Digital Realty Trust, Inc. delivered a record-breaking year in 2025, with full-year Core FFO per share reaching $7.39, a 10% increase over 2024, and Q4 Core FFO of $1.86, up 8% year-over-year. The company achieved its second consecutive year of over $1 billion in total bookings, ending with a record backlog of nearly $1.4 billion at a 100% share, driven by robust hyperscale demand exceeding $800 million and record performance in the zero to one megawatt plus interconnection segment. Strategic highlights include the successful launch of a private capital fund with over $3.2 billion in commitments and the expansion of PlatformDigital to 56 markets. Looking ahead, management provided 2026 guidance for Core FFO per share of $7.90 to $8.00, representing nearly 8% growth at the midpoint, underpinned by strong visibility from the backlog despite a power-constrained supply environment.
| Metric | Value | Change |
|---|---|---|
| Q4 2025 Core FFO | $1.86 per share | +8% year-over-year |
| Full Year 2025 Core FFO | $7.39 per share | +10% year-over-year |
| Q4 2025 Bookings (100% share) | $400 million | Top-five quarter in company history |
| Full Year 2025 Bookings (100% share) | $1.2 billion | Nearly 70% above the average of the preceding five-year period |
| Total Backlog (100% share) | $1.4 billion | Record high |
| Zero to One MW+ Interconnection Bookings | $340 million (Full Year) | +35% year-over-year |
| Same Store Cash NOI Growth (Q4) | 8.6% | +8.6% year-over-year |
| Net Debt to EBITDA (Leverage) | 4.9x | Well below 5.5x target |
| Development Pipeline | $10 billion | Record level |
Management emphasized the critical shift toward AI inference workloads as a primary demand driver for 2026. They highlighted the early adoption of their 'private AI exchange platform,' which allows enterprises to connect compute, data, and models privately. This strategic focus on inference is supported by their 'connected campus approach' and high-density colocation offerings, positioning DLR to capture enterprise demand for low-latency, secure AI infrastructure in tier-one metros where 'sovereignty matters most.'
The evolution of DLR's funding strategy through the launch of an inaugural closed-end fund marks a significant capital allocation pivot. By raising over $3.2 billion in LP equity commitments, DLR is effectively leveraging private capital to fund hyperscale development, thereby recycling public capital and reducing balance sheet strain. This 'dual-engine' approach allows them to support the heavy capital requirements of hyperscale capacity while maintaining a strong liquidity position of nearly $7 billion.
Power availability has emerged as the single most critical competitive moat. Management repeatedly noted that 'power became the industry's primary constraint' and that hyperscalers are prioritizing operators with 'verified visibility into the future supply of power.' DLR's 'five-gigawatt power bank' and long-term track record in core markets provide a distinct strategic advantage, allowing them to secure and deliver capacity on predictable schedules while competitors struggle with grid interconnection delays.
The 'zero to one megawatt plus interconnection' segment is increasingly driving growth and diversifying revenue. This segment posted a record $340 million in bookings for 2025 (up 35% year-over-year), fueled by the expansion of PlatformDigital to 31 countries and 56 markets. The network effects of Service Fabric, which now enables access to over 300 cloud on-ramps, are creating a sticky ecosystem that attracts new logos (nearly 600 added for the second consecutive year) and drives recurring interconnection revenue.
Rising costs and execution risks associated with the power-constrained environment pose a threat to margins. Management acknowledged that 'this race for scaling critical digital cloud computing... comes with a cost,' specifically citing labor and build costs. Furthermore, the refinancing of €1.4 billion in Eurobonds at rates 160 basis points higher than the redeemed debt will create a 'modest interest expense headwind starting in 2026,' potentially pressuring net income.
Supply chain and labor bottlenecks are intensifying, which could delay the delivery of capacity needed to capture the current demand wave. During the Q&A, CEO Andrew Power admitted, 'It's getting challenging more challenging by the day' regarding labor. While DLR has secured supply for its current $10 billion pipeline, the tightening labor market introduces execution risk for future projects and could inflate development CapEx beyond the guided $3.25–$3.75 billion range.
The rapid shift in reporting metrics from square footage to power-based metrics (IT load) creates a potential opacity risk for investors. While management argues this aligns better with business operations, the transition—moving from 83.7% occupancy on a square feet basis to 89% on an IT load basis—requires investors to recalibrate historical valuation models. This complexity could mask underlying utilization trends in specific asset classes or make quarter-over-quarter comparisons more difficult until a new baseline is established.
Geopolitical and regulatory risks regarding power consumption are escalating. Management addressed 'growing list of efforts to reduce the impact of the data center industry on consumer electric rates,' including potential requirements for behind-the-meter solutions. While DLR is pushing back against 'nimbyism,' the threat of deprioritization by utilities or stricter regulations in key markets like Northern Virginia or Europe could significantly slow down the 'gradual' arrival of new supply they are forecasting.
Overall: Management exhibited a highly confident and enthusiastic demeanor throughout the call, frequently using superlatives such as 'exceptional execution,' 'record financial performance,' and 'pivotal year.' They appeared poised and direct during the Q&A, offering detailed technical explanations regarding power constraints and AI inference without deflecting difficult questions about costs or competition.
Confidence: HIGH - Management's confidence was evidenced by their specific guidance raises, detailed discussion of a record backlog, and assertive language regarding their competitive positioning in a power-constrained market. They explicitly stated they are 'never been better positioned' and are 'seeing more customers coming called for the same capacity blocks.'
