REITs trade at a significant premium compared to the broader market and other real estate subsectors. As of early 2026, the median NAV discount for REITs overall is roughly 12–18%. In contrast, data centers trade at an 18.9% premium to NAV. This reflects the high market value of existing “powered” shells, which are considered irreplaceable in a capacity-constrained environment.
2026 Price to NAV by REIT Sector
- Data Centers: +18.9% premium
- Healthcare: +30–58% premium
- Office: -29.9% discount
- Industrial: -26.7% discount
- Self-Storage: -9.0% discount
Despite these premiums, the average P/FFO (Price to Funds From Operations) for data center REITs contracted slightly to 22x in late 2025. Investors rotated into other sectors, including healthcare and advertising. The outlook for 2026 points to a re-expansion of these multiples as record pricing reversions, up to 45% in some APAC markets, begin to improve the bottom line.
Comparative Performance: Digital Realty vs. Equinix
The two largest players, Digital Realty Trust (DLR) and Equinix (EQIX), serve as benchmarks for sector performance. Both underperformed relative to the S&P 500 in 2025. Analysts maintained “Overweight” ratings heading into 2026, citing pricing strength and new lease signings.
Comparison of Digital Realty (DLR) vs. Equinix (EQIX)
- Market Cap (Early 2026): DLR $62.17 billion; EQIX $93.95 billion
- Annual Revenue: DLR $6 billion; EQIX $9.2 billion
- P/FFO Ratio: DLR 50x; EQIX 69x
- Dividend Yield (est.): Both 2.3%
- EBITDA Margin: DLR 21.4%; EQIX 14.7%
While Equinix has higher market cap and revenue, Digital Realty is often viewed as having more immediate upside. It benefits from $126 million in annualized rent expected to commence in early 2026. Equinix, however, leads in profitability metrics, including return on equity (9.6%) and return on assets (3.5%).
The Iron Mountain Transformation
Iron Mountain (IRM) has emerged as a top-performing data center REIT in 2026. After moving from a legacy records-management focus, its data center segment grew 30% in 2025. Data center revenue is projected to exceed $1 billion in 2026, supported by a 400 MW backlog and a high-margin (50%+) EBITDA profile. IRM’s cross-selling to its 240,000 existing customers allowed it to increase its dividend by 10% in 2026, reflecting confidence in its transformation strategy.
Global REIT Performance and M&A
The broader REIT market in 2026 is shifting toward “non-core” sectors, including data centers. These assets are expected to outperform traditional office and residential sectors in near-term growth. Consolidation continues, with 70% of respondents expecting M&A activity to rise. High interest rates have favored large, listed REITs, allowing them to acquire smaller, distressed platforms at discounts to replacement cost.
Case Studies: Leading Data Center REITs in 2026
The strategies of the sector’s top companies illustrate how data center REITs are capturing the AI opportunity while managing structural challenges.
Digital Realty Trust (DLR): Scale and Strategic Joint Ventures
Digital Realty has maintained 20 years of top-line growth by focusing on large-scale AI requirements. In 2026, DLR prefers “full-building configurations” in new developments rather than subdividing space. This approach serves AI startups and GPU-as-a-service providers who pay premiums for scale.
A key part of DLR’s strategy is using external capital. The company contributes assets to joint ventures, such as its U.S. Hyperscale Fund. This allows DLR to recycle capital into new developments while retaining management fees and a portion of the upside. This model helps DLR maintain a financial health score of 2.62 despite the capital-intensive nature of its 500-MW+ AI campus developments.
Equinix (EQIX): Interconnection and the SDN Edge
Equinix leads the colocation and interconnection market. In 2026, it focuses on software-defined networking (SDN) solutions, enabling enterprise clients to update interconnectivity in real-time. This is critical for hybrid-cloud environments, where data moves between on-premise servers and multiple public clouds.
Equinix manages construction risk more conservatively than before. It often secures long-term customer commitments of 10-15 years or relies on backstops from hyperscale partners to ensure project financing. With an anticipated 2026 price target of $950–$965, Equinix remains a top choice for investors seeking a pure-play interconnection leader.
Iron Mountain (IRM): Leveraging the Legacy Base
Iron Mountain’s strategy leverages its global footprint of 1,300+ facilities and its trusted status with 90% of the Fortune 1000. By late 2025, IRM’s data center revenue reached a $204 million quarterly run rate, with management aiming for it to dominate consolidated profitability.
The “Matterhorn” restructuring program concluded in 2025, improving operating leverage and reducing net lease adjusted leverage to 4.9x, the lowest since the 2014 REIT conversion. In 2026, IRM is emphasizing its Asset Lifecycle Management (ALM) business, which grew 63% in 2025. This unit manages decommissioning and remarketing of old server components, a high-growth area as AI hardware replacement cycles shorten.
Keppel DC REIT (KDCREIT): APAC Dominance and Singapore Resilience
Keppel DC REIT is a key gateway to the APAC digital economy. In 2026, it consolidated ownership of two major Singapore data centers, achieving 100% control of Keppel DC Singapore 3 and 4. This increased assets under management to S$5.9 billion and raised hyperscale client exposure to nearly 70% of rental income.
KDCREIT benefits from Singapore’s resilient market, achieving a 45% positive rental reversion in FY25. With a gearing ratio of 35.3%, the REIT is well-positioned for expansion in Tokyo, Seoul, and established European hubs.
Brookfield and Peakstone: The “Industrial to Digital” Pivot
Brookfield’s $1.2 billion acquisition of Peakstone Realty Trust in February 2026 reflects a trend of industrial-to-digital repositioning. Peakstone had shifted from an office-heavy portfolio to focus on industrial outdoor storage (IOS) and warehouses.
For Brookfield, the deal was strategic. Industrial sites serve as staging grounds for data center infrastructure. The flexible, lower-cost nature of outdoor storage meets the logistics demands of data center fit-outs. This acquisition allows Brookfield to expand its industrial platform while securing land that could be converted to digital use once power becomes available.
Future Trends and Strategic Shifts
As the sector moves into the second half of the decade, REIT strategies are increasingly defined by a “speed eclipses efficiency” mindset. Winners will be those who innovate around grid limitations and the evolving needs of AI.
Onsite Power and Energy Autonomy
A major strategic trend is the shift toward onsite generation. By 2030, nearly one-third of data centers are expected to operate fully onsite-powered campuses. This is now a core component of long-term power strategy rather than a temporary solution.
- Fuel Cells: Emerging as a leading technology due to shorter lead times and lower local emissions, supporting faster permitting.
- Small Modular Reactors (SMRs): 2026 marks a pivotal year for financing “firm capacity,” with institutional focus on nuclear technologies to support gigawatt-scale AI facilities.
- Hydrogen and Natural Gas: In deregulated markets, operators adopt “Bring Your Own Power” (BYOP) strategies, using natural gas as a permanent source while waiting for grid upgrades.
The Edge and AI Inference Distribution
As inference workloads scale, the “edge” is a $19 billion market in 2026. Hardware, including AI-optimized servers, storage, and networking, represents 63% of this market. The rollout of 5G networks enables real-time analytics in healthcare, manufacturing, and smart cities. REITs are increasingly building smaller, distributed “hyperscale nodes” near major metros to reduce latency for enterprise AI applications.
Modular Construction and Design Flexibility
To counter rising costs and labor shortages, the modular data center market is expected to reach $85 billion by 2030.
- Prefabricated Facilities: Modular edge sites reduce deployment time and costs, allowing rapid urban scale-out.
- Future-Proofing: Even non-AI tenants now demand flexible facilities that can support high-density racks and future liquid cooling upgrades.
Quantum Resilience and Cybersecurity
Quantum computing is creating a new frontier in data center security. By 2026, 47% of respondents expect quantum technology to disrupt conventional encryption within 2–10 years. Forward-looking REITs are adopting quantum-resilient cybersecurity infrastructure to protect long-term facility integrity.
Sustainability and the “Affordability” Pivot
In 2026, energy affordability replaces “energy transition” as the central theme. Data centers must balance massive power needs with sustainability pressures. Operators are implementing diverse generation strategies and increasing waste heat recovery. Some European cities now require data centers to redirect heat into municipal heating networks.
Conclusion and Recommendations
The 2026 data center REIT sector has favorable growth potential but faces unprecedented complexity. Success is no longer just about leasing space—it requires managing a convergence of power, policy, and technology.
Nuanced Conclusions
The industry is experiencing a historic divergence. Companies that master strategy, integration, and execution are seeing outsized success. Those that fail to adapt to AI technical requirements fall behind despite market hype. The 18.9% valuation premium over NAV reflects the scarcity of power-ready capacity, but it is sustainable only if operators continue to deliver new supply.
The “power expectation gap” is the greatest threat to sector growth. Widening delivery timelines may trigger market distress as smaller players struggle with capital intensity and regulatory hurdles. The shift to inference is reshaping the geographic map, favoring REITs in Tier 2 markets and metro edges.
Actionable Recommendations
For institutional investors and operators, key strategic imperatives include:
- Prioritize Energy Autonomy: Invest in behind-the-meter generation and onsite storage. “Speed to power” now outweighs connectivity in site selection.
- Embrace Geographic Diversification: Focus on power-advantaged U.S. states like Texas and the Sun Belt, as well as emerging APAC hubs such as Taiwan.
- Invest in Modular and Flexible Design: Integrate modular components to counter labor shortages and construction inflation. Future-proof all builds for high-density liquid cooling.
- Mitigate Local and Regulatory Risk: Engage with communities and utilities years before acquisition. With “by-right” zoning disappearing, community support is as critical as infrastructure capability.
- Focus on Post-Training Infrastructure: Optimize portfolios for low-latency delivery and secure data boundaries to serve the growing demand for Sovereign AI clouds.
In summary, data center REITs in 2026 remain a compelling way to participate in the AI expansion. The era of easy growth is over. The next decade will favor operators who turn infrastructure, regulatory, and technical challenges into durable competitive advantages.
