STACK Infrastructure expands AI data centres across three continents
STACK Infrastructure has climbed to number 20 in the Top 100 Data Centre Companies for 2025, reflecting a fast build-out aimed squarely at AI and cloud demand. The Denver-based operator now manages over 2.5 gigawatts (GW) built or under development, with another 4 GW in the pipeline.
The company delivers powered land, build-to-suit, wholesale colocation and rapid deployment capacity across key hubs: Northern Virginia, Silicon Valley, Phoenix, Portland, Atlanta, Dallas/Fort Worth, Chicago, Tokyo, Melbourne, Johor Bahru and Seoul.
Green financing accelerates scale
STACK secured more than US$6bn in green financing during 2025 to fund new builds across North America. That includes a US$900m facility for the 200MW NVA05 campus in Manassas, Virginia, and financing for the 500-acre Stafford Technology Campus-planned for 19 facilities and more than one gigawatt of capacity.
In March 2024, the company issued US$240m in green securitised notes, achieving the lowest blended spread over the treasury rate seen by a data centre issuer since October 2021. This pricing signals investor confidence in its sustainability approach and project execution. For context on green bond standards, see the Green Bond Principles from ICMA here.
"Our clients prioritise minimising environmental impacts, which parallels our mission to advance a sustainable digital future," says Brian Cox, CEO of STACK Americas.
Strategy: hyperscale-first, with focus on AI
STACK divested its European colocation business to Apollo-managed infrastructure funds in April 2025. The deal covered seven facilities across five markets and frees up capital and teams to focus on hyperscale development in North America and Asia-Pacific.
The company's design framework targets high-density deployments required by AI and machine learning. Think more power per rack and precise cooling to keep GPUs productive without thermal throttling-delivered with a flexible blueprint that can scale as workloads change.
"STACK was built for the world's largest innovators… we have prioritised long-term scalability with a flexible data centre design," says Matt VanderZanden, President of STACK Americas. "As an AI-ready digital infrastructure company with a proven track record of high-density deployments, we're building on this foundation to meet evolving client needs."
New builds and markets
In September 2025, STACK announced it will deliver digital infrastructure for Oracle's planned Stargate campus in Doña Ana County, New Mexico, where Oracle has committed to deploy up to 4.5 GW of capacity. The company also brought its first Japanese facility online: a 36MW campus in Tokyo's Inzai District.
APAC growth continues in Melbourne, Johor Bahru and Seoul-locations chosen for proximity to cloud regions, strong grid access and growing demand from AI workloads.
Leadership aligned to scale
Brian Cox joined as CEO in 2018 after helping grow Cologix from one data centre to 25. He brings more than two decades of experience building and operating large-scale infrastructure.
Matt VanderZanden was promoted to President in February 2025 after serving as Chief Commercial Officer; previously, he led global data centre site selection at Facebook (now Meta). "We've built a reputation for speed, scale, certainty and responsibility," he says. "I'm excited to work with our team to support clients' growth and success."
Why this matters for IT and development teams
AI projects are bumping into real limits: power availability, cooling density and lead times for capacity. STACK's expansion suggests more options for high-density deployments across multiple metros, which can shorten timelines for AI rollouts and multi-region architectures.
- Capacity where you need it: Coverage across top US markets plus Tokyo, Melbourne, Johor Bahru and Seoul.
- High-density ready: Facilities built for GPU-heavy racks, liquid-ready designs and scalable power blocks.
- Sustainability: Access to green financing can align with corporate carbon goals and reporting requirements.
- Focus and execution: Divestment of smaller colocation assets points to priority on scale and speed for hyperscale builds.
Practical takeaways
- Validate density early: Confirm rack power budgets, cooling method and floor loading before committing hardware roadmaps.
- Model energy and cost: Tie PUE assumptions, utility rates and demand charges to your TCO for AI clusters.
- Plan for staged growth: Lock in expansion paths (power and whitespace) to avoid re-architecting your cluster in year two.
- De-risk locations: Check grid upgrades, water availability and permitting timelines for each market on your shortlist.
- Skill up the team: If you're building AI capability in-house, see curated training paths by role here.
Bottom line
With 2.5 GW delivered or in development and 4 GW queued, STACK is operating at meaningful scale across three continents. Combined with fresh green financing, high-density designs and a hyperscale-first focus, the company is positioned to serve the next wave of AI and cloud demand.
Your membership also unlocks:
