Annualized Run-Rate (ARR) $111.4M $9.28M monthly billings
Adjusted EBITDA Run-Rate $64.6M 58.0% EBITDA Margin
Active Deployable Racks 2,000 @ 32 kW avg design
Implied Enterprise Value $1,421M $22.2M / active MW

Data Hall Row & Rack Cluster Topology

Real-time thermal load & busway power distribution across modular PODs. Hover/tap racks to inspect.

Nominal AI Rack
Cooling CRAH
High Thermal
Unleased / Buffer
Hover or touch a rack cluster to inspect telemetry

P&L Run-Rate Anatomy

Annualized baseline from contracted colocation capacity and ancillaries

Line Item Component Monthly Annualized
Wholesale Colocation Space & Power Reservation $8.17M $98.1M
Interconnection & Cross-Connect Carrier Fees $1.11M $13.3M
Gross Operating Revenue $9.28M $111.4M
Facility OPEX & Engineering Maintenance (18%) ($1.67M) ($20.0M)
Corporate SG&A, Security & Compliance (12%) ($1.11M) ($13.4M)
Utility Grid Power Expense (Net of pass-through) ($1.12M) ($13.4M)
Adjusted EBITDA $5.38M $64.6M

IPO Valuation Sensitivity

Enterprise Value (EV) matrix across varied multiple expansions

Multiple Adj. EBITDA Enterprise Value

Context: Hyperscale Data Center IPO Economics

Data center operators (such as DayOne in recent filings) evaluate business performance through contracted megawatt (MW) capacity, lease rate per kilowatt ($/kW/month), and annualized run-rate (ARR). With the surge in dense generative AI clusters demanding 30 kW to 80+ kW per liquid-cooled rack, critical IT load efficiency (measured via PUE) and utility substation capacity have become primary constraints on valuation multiples.

Why does PUE matter for revenue?

Power Usage Effectiveness represents the multiplier between critical IT power and total utility draw. A PUE of 1.20 means that for every 100 MW of GPU compute, the facility draws 120 MW from the grid. Lower PUE saves utility expenses and frees capacity for leased customer racks.

Pass-Through vs Bundled Power

Hyperscale leases for major cloud operators (e.g. AWS, Microsoft, Meta) almost universally structure electricity as a direct utility pass-through, insulating operators from grid tariff fluctuations and stabilizing EBITDA margins near 55–65%.

Rack Density & AI Workloads

Traditional enterprise colocation averages 6–10 kW per cabinet with standard raised-floor perimeter cooling. Modern AI clusters (NVIDIA H100/B200 NVL72) require 40–120 kW per rack, necessitating direct-to-chip liquid cooling manifolds and dedicated rear-door heat exchangers.