Vantage Data Centers
Develops, owns, and operates large-footprint data-center campuses leased to hyperscalers and AI companies under long-term (typically 10-15yr) triple-net-style wholesale leases; capital-intensive build-to-suit + speculative development funded by sponsor equity plus asset-backed securitizations and green project loans. Separately capitalized regional platforms (North America, EMEA, Asia-Pacific).
Vantage does NOT disclose company post-money valuations. The primary points are the disclosed SIZES of equity raises (several at the regional-platform level - EMEA/APAC - not the whole company), used as directional capital-formation markers, not valuations. The 2025 point is the secondary press EV estimate midpoint. Treat magnitudes as raise/EV figures, not equity valuations.
Earnings, margins, COGS & capex
Private hyperscale developer with no published P&L. The disclosed financial story is a capital-formation story: multi-billion sponsor equity (DigitalBridge/Silver Lake $9.2B), plus a deep and repeated debt stack (green project loans, asset-backed securitizations across US, UK, and Continental Europe). Cash flows are dominated by growth capex on GW-scale AI campuses; contracted long-term hyperscale leases underpin the debt but company-level revenue, margins, and FCF are not public.
Revenue trend
Margins
n/a
n/a
COGS structure
Not disclosed. Principal cost drivers for the model: land, electrical/mechanical fit-out (transformers, switchgear, generators, liquid-cooling), construction labor, and power procurement. Frontier is engineered for 250kW+-per-rack ultra-high-density AI with liquid cooling, raising per-MW build cost vs legacy air-cooled colocation.
Capex
The defining line item. Frontier (Shackelford County, TX): $25B, 1.4GW across 10 buildings / 3.7M sq ft on 1,200 acres, first building live 2H2026, full build 2028. Port Washington, WI (Project Lighthouse): $15B, ~1GW (~902MW), four buildings / ~2.5M sq ft, broke ground Dec 2025, complete 2028. Both are OpenAI/Oracle Stargate sites; combined stated investment >$40B. Plus New Albany, OH (192MW / 1.5M sq ft, three pre-leased buildings, funded within the $5B Jun-2025 green loans - a subset, not additive) and the EMEA/APAC platforms.
Latest earnings
n/a
No public guidance. Disclosed operating ambition: drive ~$30B in data-center development; >$40B committed to Frontier + Wisconsin.
- Frontier campus capacity
- 1.4 GW / 3.7M sq ft / 10 buildings ($25B)
- Wisconsin (Port Washington) capacity
- ~1 GW (~902MW) / 4 buildings ($15B)
- 2024 incremental funding
- ~$13B (record)
- 2024 sponsor equity round
- $9.2B (DigitalBridge/Silver Lake, completed Jun 13, 2024)
- 2025 North America green loans
- $5B incremental (incl. $2.25B New Albany, a subset)
Growth drivers
- AI training/inference demand driving GW-scale, high-density build-to-suit leases
- OpenAI + Oracle 'Stargate' partnership — Vantage is a named developer for Texas (Frontier) and Wisconsin (Port Washington) sites
- Hyperscaler capex supercycle (Microsoft, Oracle, Google, Amazon, Meta) requiring third-party developed capacity
- Liquid-cooling / high-density design capability positioning it for GPU workloads legacy colo can't serve
- Deep, repeatable access to project financing (green loans + ABS) to fund the build ahead of demand
- Geographic diversification across North America, EMEA (first euro DC ABS), and Asia-Pacific (GIC/ADIA-funded)
Bull & bear
Vantage is a prime, institutionally-backed picks-and-shovels owner of the physical AI buildout, with anchor Stargate demand, scarce land+power, and demonstrated capital access - a private way to own the AI-infrastructure supercycle.
- Directly levered to the largest AI infrastructure program announced (Stargate: a planned $500B / 10GW by 2029, ~7GW / $400B+ already committed) as a named Texas and Wisconsin developer
- DigitalBridge + Silver Lake sponsorship plus GIC/ADIA/MEAG/AustralianSuper provides equity depth few private peers match
- Repeatable, innovative financing (first euro-based DC ABS; $5B incremental green loans) lowers cost of the capital-heavy model
- Land + power + high-density design is exactly the scarce bundle hyperscalers can't source fast enough
- >$40B committed pipeline gives multi-year contracted-revenue visibility as buildings deliver from 2H2026
- Optionality on an eventual IPO/monetization at a premium DC-infrastructure valuation
A maximally capital-intensive, opaque, highly-levered spec developer whose fortunes ride on an AI-capex cycle and a handful of cash-burning counterparties; any demand or financing wobble is amplified by the balance sheet.
