The conventional case for a data center sale-leaseback has always rested on a straightforward capital efficiency argument: unlock the balance sheet, redeploy capital into core business operations, and retain operational continuity through a long-term lease. That logic has not changed. What has changed, fundamentally and perhaps permanently, is the asset being valued.
In 2026, a data center with secured grid capacity is not merely a real estate asset. It is an energy infrastructure asset — and the market is pricing it accordingly.
Understanding this distinction is the starting point for any serious sale-leaseback analysis today.
Power as the Primary Valuation Driver
The data center industry crossed a structural threshold in 2025 that is reshaping how buyers underwrite existing assets. The primary constraint on new supply is no longer construction labor, equipment lead times, or regulatory permitting — it is power.
According to Sightline Climate, 190 GW of large data center and AI factory capacity has been announced since 2024, representing 777 individual projects. Of the 16 GW slated for delivery in 2026, only 5 GW is currently under construction. The remaining 11 GW sits in the announced stage — subject to power availability, interconnection approvals, and utility co-development timelines that routinely extend three to five years.
The interconnection queue is the decisive variable. As Enverus documented in its 2026 Interconnection Queue Outlook, grid congestion and study delays persist despite FERC reform efforts, with outcomes increasingly dependent on regional ISO market structure rather than simply queue position. RMI analysis puts the average time from interconnection request to commercial operation at nearly five years — more than double the sub-two-year timeline that prevailed in 2008. More than 2.2 terawatts of generation and storage projects currently sit in queue, yet only a fraction reach commercial operation annually.
For an enterprise that already owns a powered, grid-connected data center, this bottleneck is not a liability. It is leverage.
What Buyers Are Actually Paying
Lease rates have repriced materially, and the repricing is most acute at scale. CBRE’s North America Data Center Trends H1 2025 report documented pricing increases of up to 19% for 10 MW+ requirements in Silicon Valley, 15.4% in Chicago, and 13.8% in Northern Virginia — all driven by the same underlying constraint: contiguous, power-ready capacity is unavailable at any development timeline that matters to an enterprise buyer.
Quoted wholesale lease rates for 250+ kW requirements in Tier-1 U.S. markets ranged from $120 to $150 per kW/month in 2024–2025, with DatacenterHawk analysis noting tenants who signed deals seven to eight years ago at $75/kW are now facing renewal conversations at $130–$150/kW — in markets where the landlord holds nearly all negotiating leverage. CBRE projects pricing comparable to 2011–2012 peaks above $200/kW/month is increasingly plausible as vacancy across primary markets sits at 1%, the second consecutive year at that floor.
On the investment side, development yield benchmarks have firmed. The Alantra 2024 Data Centre M&A Review documented development yields for 15-year NNN leases with major hyperscaler tenants running 7.5%–8.5%, with exit cap rates in the 5.25%–6.0% range for Class-A assets. For AI-infrastructure-grade facilities with power certainty, the spread between development yield and exit cap reflects the scarcity premium buyers are prepared to underwrite.
CNBC reporting on 2025 deal activity confirmed that global data center M&A and investment hit a record $61 billion in 2025, with analysts explicitly noting that power supply constraints “enhance the value of existing facilities” and that companies not treating data centers as core business assets are increasingly likely to surface those assets into the market. That trend will accelerate in 2026.
The Brownfield Premium: Why Existing Assets Outcompete Greenfield
New development is not simply expensive — it is slow in ways that matter more than price in a power-constrained environment. Greenfield campus delivery at $7–$10 million per MW for core/shell, before tenant fit-out costs of an additional $3–$5 million per MW, requires interconnection queue position, utility co-development agreements, and construction timelines that rarely compress below 36 months even in favorable markets.
Brownfield retrofits and existing asset acquisitions are gaining momentum because they bypass the hardest part of the development timeline: securing the power allocation. An existing facility with a live grid connection, established utility relationships, and on-site power infrastructure represents a years-ahead position relative to any announced greenfield project that has not yet broken ground.
Sightline Climate’s analysis estimated 30–50% of the 2026 announced pipeline may not materialize — projects that will be delayed or canceled precisely because the power was never secured. For the owner of an operating facility in a supply-constrained market, every project that fails to come online is a data point that reinforces the scarcity premium on their asset.
The hyperscaler demand signal further validates this logic. Mintz’s March 2026 market review noted that 35 GW of data center capacity is currently under construction in North America, with 60% fully pre-leased — and that SASB (single-asset, single-borrower) stabilized loans exceeded $11 billion in 2025, a 3x increase over 2024 levels. Capital is aggressively seeking stabilized, operational assets with creditworthy tenants, precisely the profile a sale-leaseback transaction is designed to create.
The Advisory Framework: Sell, Lease Back, or Hold
The decision to pursue a data center sale-leaseback valuation in 2026 is not binary. The framework we apply at StackedAI evaluates four variables before a recommendation is issued.
