In this section, we hear from some of the major data center providers about how they are transitioning these ideas into tangible, AI- and neocloud-ready infrastructure.
In 2025, neocloud revenues exceeded 23 billion USD, including approximately five billion USD in Q2 alone—a figure that represents over 200 percent year-on-year growth. Projections indicate revenues will reach nearly 180 billion USD by 2030, with an average annual growth rate of 69 percent.*
This rapid growth is driving exceptional demand for highly specialized, high-density capacity, with current requirements ranging from 50–100kW per rack and future designs targeting 200–300kW per rack. Traditional air-cooled data halls, typically designed for 5–10kW per rack, are inadequate for supporting this new class of GPU infrastructure.
So, needless to say, the emergence of neocloud providers is quickly changing the game in both the cloud market and its underlying data center infrastructure. Specifically, leading colocation and wholesale data center providers are accelerating innovation in power, cooling, interconnection, and campus design in order to meet evolving needs. For instance, adapting from 5kW racks to AI factories involves more than simply increasing power supply—it also means redesigning white space, implementing new cooling topologies, and aligning more closely with neocloud build cycles to accelerate AI deployments. Likewise, it also involves ensuring all expansions and adaptations are not only engineered with AI factories in mind but also designed to enhance operational sustainability.
ALIGNED DATA CENTERS
Next-Generation Cooling for the Most Demanding Workloads
Aligned has consistently advocated for adaptive cooling solutions to meet the demands of the highest-density deployments, sustainably. For instance, neoclouds like Lambda find significant opportunities in Aligned’s adaptive data centers and award-winning cooling innovation capable of supporting the highest-density AI accelerators: Delta Cube, an advanced air-cooling solution that supports power densities of up to 50kW per rack, and DeltaFlow~, a liquid cooling technology designed to support rack densities beyond 350kW per rack.
“In an era where neocloud platforms are pushing compute densities far beyond legacy capabilities, Aligned remains ahead of the curve through relentless innovation in adaptive design. From pioneer hybrid cooling to grid-integrated power solutions, our infrastructure is built to support the rapid deployment of today’s most intensive workloads while providing the scalability to meet tomorrow’s unknown requirements.”
—Phill Lawson-Shanks, Chief Innovation Officer, Aligned
CORESITE
GPUaaS, DGX Readiness, and Custom Cooling
AI, HPC, and hybrid IT are reshaping how enterprises deploy infrastructure, driving demand for data centers that can support higher power densities, advanced cooling, and resilient designs—while also enabling seamless interconnection. CoreSite is an NVIDIA DGX-Ready certified data center developer and operator, with facilities designed to support high-performance AI and neocloud infrastructure. These facilities are engineered for high power densities, energy-efficient configurations, and robust interconnection to support AI workloads. For example, STN Inc. is using CoreSite’s customized liquid cooling system to support GPU One, a GPU-as-a-Service platform.
“Our data center campuses serve as gateways to the neocloud ecosystem. Through our direct cloud onramps and native network deployments, customers can interconnect directly to leading neocloud platforms using high-speed connectivity up to 400G. As we expand capacity in key markets, we continue to deliver dense interconnection, deploy additional direct cloud onramps, and customizable cooling and redundancy systems—enabling enterprises to both host AI infrastructure and directly access the compute powering today’s AI factories.”
—Juan Font, CEO, CoreSite, and SVP of American Tower
DIGITAL REALTY
Industrializing Liquid-to-Chip Cooling at Global Scale
With its high-density colocation solutions integrating power, advanced cooling, connectivity, and compliance into a single package for AI and HPC workloads, Digital Realty has made AI-ready and neocloud-ready design a priority. Neoclouds like CoreWeave and Vultr are hosting large deployments in Digital Realty infrastructure to make use of its distributed cloud architecture and expedite their time-to-market.
“AI and HPC represent a shift larger than cloud—moving intelligence from code into permanent, physical infrastructure. These workloads demand dense power, advanced cooling, and globally interconnected systems. Digital Realty’s strength lies in more than two decades of operating the world’s most complex environments, converting bespoke engineering into resilient, repeatable platforms. Purpose-built for AI inference—the long-term economic engine of AI—our modular architecture supports hybrid air and liquid cooling, direct-to-chip deployments, and continuous retrofits. In partnership with NVIDIA and Lenovo, we deliver future-proof infrastructure that scales with silicon, not against it.”
—Chris Sharp, CTO, Digital Realty
EDGECONNEX
AI-Enabled Data Center Infrastructure Anchored by Ingenuity
The next-generation data center solution, Ingenuity, has been engineered explicitly by EdgeConneX to meet the demands of HPC, AI training, and AI inference workloads. EdgeConneX and Lambda are developing AI data center infrastructure that leverages cooling technologies combining liquid-to-the-chip direct cooling and air cooling.
