E03 Why Data Centres Are Built This Way

Key Takeaways

• Modern data centres are complex infrastructure systems rather than simply buildings filled with computers.
• Most of the infrastructure inside a data centre exists not to perform computing, but to ensure computing can continue reliably under almost all circumstances.
• Every major system—from electrical power and cooling to security and fire protection—is the result of balancing reliability, performance, cost, safety, sustainability and community expectations.
• Many environmental and social impacts associated with data centres originate from the supporting infrastructure required to achieve continuous operation, rather than from computing itself.
• Understanding why these systems exist helps explain both the remarkable capabilities of modern data centres and the challenges surrounding their development.

More Than a Building Full of Computers

In the previous articles of the MDCO Explain Series, we explored what a data centre is and how it exchanges electricity, water, connectivity and other resources with the outside world to deliver the digital services that underpin modern society.

The next logical question is perhaps the most important one.

Why are data centres built the way they are?

Why do they require enormous substations, rows of diesel generators, kilometres of electrical cables and cooling pipes, sophisticated fire protection systems, multiple layers of physical security, and extensive monitoring and control systems? Why can a building that appears externally quiet and unremarkable contain engineering systems comparable in scale and complexity to those found in major industrial facilities?

To many people, these features appear excessive. After all, if the purpose of a data centre is simply to run computers, why does so much of the building appear to have nothing to do with computing?

The answer is surprisingly simple.

Very little of a modern data centre actually exists to perform computation.

The servers, storage devices and networking equipment may be the reason the facility exists, but they occupy only one part of a much larger ecosystem of supporting infrastructure. Every other major system is there for one purpose: to ensure that those servers can continue operating safely, reliably and continuously.

A useful comparison is the human body. Although the brain performs the functions we most closely associate with intelligence, it cannot function independently. It depends on the heart to circulate blood, the lungs to provide oxygen, the nervous system to coordinate activity, the skeleton to provide structure, and the immune system to defend against threats. If any of these supporting systems fail, the brain cannot continue functioning regardless of how healthy it may be.

A modern data centre operates in much the same way. Servers represent the “brain” of the facility, but they rely upon an extensive network of electrical, mechanical, telecommunications, safety and control systems that work together as a single integrated organism.

In fact, many of the systems that dominate the physical appearance of a hyperscale data centre—large cooling plants, electrical substations, generators, pumps, cooling towers and security infrastructure—perform no computing whatsoever. Their role is simply to create the conditions under which computing can continue uninterrupted.

This distinction is important because it explains why public discussions about data centres often focus on infrastructure that appears unrelated to digital services. Noise typically originates from cooling equipment rather than servers. Water consumption is associated primarily with cooling technologies. Air emissions arise mainly from backup power systems. Land requirements reflect not only computing halls but also electrical infrastructure, cooling equipment, fire protection systems and security setbacks.

Understanding these supporting systems therefore helps explain not only how data centres operate, but also why they create particular environmental, social and planning considerations.

MDCO Insight: Most of the infrastructure inside a data centre exists not to perform computing, but to ensure that computing can continue under almost all foreseeable conditions.

Reliability: The Design Principle Behind Everything

If there is one idea that explains almost every engineering decision within a modern data centre, it is reliability.

Most buildings are designed on the assumption that occasional interruptions are acceptable. An office building can usually tolerate a brief power outage or a temporary failure of its air-conditioning system. Occupants may experience inconvenience, but work can often resume once the problem is resolved.

Digital infrastructure operates under very different expectations.

Today, data centres host cloud computing platforms, banking systems, telecommunications networks, healthcare applications, government services, logistics systems, online retail platforms and increasingly artificial intelligence workloads. Millions of people and businesses now expect these services to remain available every hour of every day, regardless of equipment failures, utility interruptions or extreme weather events.

This expectation has fundamentally shaped the design of modern data centres.

Rather than asking how a system should operate under normal conditions, engineers begin by asking a different question:

What happens if something fails?

What if the incoming utility supply is interrupted?

What if a transformer develops a fault?

What if a cooling unit stops operating?

What if a fibre-optic cable is accidentally severed?

What if maintenance needs to be performed while the facility remains operational?

The answer to each of these questions is usually the same.

The service should continue.

Achieving this objective requires a design philosophy built around redundancy, resilience and fault tolerance. Critical equipment is duplicated. Independent power paths are provided. Cooling systems are backed up. Telecommunications connections enter the facility through multiple routes. Maintenance can often be performed without interrupting operations.

These concepts are reflected in internationally recognised design standards such as the Uptime Institute Tier classifications, which describe increasing levels of redundancy and operational resilience. Although not every data centre is designed to the highest tier, the underlying philosophy is consistent across the industry: critical services should remain available despite equipment failures or planned maintenance.

Of course, higher reliability does not come without consequences.

Every additional backup generator requires land, fuel storage and maintenance. Every duplicate cooling system consumes additional capital and space. Every redundant electrical path increases engineering complexity. Greater resilience often means greater resource consumption, higher construction costs and larger physical footprints.

Here lies one of the recurring themes throughout the MDCO Explain Series.

Many of the systems that later become subjects of environmental assessment or community concern were originally introduced to satisfy society’s own expectations for uninterrupted digital services. The challenge facing designers is therefore not simply to maximise reliability, but to achieve an appropriate balance between reliability, cost, environmental performance, safety, sustainability and community impact.

Different stakeholders naturally place different emphasis on these objectives. Customers may demand near-perfect availability. Investors seek commercially viable projects. Regulators focus on public safety and compliance. Utilities prioritise grid stability. Communities may be concerned about noise, emissions or land use. Governments seek both economic development and sustainable infrastructure.

The modern data centre is ultimately the product of balancing all of these legitimate interests rather than optimising any single one.

MDCO Insight: The architecture of a modern data centre reflects society’s expectation that digital services should remain continuously available, even when individual components fail.

The Electrical System: Feeding Continuous Computation

Electricity is the foundation upon which every other system inside a data centre depends.

Without electrical power, servers stop processing information, storage systems become inaccessible, cooling systems cease operating, security systems fail, and telecommunications equipment loses connectivity. In practical terms, electricity is the lifeblood of the entire facility.

As data centres have grown from small enterprise server rooms into hyperscale campuses supporting cloud computing and artificial intelligence, their electrical requirements have expanded dramatically. Modern hyperscale facilities may require tens or even hundreds of megawatts of continuous electrical demand. Emerging AI campuses are expected to require even greater capacities.

Meeting these requirements involves far more than connecting a building to the local electricity supply.

In Malaysia, large hyperscale data centres increasingly connect directly to the national transmission network through dedicated 132 kV or 275 kV substations. High-voltage electricity is received from the transmission grid before passing through transformers, switchgear and multiple distribution systems that progressively reduce the voltage to levels suitable for servers and other equipment.

Yet delivering electricity is only part of the challenge.

The greater challenge is ensuring that electricity remains available continuously.

To achieve this, modern data centres typically incorporate multiple layers of resilience. These may include independent utility feeds, redundant transformers, duplicate switchboards, uninterruptible power supply (UPS) systems equipped with batteries, standby diesel generators, automatic transfer systems and sophisticated power monitoring platforms that continuously assess the health of the electrical network.

Each layer serves a different purpose. Batteries provide immediate power during very short interruptions. Generators sustain operations during prolonged outages. Redundant distribution systems allow maintenance to be performed without interrupting critical services. Monitoring systems detect abnormalities before they develop into failures.

Collectively, these systems transform ordinary electricity supply into highly reliable electrical infrastructure capable of supporting continuous digital services.

However, every engineering solution also creates wider consequences beyond the facility itself.

Large substations occupy land and require significant investment. Transformers and generators generate noise. Backup generators produce emissions during routine testing and emergency operation. High electricity demand requires reinforcement of transmission infrastructure and influences long-term energy planning. Even the electricity itself carries different environmental implications depending on how it is generated.

