Should You Retain Digital Realty Stock in Your Portfolio Now?

Digital Realty is showing healthy leasing momentum and strategic growth investments—signing $162 million of new leases in Q3 2025 (including $76.1M >1MW, $64.9M 0–1MW and $19.6M interconnection bookings), operating 311 data centers across 50+ metros with >5,000 customers, and advancing a development pipeline of 10.2M sq ft active plus 4.8M sq ft held for future development. The company reported ample liquidity with $3.30B cash, ~$3.20B available on revolving credit, recent land buys ($49M LA site supporting 32 MW; $18M for two parcels), and a 2.75% weighted average coupon, but carries $18.2B total debt, a net debt/adjusted EBITDA of 4.9x, fixed charge coverage of 4.6x and a 26.4% YoY rise in interest expense—heightening sensitivity to pricing pressure and competition even as the Zacks 2025 FFO/share consensus sits at $7.35. Notable product/partner momentum includes a November 2025 collaboration with NVIDIA at the Manassas campus to support AI factory infrastructure.

Analysis

Market structure: Digital Realty (DLR) is a clear beneficiary of AI/cloud secular demand — $162m of new leases in Q3 (>$76m in >1MW deals), a 10.2M sq ft active development pipeline and a recently publicized NVIDIA partnership meaningfully increase its addressable revenue per campus. Winners: hyperscalers, NVDA (infrastructure buyers) and large-scale colos with capital (DLR). Losers: smaller/new entrant data-center builders and margin‑sensitive regional operators facing aggressive price competition.

Risk assessment: Leverage (total debt $18.2B, net debt/adjusted EBITDA 4.9x) and interest expense (+26% YoY) are the dominant financial risks; liquidity ($3.3B cash + $3.2B revolver) buys 12–24 months of runway assuming stable leasing but not a demand shock. Tail events include a 20–30% cutback in AI capex by hyperscalers, faster-than-expected supply delivery (oversupply from ~10M sq ft pipeline industry-wide) or a credit shock that widens DLR spreads >200bps.

Trade implications: Expect continued rent/term compression in competing micro-markets; DLR’s scale, land ownership (LA 32MW land buy) and investment-grade tenant mix justify a modest structural premium but not full insulation from cyclical weakness. Cross-asset: worsening credit or rate re-pricing would push DLR equity down while widening its bond CDS; NVDA equity and select long-dated tech hardware names are second-order beneficiaries.

Contrarian angles: Consensus underestimates execution/capex risk — building AI-ready campuses (power, PPA, cooling) compresses near-term FCF even if long-term ARR grows. If rates fall or NVDA-led pilots scale into multi-facility deals within 6–12 months, DLR could rally 15–30% from current levels; conversely, a 100–150bp rise in corporate spreads would be a fast catalyst to re-rate downward.

Key Decisions for Investors

• Establish a 2–3% long position in DLR over the next 1–3 weeks (buy shares), target a 12‑month total return of +20% and set a hard stop-loss at -12%. Hedge the position by buying a 6‑month 10% OTM put (or a protective collar) sized to cap downside to ~10% of position value.
• Allocate 1% portfolio to NVDA via a 6‑month 10/25% OTM call spread (debit) to capture continued infrastructure spend tied to AI deployments; exit if NVDA underperforms the Semiconductor ETF by >8% over 3 months or if GPU cycle commentary turns negative.
• Pair trade: size 2% long DLR vs 2% short CUZ (Cousins Properties) to express preference for AI‑exposed colo scale over traditional office/residential REIT growth; rebalance or close after 6–12 months or if the DLR/CUZ relative return narrows by >10%.
• Reduce exposure to DLR corporate credit: trim direct DLR bond holdings by 50% and buy protection (5‑year CDS or long DLR CDS ETF) if the 5‑year DLR spread widens >75bps from current levels or exceeds 200bps over Treasuries — sized 0.5–1% portfolio as tail‑risk insurance.