A New Era of Scale
The data center industry is undergoing a dramatic transformation. In the past, a 5MW colocation deal was considered a major win. Today, hyperscale giants like Amazon, Google, and Microsoft are locking in colocation leases exceeding 100MW — a scale that would’ve seemed inconceivable just five years ago. This shift is not just a matter of demand — it's a fundamental reshaping of how data center infrastructure is sourced, deployed, and monetized.
This blog explores the reasons behind this surge in 100MW+ colocation deals, the driving forces from the hyperscalers themselves, how colocation providers are adapting to this mega-scale reality, and what it means for the global data infrastructure market.
The Hyperscale Tidal Wave
Who Are the Hyperscalers?
Hyperscalers refer to companies that build and operate massive-scale computing infrastructure to support global cloud services, AI workloads, and digital platforms. The top hyperscalers include:
- Amazon Web Services (AWS)
- Microsoft Azure
- Google Cloud
- Meta (Facebook)
- Apple
- Oracle
- Alibaba Cloud
These companies operate hundreds of data centers globally and spend billions annually on infrastructure. In 2024 alone, Amazon announced plans to invest over $100 billion in its global data center expansion over the next decade — setting a new precedent for scale and speed.
Why 100MW+ Deals Are Becoming the New Norm
1. AI and Machine Learning Workloads
The explosive rise of AI, especially large language models like OpenAI's GPT-4 and future successors, has exponentially increased compute demands. Training and running these models require vast amounts of processing power, which translates into more GPUs, more power, and more space.
- NVIDIA H100 clusters draw over 40kW per rack — double the power density of traditional racks.
- AI clusters are often deployed in hyperscale configurations, requiring large, contiguous space and dedicated power infrastructure.
A 100MW deal ensures hyperscalers have the space and power for multi-generational AI infrastructure rollouts.
2. Supply Chain Preemption
In today’s market, power is the bottleneck. Grid limitations, permitting delays, and sustainability requirements mean securing power and capacity years in advance is now standard. Hyperscalers are signing massive deals to lock in future availability and avoid being caught in capacity shortages.
3. Operational Efficiency and Economics of Scale
Hyperscalers prefer consolidating workloads into larger, more efficient campuses. A single 100MW facility offers:
- Lower per-kW costs through economies of scale
- Simplified operations, network management, and security
- Greater control over cooling, energy sourcing, and sustainability initiatives
How Colocation Providers Are Responding
1. Building Bigger and Faster
The traditional model of building 10MW–20MW facilities and scaling incrementally is fading. Operators like QTS, Vantage, Digital Realty, and STACK Infrastructure are constructing campuses with initial phases of 60MW–120MWand expansion potential beyond 200MW.
Examples include:
- QTS’s Virginia campus with 250MW capacity
- Digital Realty’s upcoming 1GW portfolio expansion in Europe
- STACK’s hyperscale buildouts in Phoenix and Northern Virginia
2. Land Banking and Power Procurement
To meet hyperscaler expectations, providers are:
- Acquiring hundreds of acres of land in emerging and Tier 1 markets
- Partnering directly with utilities and renewable energy developers
- Securing grid interconnects and power purchase agreements (PPAs) 3–5 years in advance
3. Modular and Prefabricated Builds
Speed is critical. Colocation operators are increasingly adopting:
- Prefabricated modular designs
- Offsite construction and plug-and-play components
- Streamlined entitlements and standardized permitting across jurisdictions
This reduces time-to-market and enables hyperscaler clients to deploy faster.
Where the 100MW Deals Are Happening
1. Northern Virginia (Ashburn)
Still the world’s largest and most active data center hub. Despite power grid challenges, 100MW+ deals continue due to unmatched fiber density and proximity to federal contracts.
2. Phoenix, Arizona
A rising favorite due to available land, competitive power costs, and dry climate. Multiple providers have secured 200MW+ campuses here.
3. Dallas-Fort Worth, Texas
Strong utility infrastructure, no state income tax, and abundant land make DFW a hotspot for hyperscale expansions.
4. Chicago, Illinois
Chicago is seeing renewed interest due to its central U.S. location and robust grid capacity, making it ideal for AI and ML workloads.
5. International Markets
- Ireland, Northern England, and Sweden in Europe
- Singapore, Indonesia, and India in APAC
- Chile and Brazil in LATAM
Global providers are now designing campuses to serve regional and edge deployments at scale.
Challenges in Meeting 100MW+ Demands
1. Power Constraints
Even with land and capital, grid access is a limiting factor. Cities like Dublin, Singapore, and even parts of Virginia are pausing new power connections due to strain.
2. Cooling and Environmental Limits
High-density workloads generate intense heat. Providers must now integrate:
- Liquid and immersion cooling systems
- Water-free cooling in arid climates
- Sustainable site design to meet ESG and emissions mandates
3. Workforce and Construction Labor Shortages
The construction of hyperscale campuses requires highly specialized labor — electricians, HVAC engineers, and fiber techs — all of whom are in short supply globally.
The Strategic Implications
For Hyperscalers:
- Colo becomes an extension of cloud: many now view wholesale colocation as part of their infrastructure rather than a third-party service.
- Vendor consolidation: fewer, larger deals mean long-term relationships with select providers who can build fast and deliver globally.
- Sustainability pressure: ESG mandates are pushing hyperscalers to demand carbon-free, renewable-powered builds.
For Colo Providers:
- Shift to campus-based models: smaller sites are being phased out for mega-campuses with cross-market integration.
- Increased competition: only the most well-capitalized providers can win 100MW+ deals — expect M&A and partnerships to rise.
- Revenue predictability: large anchor clients provide stable long-term income, improving operator financial health.
The Bigger Picture: 100MW Is the New Normal
100MW is no longer an outlier — it’s the new benchmark. This size represents the evolving needs of a cloud-first, AI-accelerated world. Whether it’s training the next frontier of AI models, delivering real-time services at global scale, or enabling automation and analytics, massive compute power is the backbone.
Hyperscalers aren’t just demanding space and power — they’re reshaping the geography, economics, and engineering of colocation.
Prepare for the Gigawatt Era
The era of megawatts is quickly giving way to the gigawatt era, where entire regions are being transformed into digital infrastructure ecosystems.
For data center professionals, brokers, infrastructure vendors, and investors, the rise of 100MW+ colocation deals signals both opportunity and responsibility — to scale sustainably, plan globally, and innovate boldly.
The next five years will be defined by those who can meet this scale not just with hardware, but with vision.