These wider consequences explain why the design of electrical systems involves far more than engineering calculations alone.

Customers seek uninterrupted service. Utilities focus on network stability. Investors evaluate cost and return. Regulators establish safety and environmental requirements. Governments consider energy security and national competitiveness. Communities may focus on local impacts such as noise, visual appearance and air quality.

None of these perspectives is inherently right or wrong. Each reflects legitimate interests within a much larger system.

The role of engineers is therefore not simply to maximise reliability or minimise cost, but to balance these competing objectives in a way that allows the digital economy to function while managing broader social, environmental and economic outcomes responsibly.

As data centres continue to expand in size and importance, this balancing exercise is likely to become even more significant than the engineering technologies themselves.

MDCO Insight: The scale of a data centre’s electrical infrastructure reflects not only the demand for computing power, but society’s collective expectation that digital services should remain available under almost all circumstances.

Cooling Systems: Managing the Heat Produced by Computing

Every unit of electricity consumed by computing equipment ultimately becomes heat. This is a fundamental principle of physics and one of the defining engineering challenges of modern data centres.

Without effective cooling, server temperatures would quickly rise beyond safe operating limits, reducing performance, shortening equipment lifespan, and increasing the likelihood of service interruptions. Cooling systems therefore exist not as optional additions, but as essential infrastructure that enables reliable computing.

Historically, most data centres relied on air-based cooling, where chilled air is circulated through server rooms to remove heat. As computing densities have increased, particularly with the growth of artificial intelligence (AI), newer technologies such as chilled-water systems, liquid cooling, and direct-to-chip cooling are becoming increasingly important.

Each approach involves different engineering considerations. Air cooling generally requires less water but may consume more electricity. Liquid cooling can remove heat more efficiently from high-density equipment but introduces additional complexity in system design and operation. Designers must therefore balance reliability, energy efficiency, water availability, climate conditions, operating costs, and future scalability.

Cooling systems also explain many of the environmental discussions surrounding data centres. Fans, pumps, chillers and cooling towers consume electricity, generate noise, reject heat to the surrounding environment, and in some cases require significant quantities of water. These impacts are not the purpose of the facility, but rather the consequence of maintaining computing equipment within safe operating conditions.

As AI workloads continue to increase computing densities, cooling technologies are likely to become one of the industry’s most important areas of innovation.

MDCO Insight: Many public discussions about data centre sustainability are, at their core, discussions about how best to manage the heat generated by modern computing.

Water Systems: Balancing Reliability, Efficiency and Resource Stewardship

Water is one of the most widely discussed aspects of modern data centre development, yet its role is often oversimplified.

Depending on facility design, water may support cooling systems, fire protection, equipment maintenance, sanitation, and landscape management. While cooling usually accounts for the largest operational demand, the amount of water consumed varies significantly depending on the technology selected and local climatic conditions.

Design decisions frequently involve trade-offs rather than simple optimisation. Systems that minimise water consumption may require additional electricity, while systems that reduce electricity use may consume more water. The most appropriate solution therefore depends not only on engineering considerations but also on local resource availability, environmental priorities, and regulatory requirements.

For communities, water represents more than an engineering input. It is a shared public resource. Questions surrounding water use therefore often extend beyond operational efficiency to broader issues of sustainability, resource allocation, and long-term resilience.

Increasingly, operators are responding through water-efficient cooling technologies, recycled water schemes, rainwater harvesting, and improved monitoring of water performance.

MDCO Insight: The question is increasingly shifting from how much water a data centre uses to how responsibly that water is managed within its local context.

Monitoring and Control Systems: The Facility’s Nervous System

A modern data centre contains thousands—sometimes millions—of individual components that must operate together with extraordinary precision.

Monitoring and control systems act as the facility’s nervous system. Sensors continuously measure electrical loads, temperatures, humidity, equipment status, network performance, security events, and environmental conditions throughout the site.