“With our innovative industry-leading Ingenuity data center solution, we can support up to 600kW+ per rack for dedicated workloads and mixed workloads for AI / HPC.EdgeConneX’s partnership with Lambda, the Superintelligence Cloud, is built on our fundamental promise to deliver cutting-edge infrastructure solutions for Lambda’s AI-driven technologies to its cloud and hyperscale customers. We are building 30MW+ of state-of-the-art high-density data center infrastructure in Chicago and Atlanta with Lambda, representing a strategic advancement toward its 2GW+ vision, which will lead to the deployment of a million GPUs by the end of the decade. Our innovative hybrid cooling design will effectively support the high-power demands of next-generation AI accelerators while ensuring energy efficiency and environmental sustainability.”
—Phillip Marangella, Chief Marketing and Product Officer, EdgeConneX
EDGECORE
Preparing for Liquid Cooling at Scale
EdgeCore is developing massive data center campuses that are built with power and scalability in mind and engineered for very high power density and large critical loads. These factors make them ideal candidates for neoclouds, that need both space and opportunity to grow quickly without long lead times.
“Power, latency, scale, and chip supply are the four defining bottlenecks in the modern infrastructure race and neoclouds are pushing all four at once. It’s a formidable challenge, but a clear structural opportunity. EdgeCore is building for an AI-driven era in which specialized cloud deployments fuel outsized growth and require multi-hundred-megawatt campuses delivered with certainty. Neoclouds are accelerating the adoption of direct-to-chip liquid cooling, high-capacity power, metro adjacency, and modular design. The winners are building for this reality now and do so with minimal technical debt.”
—Clint Heiden, Chief Commercial Officer, EdgeCore
FLEXENTIAL
High-Density Colocation as a Neocloud Launchpad
Flexential has become a significant infrastructure partner for GPU-focused cloud platforms. For example, both CoreWeave and DigitalOcean have expanded operations into several Flexential facilities in order to access the high-density GPU support, rapid scaling of cloud services, and seamless connectivity and integration that can support their extensive GPU-accelerated cloud expansion.
“The demand for data center capacity is at a record-level high, and that presents a huge opportunity for providers that can build and interconnect efficiently. This surge isn’t a one-off but a multi-stage build for the next decade of infrastructure. Elements like power and cooling constraints, network density demands, and longer lead times for critical infrastructure components certainly pose challenges, but through Flexential’s modular, purpose-built, high-density colocation infrastructure, we are able to implement innovative cooling solutions (which scale energy and cooling resources flexibly according to workload demands) and software-defined, highly flexible connectivity (which provides distributed, on-demand networking) to support next-generation cloud environments.”
—Ryan Mallory, CEO, Flexential
NTT GLOBAL DATA CENTERS
Gigawatt-Scale, AI-Ready Campuses
NTT Global Data Centers, recognized as one of the world’s largest data center operators, is expanding its capacity significantly by adding approximately 1GW across seven markets and prioritizing higher power densities, advanced cooling systems, and robust global supply chains to support both neocloud and hyperscale clients.
“Neoclouds are redefining how AI is built and delivered, and our role is to provide a trusted, AI‑ready foundation that accelerates innovation. At NTT Global Data Centers, we co‑design capacity with neocloud partners, combining high‑density power, advanced liquid cooling, and global campus scale with speed, security, and sustainability. This approach enables neoclouds to move quickly, interconnect seamlessly across regions, and scale responsibly while meeting performance and environmental goals. Together, we’re shaping the infrastructure backbone for the next era of AI.”
—Doug Adams, CEO, NTT Global Data Centers
STACK INFRASTRUCTURE
Designing for 100–300kW Racks
STACK Infrastructure has specifically targeted AI and neocloud workloads with its AI-ready portfolio. The company’s cooling architecture accommodates multiple tiers, including up to 30kW per rack with traditional air cooling, 50kW with rear-door heat exchangers, 100kW with direct-to-chip liquid cooling, and planned capability to support 300kW or more per rack through immersion cooling. This range of cooling options enables neocloud tenants to increase rack density over time without relocating from an existing campus and transitioning from air-assisted deployments to fully liquid-cooled clusters as their AI workloads evolve.
“Neoclouds are purpose-built for AI, and they require data centers built the same way. At STACK APAC, we are an engineering-led developer, focused on the fundamentals of power, cooling, and execution that GPU-intensive workloads demand. We build campuses designed for high-density deployments today, which are liquid cooling native for higher rack densities but also capable of supporting workloads that require hybrid environments supporting CPU and ASIC capacity as workloads evolve. By delivering reliable, future-ready infrastructure at speed and at scale, we enable neocloud providers to focus on availability, flexibility, performance, and growth.”
—Preet Gona, CEO, STACK APAC
