Chapter 3

Chapter 3: Data Center Case Studies

Global Models for AI-Driven, Sustainable Infrastructure

As AI workloads become increasingly energy- and compute-intensive, the world’s leading technology companies are rethinking how and where data centers are built. This chapter showcases three case studies—Meta’s Project Cosmo, DayOne’s SIJORI strategy, and Kao Data’s Harlow campus—each demonstrating a different approach to high-performance, sustainable, AI-centric infrastructure development.

 3.1 Meta’s Project Cosmo (Cheyenne, Wyoming)

Meta’s newest facility, Project Cosmo, exemplifies a next-generation, AI-specific data center model.

Key Facts:

• Size: 715,000 sq. ft

• Investment: $800 million

• Operational Year: 2027

• Power Source: 100% renewable (solar farm + energy trading partnership)

• Cooling: Closed-loop liquid system (minimal water use)

Strategic Advantages:

• Located in Cheyenne, Wyoming, selected for its cool climate, low taxes, and supportive regulatory environment.

• Backed by a $1.2 billion solar energy project from Enbridge with 1.2 million panels and battery storage.

• Integrated into Black Hills Energy’s real-time trading market (California ISO’s WEIM) for flexible energy sourcing.

Community Impact:

• Meta committed to building sewer and water lines, and to not exceed the water usage of a typical 100-person business.

• Grazing-compatible solar array design preserves some agricultural land use.

This case shows how AI-specific infrastructure, sustainability, and community partnerships are converging in modern hyperscale builds.

3.2 DayOne in the SIJORI Region (Singapore–Johor–Riau Islands)

DayOne, a Chinese technology infrastructure company, is executing a hub-and-spoke data center strategy across the SIJORI triangle — combining connectivity in Singapore, low-cost operations in Malaysia, and expansion capacity in Indonesia.

SIJORI Strategy Highlights:

• Singapore: High-density hub in Jurong East; 20MW capacity; powered by hydrogen-ready fuel cells.

• Johor, Malaysia: 120MW across two campuses; serves as cost-effective overflow for Singapore.

• Batam, Indonesia: 72MW campus with direct subsea cable access to Singapore; built in collaboration with the Indonesia Investment Authority (INA).

Why it Matters:

• DayOne leverages geographic synergy: regulatory strength in Singapore, low costs in Johor, and abundant land in Batam.

• Its Batam campus is the first in Asia specifically designed for AI workloads.

• DayOne aims to reach 100% renewable energy use by 2030, pioneering sustainability in Asia’s digital infrastructure market.

This model demonstrates how companies can de-risk energy and capacity challenges through regional distribution and diversification.

 3.3 Kao Data’s Harlow Campus (United Kingdom)

Kao Data, based in the UK, has built the region’s first AI-native data center platform, optimized for high-density, GPU-intensive workloads like machine learning and high-performance computing (HPC).

Facility Highlights:

• Location: Harlow, UK (15-acre campus)

• Capacity: 40MW IT load

• Design: Supports up to 80kW per rack and 480kW across AI pods

• Cooling: Direct-to-chip liquid cooling (30% energy savings)

• Power Use Effectiveness (PUE): 1.2

Sustainability Commitments:

• 100% renewable electricity since 2018

• Direct power matching with a wind farm in Kent

• Use of hydrotreated vegetable oil (HVO) for backup generators

• Transitioning to local Power Purchase Agreements (PPAs)

Strategic Connectivity:

• Carrier-neutral with ultra-low latency to major cloud providers

• <2ms latency to major European exchange points (e.g. London Internet Exchange)

Regulatory Alignment:

• Fully compliant with UK energy and data protection laws

• Classified as Critical National Infrastructure (CNI)

Kao Data illustrates how purpose-built AI data centers can balance cutting-edge performance, compliance, and environmental goals in mature markets.

Chapter Takeaways

• Meta is setting a new bar for sustainability and local integration in the U.S.

• DayOne shows the power of regional optimization and speed-to-market in Asia.

• Kao Data exemplifies how Europe is addressing AI demands with low-latency, high-density, and climate-conscious facilities.

Each case reflects a shift toward AI-optimized, renewable-powered, and policy-aligned infrastructure—providing blueprints for the global data center industry.