A 520 MW power-secured data center platform across the U.S. Southeast and Midwest — where power is secured first, campuses are standardized, and capacity is leased to credit-quality tenants.
Phoenix Power LLC was formed to advance a portfolio of strategically located, grid power-first inference sites. The platform comprises a 520 MW critical-IT portfolio across ten campuses, anchored by Project Panther in North Texas and the Starboard joint venture in Central Florida.
Campuses are leased on long-dated, triple-net (NNN) terms to investment-grade hyperscale and credit-worthy neocloud tenants — pairing infrastructure-grade cash flows with the fastest-growing demand segment in the market.
Our sites are concentrated in power-rich Southeast and Midwest markets, where generation, land, and pro-growth regulation align — and where the time-to-power gap is widest.
Power is the scarce input. Phoenix owns the power-first path to capacity — the single most decisive factor in site selection today.
Between scaled incumbents and hyperscaler self-build sits an underserved mid-tier. Phoenix occupies it with a differentiated, power-first model.
Utility capacity and interconnection queue position locked before vertical build — de-risking the scarcest input in the market.
Sites concentrated across the Southeast and Midwest, where generation, land, and pro-growth regulation align.
A repeatable 52 MW campus blueprint with modular, phased build-out — standardization that compresses cost and schedule.
Long-dated NNN leases to investment-grade hyperscalers and vetted neocloud tenants, with contractual escalators.
Design-build, power, and engineering partners contracted across the platform to deliver at institutional standard.
Purpose-built for the 2026–2032 AI build cycle — modular 24 MW blocks sized to modern GPU clusters, delivered fast in mid-tier markets.
AI is the structural driver of data center capacity growth — and power availability is now the foremost site-selection criterion, pushing development into power-rich secondary markets.
U.S. data center power demand is projected to grow from ~25 GW in 2024 to 80+ GW by 2030.
Primary-market vacancy sits at record lows, with ~74% of under-construction capacity already preleased.
Total demand grows 3.5× this decade — with AI workloads driving the majority of incremental capacity.
Sources: McKinsey & Company (Aug 2025) · CBRE North America Data Center Trends H1 2025 · Knight Frank Data Centres Global Forecast 2026
AI demand is diverging into two distinct asset classes — gigawatt-scale training campuses and metro-proximate inference sites. Phoenix's standardized 52 MW campus sits squarely on the inference side of the divide.
Gigawatt-scale, remote, single-tenant builds — gated by power, not demand.
Demand grows with a handful of frontier labs, at ~13% CAGR — a fraction of the inference curve.
Metro-adjacent, latency-sensitive 20–100 MW blocks — demand scales with every user, agent, and query.
Inference co-locates with users — favoring power-rich, metro-proximate markets like Texas and the Southeast, and reaching ~2× the training base by 2032.
Phoenix's ten 52 MW inference campuses are purpose-built for the fastest-growing segment of AI demand.
Sources: McKinsey & Company (2025) · Goldman Sachs Research · Epoch AI · JLL North America Year-End 2025
A disciplined, repeatable playbook applied across every campus in the portfolio.
Lock utility capacity and interconnection queue position before committing capital to vertical build.
A repeatable 52 MW reference design with modular, phased halls — compressing both cost and time-to-energization.
Long-dated NNN leases to investment-grade and vetted neocloud tenants, with contractual rent escalators.
Five cornerstone projects across Texas, Florida, and North Carolina — more than 2 GW of identified utility capacity.
A team with a track record of closing utility negotiations, leasing capacity, and delivering data center infrastructure at scale.
A distinguished group of industry leaders spanning energy transition, power systems, enterprise technology, capital markets, and hyperscale infrastructure.
Team experience spans leading platforms and operators including AWS, Meta, Google, Microsoft, Apple, Akamai, Vantage, Oracle, EdgeConneX, and CoreWeave. Names are representative of reference operators and do not imply an existing commercial relationship or endorsement.
Design-build, power, and engineering partners contracted across the platform.
Whether you're a hyperscale or neocloud operator seeking capacity, a landowner in a power-rich market, or a prospective partner — we'd like to hear from you.