1. Power Certification
The single most important input is the nature of the facility’s power position. An asset with a dedicated utility feed, contracted capacity, and on-site redundancy that has survived load growth scrutiny commands a materially different valuation than a facility operating at the margin of available grid capacity. Quantify the contracted MW position, the interconnection agreement status, and the utility’s stated capacity headroom before any buyer conversation.
2. Lease Structure and Credit Quality
Sale-leaseback transactions generate valuation multiples that are directly sensitive to lease term and tenant credit. A 15-year NNN structure with an investment-grade tenant (or a creditworthy hyperscaler) will price to a 5.25%–6.0% exit cap. A shorter-term, gross lease structure to an unrated enterprise tenant will price materially wider. The advisory question is whether the leaseback terms can be structured to support the valuation the asset’s power position would otherwise command.
3. Market Position Relative to Announced Pipeline
An existing asset in a primary market — Northern Virginia, Silicon Valley, Chicago — benefits from near-zero vacancy and institutional buyer depth. An asset in a secondary or tertiary market benefits from lower competition for power-enabled development sites, but faces a thinner buyer pool. Cushman & Wakefield’s Americas data center market update documents the shift toward managed growth in established markets — tighter supply, fewer approvals — which reinforces the value of existing permitted and powered assets regardless of geography.
4. Capital Redeployment Opportunity Cost
The sale-leaseback is ultimately a financing decision. The net present value of the leaseback obligation must be compared against the cost of alternative capital — debt at current rates, equity dilution, or foregone investment in core business operations. With net lease investment volume running 24% higher year-over-year through Q3 2025 according to W. P. Carey’s 2026 outlook, buyer demand is creating favorable execution conditions for well-structured transactions. The question is whether the seller’s cost of leaseback capital is cheaper than the alternatives.
What the Numbers Tell You
A 10 MW data center in a primary market with contracted power, a Tier III configuration, and existing hyperscaler tenancy is a materially different asset than it was three years ago. Applying current market parameters:
- Gross asset value: At $7–$10M per MW replacement cost, a 10 MW facility represents $70–$100M in replacement-cost-equivalent value — before the scarcity premium for secured power is applied.
- Income capitalization: At $130–$150/kW/month lease rates on a 10 MW NNN structure, annualized NOI approaches $15.6–$18M. Capitalized at a 5.5% exit cap rate, that implies a gross asset value of $284–$327M.
- The power premium: The delta between replacement cost and income-capitalized value is the power premium — the market’s explicit pricing of the difficulty in replicating the asset’s grid position. In a market where 11 GW of announced projects cannot get to construction, that premium is structural, not cyclical.
These figures represent a range, not a comp. Actual transaction pricing depends on lease structure, tenant quality, market depth, and the specific utility relationship. But the directional message is clear: enterprise owners of operating data centers are systematically underestimating what their assets are worth in a market where power availability has become the binding constraint.
The Hold Case: When the Sale-Leaseback Math Does Not Work
A sale-leaseback is not always the right answer, even when the asset is valuable. Three conditions favor holding over transacting.
Tenant optionality: If the facility is internally operated and supports a business unit where AI or compute workloads are expected to scale materially, the opportunity cost of locking in long-term leaseback terms may exceed the benefit of capital release. Facilities in markets where moratorium proposals are advancing — New York, Michigan, Virginia, Oklahoma among others — represent defensive holds where development alternatives may be foreclosed entirely.
Below-market lease risk: If existing internal operations would require below-market lease rates to justify a leaseback, the buyer pool may demand a credit enhancement that erodes the net proceeds.
Upgrade capital intensity: An aging facility requiring significant capital investment to meet AI-workload density requirements — liquid cooling retrofits, power density upgrades — may require the owner to retain the asset and fund the improvement before a transaction will price at the premium the power position otherwise supports.
The Advisory Takeaway
The data center sale-leaseback calculus has changed. The asset being valued in 2026 is not principally a building — it is a power position, a grid interconnection, and an operational timeline advantage that cannot be replicated at any price on a reasonable schedule.
Global data center M&A hit $61 billion in 2025 and activity is expected to intensify in 2026 as hyperscalers continue to externalize capital requirements and the 11 GW stuck in the announced pipeline demonstrates that new supply is not coming fast enough to relieve pressure on existing assets. Net lease investment volume is up 24% year-over-year. Buyers are capitalized and active.
Enterprise owners who have not recently stress-tested the market value of their owned data center infrastructure against current transaction benchmarks are operating on stale information. The analysis is worth running — not because a sale-leaseback is always the right outcome, but because the decision should be made with an accurate picture of what the asset is actually worth.
That picture has changed considerably. The market is offering a window that the interconnection queue math suggests may not remain open indefinitely.
StackedAI provides enterprise data center advisory across M&A, capital markets, and infrastructure strategy. For a confidential valuation assessment of your owned data center assets, contact the advisory team.