QTS Data Centers (Blackstone)
Develop-own-lease wholesale/hyperscale data centers: long-dated (15+ year, often 15-20 year) triple-net-style leases with contractual escalators to investment-grade hyperscale tenants; capital-intensive build funded via sponsor equity, IG project bonds, CMBS and bank debt at the asset level
QTS is 100% sponsor-owned with no primary fundraising rounds since the 2021 take-private, so the trail is deal-event based: the 2021 point is a disclosed transaction EV; the 2025 point is press framing anchored by the real ~$3B founder-stake purchase. Corroborating direction (not marks): leased capacity ~14-15x since acquisition, +18% appraisal uplift on a refinanced 10-property CMBS pool, and the ~$40B Aligned Data Centers comp (Oct 2025).
Earnings, margins, COGS & capex
QTS discloses no consolidated P&L as a private company. What is verifiable (2025-2026): leased capacity has grown ~14-15x since the Aug 2021 take-private; commissioned capacity grew from ~400 MW to ~3 GW, with ~6 GW of electricity expected at full project completion; the portfolio spans 90+ data centers in 15 US states (operational or under construction) plus sites in the UK, Netherlands, Spain and a planned Milan, Italy campus; the development pipeline grew from ~$1B to $25B+ and is 100% pre-leased to investment-grade technology companies on 15+ year leases with escalators, with a land bank reported to support up to ~$80B of further development. Financing has shifted from sponsor equity toward institutional debt: the Apr 2026 $4.6B green bond (5.700% due 2036, Moody's Baa2) drew ~$12.5B of peak demand and priced ~162.5bp over Treasuries - press hailed it as a landmark project-finance-style IG bond for a single data-center campus; a Jul 2026 ~$2B loan (Project Magnolia, Goldman-led) repays campus debt and funds a distribution to the sponsor. A CMBS appraisal showed one refinanced 10-property pool's value up 18% ($4.75B to $5.62B).
Revenue trend
Margins
wholesale hyperscale leases typically carry high, stable property-level margins; power often passed through
by design; turns on lease commencement
COGS structure
Property-level cost stack: electricity (largest input, substantially passed through; renewables sourced via a 20-year Georgia Power agreement for ~350 MW of new solar supporting the Atlanta-Metro and Suwanee campuses, signed Mar 2022), cooling/water, site staffing (headcount roughly doubled from ~650 at acquisition to 1,300+), maintenance, and property taxes.
Capex
The dominant financial fact. ~35 concurrent development projects (trade press, 2024-2026); Fayetteville GA campus alone (~615 acres, ~7M sq ft planned) absorbed a $4.6B project bond; February 2026 ~$2.05B CMBS refinanced $1.36B of debt across three campuses; UK Cambois campus of 10 data halls on 133+ acres (~1.1 GW at build-out, phase 1 opening ~2028); new US campuses advancing in Alabama, Iowa, and Colorado. Funding mix: BREIT/BIP equity + IG project bonds + CMBS + syndicated loans.
Latest earnings
n/a
None issued; sponsor framing (Blackstone 2026 mid-year update): leased capacity ~15x since acquisition, pipeline fully pre-leased
- Take-private EV (Aug 2021)
- ~$10B including assumed debt, at $78/share
- Leased capacity growth since acquisition
- ~14x (BREIT 2025 year-end letter); ~15x (Blackstone 2026 mid-year update)
- Commissioned capacity
- ~400 MW (2021) to ~3 GW (2025); ~6 GW of electricity at full project completion
- Portfolio
- 90+ data centers in 15 US states, operational or under construction (2026)
- Pre-leased development pipeline
- $25B+, 100% pre-leased, 15+ yr IG leases (BREIT, 2026); grew from ~$1B at acquisition
- Apr 2026 green bond
- $4.6B, 5.700% due 2036, Moody's Baa2 (first-time IG), ~$12.5B peak demand, ~T+162.5bp
- Share of BREIT real estate asset value
- ~20.4% (2025 year-end letter)
Growth drivers
- Hyperscaler AI/cloud capacity procurement — pre-leased, 15+ year investment-grade contracts (Microsoft named at Fayetteville, GA)
- Land-and-power banking in Atlanta/Georgia, Northern Virginia, Dallas-Fort Worth, and Chicago, plus new campuses in Alabama, Iowa, and Colorado and European expansion (UK Cambois ~1.1 GW; Netherlands; Spain; planned Milan, Italy)
- Utility relationships in anchor markets (incl. a 20-yr ~350 MW renewable-energy agreement with Georgia Power) supporting campus power supply and sustainability commitments
- Cheapening capital — landmark IG single-campus project bond (Baa2) with ~$12.5B demand lowered cost of AI build-out debt
- Blackstone ecosystem — BREIT/BIP equity, plus Blackstone Digital Infrastructure Trust (NYSE: BXDC, IPO May 2026) as a listed buyer of stabilized data-center assets - a capital-recycling channel
Bull & bear
QTS is one of the biggest winners of the AI data-center land grab: it locked up land and power before the wave, pre-leases everything to investment-grade hyperscalers for 15+ years, and now finances the build at investment-grade cost. That is a compounding development machine with contracted, bond-like cash flows on the back end.