- No public financials - investors underwrite blind vs listed EQIX/DLR with audited disclosure
- $40B+ of committed capex against undisclosed (and structurally negative) enterprise free cash flow
- Dependence on continuous ABS/green-loan issuance - vulnerable to rate spikes or DC-credit-spread widening
- Concentration in OpenAI/Oracle Stargate; OpenAI's own cash burn and program-timeline risk are existential inputs
- Hyperscaler self-build and a crowded field of KKR/Blackstone/Macquarie-backed rivals compress returns on new capacity
- AI-capacity 'digestion' fears (unleased spec MW, power/permitting delays) would strand capital mid-build
What it is worth
Private - no market price. Anchored on the last priced equity event ($9.2B DigitalBridge/Silver Lake round completed Jun 13, 2024) plus press/analyst enterprise-value estimates and read-through from public DC-infra comps (DLR, EQIX trade at premium EV/EBITDA on AI-demand; CRWV as an AI-native comp). All figures are estimates - Vantage does not disclose revenue or EBITDA, so any multiple-based value is illustrative, not derived from audited financials.
AI-capex digestion leaves spec capacity unleased and/or rates/credit-spreads widen, raising financing cost against a levered, undisclosed-cash-flow balance sheet - equity value compresses materially and monetization/IPO is deferred.
Executes the committed pipeline with normal delays; value roughly in line with the ~$15-20B press EV estimates (unverified), scaling with delivered, leased MW and a still-supportive financing environment.
Stargate demand fully materializes, campuses lease up on schedule at premium high-density rents, and capital stays cheap - enterprise value re-rates well above the ~$15-20B press range toward a scarcity premium; eventual IPO at a top-tier DC-infra multiple.
The core question is not current earnings (undisclosed, likely negative FCF mid-build) but the value of a >$40B contracted/near-contracted GW-scale AI pipeline against the cost and availability of capital to complete it. Value accrues as Frontier/Wisconsin buildings deliver (from 2H2026) into long-term hyperscale leases. Reported press EV estimates clustered ~$15-20B in 2025 (secondary/unverified); treated as a base-case reference only.
SWOT
Strengths
- Named developer on the OpenAI/Oracle Stargate program - anchor demand at unprecedented scale
- Proven ability to raise multi-billion equity (DigitalBridge/Silver Lake $9.2B) and repeat debt (green loans, first US/UK/euro DC ABS)
- High-density, liquid-cooled campus design tuned for GPU/AI workloads (250kW+/rack at Frontier)
- Large contiguous land + power positions (1,200-acre Frontier) that are increasingly the binding constraint in the industry
- Global multi-region platform (NA, EMEA, APAC) with region-specific institutional capital partners (GIC, ADIA, MEAG, AustralianSuper)
Weaknesses
- Opaque financials - no public revenue, margin, or FCF; harder to underwrite than listed REIT peers (DLR/EQIX)
- Extreme capital intensity and reliance on continuous capital-markets access; a financing-cost or credit-spread shock hits the model directly
- Heavy leverage via project/ABS debt against a still-building asset base
- Customer concentration risk toward a small set of hyperscalers/AI labs (OpenAI, Oracle, Microsoft)
- Speculative/spec-build exposure if AI capacity demand decelerates before leases are signed
Opportunities
- Sustained hyperscaler + AI-lab capex supercycle needing third-party developed GW-scale capacity
- Sovereign / international AI infrastructure buildouts (APAC via GIC/ADIA; EMEA expansion)
- Power-secured land as a scarce, monetizable moat as grid interconnection queues lengthen
- Potential future IPO or partial monetization by DigitalBridge/Silver Lake in a strong DC-infra market
- Adjacent revenue from on-site power generation / behind-the-meter and grid partnerships
Threats
- AI-capex 'digestion' or a demand air-pocket leaving spec capacity unleased
- Rising rates / widening credit spreads increasing the cost of the debt-funded model
- Power availability, grid interconnection delays, and local water/permitting opposition
- Hyperscalers self-building (Microsoft, Meta, Google, Amazon) rather than leasing
- Concentration in Stargate - program timeline slippage or counterparty stress (OpenAI cash burn) reverberates
- Well-capitalized private competitors (QTS/Blackstone, CyrusOne/KKR, Aligned/Macquarie, CoreWeave) bidding for the same land, power, and tenants
Moats, dependencies & bottlenecks
Moats
1,200-acre Frontier with contracted power is increasingly the industry's binding constraint; hard to replicate quickly.