Management established a range of $7.90 to $8.00 per share. The midpoint of $7.95 represents approximately 8% year-over-year growth. This guidance reflects underlying business strength balanced by increased investment spending.
On a normalized and constant currency basis, growth is anticipated to be 'more than 10%'.
Expected to grow between 4% and 5% on a constant currency basis.
Development CapEx net of partner contributions is expected to rise to between $3.25 billion and $3.75 billion. Development yields are expected to remain in the double digits.
Expected to be between 6% and 8%, with upside potentially mitigated by a high mix of zero to one megawatt leases expiring and fixed-rate renewals in the hyperscale portfolio.
Hedging & Uncertainty: Management generally used direct, assertive language regarding past performance and current backlog ('record financial performance,' 'clearly gaining momentum'). However, hedging appeared when discussing future power availability and the timeline for AI inference proliferation. Phrases like 'expected to scale,' 'should improve,' and 'I think we're still a good ways away' indicate uncertainty about the exact timing and magnitude of the next growth phase. They also used temporal hedges regarding supply, noting new capacity will 'continue to arrive gradually,' which softens the expectation for immediate supply relief.
"Power became the industry's primary constraint." - Andrew P. Power, President and CEO
"Inference expected to scale in 2026." - Andrew P. Power, President and CEO
"We are unquestionably taking market share." - Colin McLean, Chief Revenue Officer
"It's getting challenging more challenging by the day." - Andrew P. Power, President and CEO
"We're seeing more customers coming called for the same capacity blocks." - Andrew P. Power, President and CEO
"The total backlog is a better representation of the aggregate demand." - Matthew R. Mercier, CFO
Analyst Sentiment: Analysts were highly engaged and inquisitive, focusing heavily on the sustainability of hyperscale demand, the specific mechanics of the 'inference' growth cycle, and the practical implications of power constraints. Questions from firms like Wells Fargo, Citi, and Oppenheimer were constructive but probing, seeking to differentiate between market share gains and macro demand growth.
Management Responses: Management responses were detailed and transparent, often providing specific market examples (e.g., Northern Virginia, Tokyo, Paris) to illustrate broader trends. They effectively deflected concerns about valuation gaps by highlighting their unique private capital strategy and the 'mix of assets' argument. They maintained a consistent narrative of 'power as a constraint' and 'execution as a differentiator' throughout the session.
Hyperscale Demand & Power Constraints: Analysts sought to understand the longevity of the current hyperscale boom and how power limitations are shaping customer decisions. Management confirmed that customers are looking 'further out on the horizon' for capacity and prioritizing operators with secured power.
AI Inference vs. Training: There was significant interest in the transition from AI training to inference. Management explained that while hyperscalers are currently mixing training/inference in 'cloud zonal markets,' enterprise inference is a 'long tail demand' that will scale in 2026, driving need for dense, low-latency connectivity.
Private Capital & Valuation: Analysts questioned the disconnect between public and private market valuations. Management used this to highlight their new fund strategy, explaining that private valuations often depend on asset mix (land vs. cash-generating) and that their strategy allows them to 'pull in both private and public capital levers'.
Operational Bottlenecks: Questions regarding labor and supply chain constraints were met with honest admissions that 'it's getting challenging,' but management leveraged this to reiterate DLR's competitive advantage due to its scale and long-standing vendor relationships.
Digital Realty Trust, Inc. is executing at a high level in a supply-constrained environment characterized by explosive AI and cloud demand. The company's 'connected campus' strategy and massive 'power bank' provide a defensible moat against competitors who cannot secure grid capacity. Financially, DLR is delivering strong double-digit AFFO growth and has a record backlog of $1.4 billion, ensuring visibility for 2026 targets. The successful launch of a $3.2B private capital fund is a game-changer, allowing them to recycle capital and fund hyperscale growth without over-leveraging the balance sheet (leverage sits at a comfortable 4.9x). Key risks include rising interest expenses and labor/supply chain inflation, but the pricing power (evidenced by 6.7% cash releasing spreads) should help offset these costs. With the shift to AI inference favoring dense, interconnected urban assets—DLR's specialty—the company is uniquely positioned to capture the next leg of growth. The 8% growth guidance for 2026 appears conservative given the backlog, suggesting potential for upside.
The data center industry is facing a structural supply shortage driven primarily by power availability rather than land or materials. Management noted that 'power became the industry's primary constraint,' with new supply arriving 'gradually' due to generation and transmission upgrade delays. This bottleneck is forcing hyperscalers to lease further in advance and prioritize operators with existing power contracts, fundamentally altering the competitive landscape to favor incumbents with established infrastructure.
The adoption of AI is shifting from the training phase to the inference phase, which has different infrastructure requirements. Management stated that 'inference expected to scale in 2026,' driving demand for 'low-latency, secure, and cost-efficient' workflows located closer to data sources and end-users. This implies a long-term, durable demand driver for edge computing and interconnection-heavy data centers in major population centers, rather than just remote hyperscale training facilities.
There is a widening gap between public and private data center valuations, driven by 'severely constrained' supply and demand expected to increase 'two and a half to three times over the next five years.' This environment is allowing established operators to command premium pricing and strong returns on development, with development yields remaining in the double digits despite rising construction costs.