- Leased capacity ~14-15x since Aug 2021 with a $25B+ pipeline already 100% pre-leased - growth is contracted, not speculative (Blackstone/BREIT, 2025-2026)
- Capital-markets validation: ~$12.5B of demand for a $4.6B Baa2 bond (Apr 2026) means the market prices QTS campuses as quasi-infrastructure, structurally lowering its cost of capital versus peers
- Scale in power: commissioned capacity up from ~400 MW to ~3 GW, with ~6 GW of electricity expected at full project completion, in a market where megawatts, not money, are the scarce input
- Sponsor alignment at scale: QTS is ~20.4% of BREIT's real estate asset value - Blackstone's flagship, with first call on its equity, credit, and political capital
- Multiple exit ramps (re-IPO, stake sales, drop-downs to the now-listed BXDC vehicle) at valuations plausibly multiples of the $10B 2021 basis - press already frames QTS as a ~$60B platform, and a refinanced CMBS pool appraised +18%
QTS is a levered, opaque construction bet on uninterrupted hyperscaler AI capex. Cash flows are back-loaded, debt is front-loaded, financials are undisclosed, and the sponsor has started taking distributions funded with new debt - while community opposition has already killed its largest project.
- No disclosed revenue/EBITDA/leverage since 2021 - the '14-15x capacity' narrative is a sponsor-selected metric, not audited economics; capacity growth is not profit growth
- AI capex cyclicality: a hyperscaler pause in 2027+ would freeze new pre-leasing and strand land/power options bought at peak prices
- Structural leverage creep: project bonds, record CMBS, and a Jul 2026 ~$2B loan that partly funds a sponsor distribution - classic late-cycle dividend-recap behavior
- Concentration and key-person churn: single-tenant, single-campus financings (Microsoft/Fayetteville) mean one renegotiation or delay hits a whole capital structure, and the founder-CEO who built the platform was ousted mid-build-out (Mar 2025)
- Execution and siting risk are live, not theoretical: the 2,100-acre Virginia Digital Gateway project - pitched as the world's largest - was terminated in Jul 2026 after zoning approvals were voided, and ~35 simultaneous builds invite cost overruns in a tight power/labor/equipment market
What it is worth
No official public mark. Anchors: (1) 2021 take-private at ~$10B EV including debt; (2) leased capacity ~14-15x since, with a $25B+ fully pre-leased pipeline (Blackstone/BREIT, 2025-2026); (3) press framing of QTS as a ~$60B platform and Blackstone's ~$3B buyout of the founder's stake (Bloomberg/Bisnow, 2025) - journalistic, not disclosed marks; (4) CMBS appraisal uplift of +18% on one stabilized 10-property pool; (5) closest platform comp: the ~$40B Aligned Data Centers deal (announced Oct 2025); (6) BXDC (NYSE) as a listed vehicle for stabilized hyperscaler-leased assets. Any point estimate is inference, not disclosure.
AI capex digestion + spread widening: new pre-leasing stalls, development spreads compress, leverage taken out via distributions weighs on equity value; stabilized assets keep contracted cash flow but growth premium evaporates
Pipeline delivers on schedule; value compounds with contracted NOI as campuses stabilize; periodic CMBS appraisals (+18% pool uplift is the template) mark assets up gradually; no near-term liquidity event for QTS itself
Uninterrupted hyperscaler pre-leasing + IG-cost debt: platform value re-rates toward/above the Aligned-deal benchmark; sponsor monetizes via re-IPO, stake sales, or drop-downs to the listed BXDC vehicle at a premium to private marks
Directionally, contracted-capacity growth, the ~$3B founder-stake buyout, and the Aligned comp imply QTS's enterprise value today is a large multiple of the $10B basis, but Blackstone has published no current valuation; treat all point figures as not disclosed. QTS also carries substantially more debt than in 2021, so equity value and enterprise value diverge materially.