Long-term wholesale leases create switching costs, but tenant concentration cuts both ways.
DigitalBridge/Silver Lake equity + repeatable ABS/green-loan debt; durable only while capital markets stay open.
Real capability edge over legacy colo, but peers (QTS, CoreWeave, Aligned) are converging on it.
NA/EMEA/APAC footprint aids large multi-site tenants; not unique among top-tier developers.
Dependencies
Customer / anchor demand Frontier and Wisconsin are Stargate sites; program pace and OpenAI/Oracle financial health drive lease-up.
Model requires continuous multi-billion issuance; sensitive to rates and DC-credit spreads.
Sponsor / control equity Provide equity and strategic direction; DigitalBridge is publicly listed and itself cyclical.
Interconnection queues, water, and permitting gate delivery timelines.
switchgear, generators, cooling) Long-lead electrical gear (Schneider, Vertiv, Eaton, Caterpillar/Cummins) can bottleneck schedules.
AI-capacity demand ultimately tracks GPU availability and AI-lab economics.
Advantages
- Named participant in the flagship AI-infrastructure program (Stargate)
- Deep institutional equity + innovative repeatable debt financing
- Scarce large-scale land + secured power positions
- Purpose-built high-density, liquid-cooled campus design for AI
- Global multi-region platform with region-specific capital partners
Weaknesses
- Opaque, unaudited financials - no public revenue/margin/FCF
- Extreme leverage and capital intensity; structurally negative enterprise FCF while building
- Tenant concentration toward a few hyperscalers/AI labs
- Speculative-build exposure to an AI-demand slowdown
- Reliance on open, cheap capital markets
Bottlenecks
- Power availability and grid interconnection timelines
- Long-lead electrical equipment (transformers, switchgear, generators)
- Access to and cost of debt/equity capital for the build-ahead model
- Skilled construction labor and site permitting/community approvals (water, zoning)
- Lease-up pace vs speculative capacity delivered - the demand/supply timing gap
Top signals & trends
Top signals
Execution proof-point; on-time delivery de-risks the pipeline.
Signals sustained low-cost capital access; a spread blowout would flip bearish.
Anchor-demand health; slippage or OpenAI cash stress is a direct negative.
Public read-through on private valuation and eventual IPO intent.
Pipeline growth and contracted lease-up.
Demand air-pocket risk for spec capacity.
Trends
Primary demand tailwind for GW-scale wholesale capacity.
Favors purpose-built developers over legacy air-cooled colo.
Raises barriers (moat) but gates delivery and adds cost.
Deepens the financing toolkit Vantage pioneers (first euro DC ABS).
Largest tenants building own capacity could shrink the third-party TAM.
Determines cost of the debt-funded model; a tightening cycle would pressure returns.
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.
GPUs that drive tenant AI-capacity demand (indirect, via tenants)
Power/thermal management and liquid-cooling infrastructure
Electrical distribution, switchgear, DC power systems
Electrical gear, UPS, power distribution
Backup generators / power systems
Standby power generation
Anchor AI tenant via Stargate (Texas Frontier, Wisconsin) - private
Stargate cloud/compute partner and lease counterparty
Hyperscale cloud tenant (Azure)
Hyperscale cloud tenant class
Hyperscale cloud tenant class
Hyperscale AI/infra tenant class
Blackstone-owned hyperscale developer; the closest scale peer, often cited neck-and-neck with Vantage on US capacity
AI-native GPU cloud + DC build-out; an OpenAI compute provider whose projects count toward Stargate capacity - competes for AI capacity and tenants
Public hyperscale/colocation REIT; large global wholesale footprint
Public interconnection/colocation leader; more retail/interconnect, expanding into hyperscale (xScale)
KKR/Global Infrastructure Partners-owned hyperscale developer; direct wholesale competitor
Macquarie-backed high-density/AI-focused developer; direct competitor for GPU campuses
American Tower's colocation arm; more interconnection-oriented
IPI Partners-backed hyperscale developer; competes for build-to-suit
DigitalBridge-affiliated DC platforms; sister assets more than pure rivals