SWOT
Strengths
- One of the largest fully pre-leased hyperscale development pipelines in the US ($25B+), contracted 15+ years to investment-grade tenants with escalators
- Blackstone sponsorship — effectively unmatched equity depth plus proven access to IG bond, CMBS, and loan markets at scale
- Utility relationships and campus-level power procurement in anchor markets (incl. a 20-yr ~350 MW renewable-energy deal with Georgia Power) in an industry where grid interconnection is the binding constraint
- Land bank across top US data-center metros (Atlanta, Northern Virginia, DFW, Chicago) plus emerging markets (Alabama, Iowa, Colorado) and Europe, reported to support up to ~$80B of further development
- Track record — leased capacity ~14-15x in under 5 years of private ownership; record-setting financings (largest data-center CMBS at $3.46B; landmark IG single-campus project bond)
Weaknesses
- Zero public financial transparency — no consolidated revenue, EBITDA, or leverage disclosure; outside analysis depends on sponsor marketing and deal documents
- Heavy asset-level leverage with rising structural complexity (project bonds + CMBS + loans), including a Jul 2026 loan partly funding a sponsor distribution
- Tenant concentration in a handful of hyperscalers — single-campus financings are explicitly tied to one tenant's demand (Microsoft at Fayetteville)
- Governance churn — founder-CEO Chad Williams was pushed out in Mar 2025 (Blackstone bought his stake for ~$3B); replaced by co-CEOs David Robey and Tag Greason - key-person transition mid-build-out
- Community/political friction is real — QTS terminated the 2,100-acre Digital Gateway project in Prince William County, Virginia (Jul 2, 2026) after courts voided its zoning approvals and years of local opposition
Opportunities
- AI training/inference capacity demand still exceeds deliverable supply; pre-leasing shifts demand risk to tenants before capital is committed
- European expansion (UK Cambois ~1.1 GW campus approved — Netherlands, Spain, planned Milan) opens a second growth vector
- Project-finance-style bond pricing (Forbes, Jun 2026) keeps compressing debt costs for pre-leased AI campuses, expanding development spread
- Liquidity/recycling channels now live — Blackstone Digital Infrastructure Trust (NYSE: BXDC) IPO'd May 2026 ($1.75B at $20/sh, up to $2.0B with greenshoe) to buy stabilized hyperscaler-leased data centers; eventual QTS re-IPO or stake sales possible at marks far above the 2021 basis
- Replicating long-term utility power deals across the land bank to unlock further gigawatts
Threats
- AI capex digestion — if hyperscalers slow or renegotiate leasing, the growth engine (new pre-leases) stalls even though existing contracts hold
- Power scarcity, interconnection queues, and rising electricity politics (ratepayer backlash in Georgia/Virginia) can delay or kill projects
- Local opposition and litigation (the Digital Gateway termination is the live example) constrain the land bank's convertibility
- Rate/credit-cycle risk on a permanently refinancing capital structure; data-center credit spreads could reprice if AI sentiment turns
- Competitive supply wave — Vantage, Aligned, CyrusOne, CloudHQ, Compass, Switch, STACK, plus hyperscaler self-build, all racing for the same power and tenants
Moats, dependencies & bottlenecks
Moats
IG tenants, escalators) Converts development into bond-like contracted cash flow; pipeline 100% pre-leased (BREIT, 2026)
20-yr ~350 MW Georgia Power renewable agreement) The Georgia Power deal is a renewables-supply agreement (Mar 2022), not an interconnection guarantee; campus-level interconnection positions are not publicly quantified
Land bank in top US metros bought pre-AI-wave (reported to support ~$80B of development) Convertibility depends on zoning/community consent - the Digital Gateway termination shows the limit
Structural advantage while AI credit appetite holds; cyclical if spreads reprice
Switching costs for tenants (migration of live hyperscale workloads) Standard for wholesale DC landlords; renewals typically sticky
~35 concurrent builds, workforce doubled to 1,300+) Replicable by the best-funded peers, but few can match current scale
Dependencies
Named anchor tenant of the Fayetteville, GA campus backing the $4.6B single-campus bond; hyperscaler leasing pace drives all incremental growth
AMZN, GOOGL, META, ORCL class tenants) Pipeline is pre-leased to investment-grade global technology companies; specific non-Microsoft tenants not disclosed
Equity sponsor and strategic director; QTS is ~20.4% of BREIT real estate asset value, so BREIT redemption dynamics can influence sponsor behavior
Dominion Energy (D) in Virginia) supplier/infrastructure Gigawatt-scale delivery hinges on interconnection timelines and state utility politics
CMBS, syndicated loans via JPM, GS, WFC, C, SMBC) Business model assumes continuous multi-billion refinancing; spread widening directly hits development returns
Eaton (ETN), Schneider Electric, Caterpillar (CAT)/Cummins (CMI) gensets) Industry-wide lead times on transformers, switchgear, and generators gate delivery schedules
Digital Gateway (Virginia) terminated Jul 2026 after zoning approvals were voided amid opposition; moratoria spreading in key metros
Advantages
- One of the sector's largest fully pre-leased development pipelines ($25B+, 2026)
- Investment-grade project-bond access (Baa2 — landmark 144A single-campus DC bond) - lower debt cost than most private peers
- Blackstone's balance sheet, relationships, and land/power procurement machine
- Multi-decade utility agreements in anchor markets (Georgia Power ~350 MW renewables, 20 yr)
- Sticky, escalator-linked 15+ year hyperscaler leases insulating downside on delivered assets
Weaknesses
- Opaque financials - no consolidated public P&L or leverage disclosure since 2021
- Negative FCF through a very long build-out; returns depend on execution across dozens of concurrent projects
- Tenant and single-campus concentration (Microsoft-tied Fayetteville bond is the template)
- Debt-funded sponsor distributions introduce classic private-equity leverage-extraction risk
- Growth narrative is sponsor-reported (capacity multiples), not audited earnings; founder-CEO ousted Mar 2025
Bottlenecks
- Grid power availability and interconnection queues in core metros (the industry-wide #1 constraint)
- Transformer, switchgear, and generator lead times gating data-hall delivery
- Skilled construction and commissioning labor across ~35 simultaneous projects
- Local zoning, water, and noise politics - already cost QTS the Virginia Digital Gateway project (Jul 2026)
- Continuous access to multi-billion-dollar debt refinancing windows
Top signals & trends
Top signals
Market treats pre-leased AI campuses as investment-grade infrastructure
Sponsor-reported but directionally corroborated by financing volumes
Third-party appraisal uplift on stabilized assets
Dividend-recap-style leverage extraction while build-out FCF is negative
First visible crack in land-bank convertibility; NoVa politics tightening
Confirmed liquidity/valuation-discovery channel for QTS-class assets; BXDC itself does no ground-up development
Sponsor consolidating control for a faster build-out; key-person transition mid-cycle
Sector-wide cost-of-capital compression benefits the largest pre-leased developers most
Trends
QTS's entire model; demand currently exceeds deliverable supply
positive for incumbents with secured power, negative for pipeline conversion · QTS's utility relationships are an edge; ungated land is worth less
QTS is the category-defining issuer
Already killed QTS's largest proposed project; raises siting cost and time
Competes for power/land but also validates third-party leasing when speed matters
negative if it materializes · Existing 15+ yr leases hold; new pre-leasing pace is the exposed variable
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
20-year agreement for ~350 MW of new renewable (solar) energy supporting Atlanta-Metro and Suwanee campuses (Mar 2022); utility for the Fayetteville build
Power for Virginia footprint
Power/cooling equipment for AI data halls
Switchgear, UPS, electrical distribution
Electrical/cooling infrastructure
Backup generation
Backup generation
Named anchor tenant, Fayetteville GA campus ($4.6B bond collateral)
BREIT describes the $25B+ pipeline as 100% pre-leased to large, investment-grade global technology companies on 15+ yr leases; specific names beyond Microsoft not disclosed
Legacy QTS federal/compliance-heavy colo business (pre-2021 disclosure)
Closest analog: hyperscale wholesale developer with large IG-tenant pipeline and heavy ABS issuance
Hyperscale/AI-focused developer (50 campuses, ~5 GW); the deal is the closest public valuation comp for QTS-scale platforms
Hyperscale wholesale peer taken private in the same 2021-22 wave
NoVa-rooted hyperscale developer competing for the same tenants and power
Build-to-suit hyperscale campuses
Hyperscale developer with aggressive US + international pipeline
Large US wholesale campuses (Nevada, Michigan, Georgia)
Largest listed wholesale/colo REIT; competes for hyperscale allocations globally
Retail colo/interconnection leader moving into xScale hyperscale JVs
Global wholesale platform
The customers are also the largest builders; make-vs-lease decisions cap third-party pricing