These data are integrated through systems such as Building Management Systems (BMS), Electrical Power Monitoring Systems (EPMS), and Data Centre Infrastructure Management (DCIM) platforms, enabling operators to detect abnormalities before they develop into service disruptions.

Increasingly, artificial intelligence and predictive analytics are being incorporated to forecast equipment failures, optimise maintenance schedules, and improve operational efficiency.

Although these systems receive relatively little public attention, they play a significant role in improving reliability while reducing unnecessary resource consumption through better operational decision-making.

MDCO Insight: Modern data centres increasingly rely on digital intelligence not only to deliver computing services, but also to operate themselves more safely and efficiently.

Fire Protection and Safety Systems: Preparing for Rare but High-Consequence Events

Modern data centres are designed to minimise the likelihood and consequences of fire.

Fire protection systems combine early detection technologies, compartmentation, suppression systems, emergency procedures, and coordination with emergency services to protect both human life and critical digital infrastructure.

Different facilities may employ different suppression methods depending on operational requirements. Water-based systems remain common in many building areas, while specialised clean-agent suppression systems are often used to protect sensitive computing equipment.

Like many engineering systems within a data centre, fire protection involves balancing multiple objectives. Systems must provide effective protection while minimising unnecessary damage to equipment, environmental impacts, operational disruption, and lifecycle costs.

Although major incidents are relatively uncommon, society’s increasing dependence on digital services means that preparedness remains essential.

MDCO Insight: Fire protection systems are designed not only to protect buildings, but also to safeguard the digital services upon which modern society increasingly depends.

Telecommunications Infrastructure: Connecting Malaysia to the Global Digital Economy

Computing has little value if users cannot access it.

Telecommunications infrastructure connects data centres to customers, businesses, cloud platforms, internet exchanges, telecommunications operators, and other data centres across the world.

Multiple fibre routes are often installed to improve resilience and reduce the risk of service interruption should a cable be damaged.

Unlike power and cooling systems, telecommunications infrastructure produces relatively few direct environmental impacts. Nevertheless, it determines much of a data centre’s commercial value, resilience, and strategic importance.

Malaysia’s growing international connectivity has become one of the factors supporting the country’s emergence as a regional data centre hub.

MDCO Insight: A data centre’s value depends not only on its computing capacity, but also on how effectively it connects people, businesses and digital services.

Security Systems: Balancing Protection and Transparency

Data centres protect some of society’s most valuable digital assets.

Physical security measures typically include perimeter fencing, surveillance systems, access controls, security personnel, and secure operating procedures. These complement cybersecurity measures designed to protect customer information, financial transactions, government systems, and critical digital services.

Security illustrates one of the recurring tensions within modern infrastructure.

Communities often seek greater transparency regarding nearby developments, while operators must protect sensitive information that could create security vulnerabilities.

Both objectives are legitimate. Achieving an appropriate balance requires careful governance, regulatory oversight, and mutual trust.

MDCO Insight: Security measures often protect risks that are largely invisible to the public, making communication and trust particularly important.

Sustainability Systems: From Operational Efficiency to Long-Term Resilience

Sustainability has become an integral design objective rather than an additional consideration.

Many modern facilities now incorporate renewable energy procurement, battery energy storage, energy optimisation software, water conservation technologies, waste management programmes, environmental monitoring, and carbon reporting systems.

These initiatives are driven by multiple factors, including customer expectations, investor requirements, government policies, international standards, and corporate sustainability commitments.

Importantly, sustainability is no longer viewed solely through an environmental lens. Increasingly, operators recognise that environmental stewardship, operational resilience, economic competitiveness, and long-term business performance are closely interconnected.

The pace of innovation in these areas is expected to accelerate as computing demand continues to grow.

MDCO Insight: Sustainability is increasingly becoming a core design principle rather than simply a reporting requirement.

The Observatory Perspective

Viewed individually, each system within a data centre performs a specific engineering function. Viewed collectively, they reveal something much broader.

Electrical systems reflect society’s expectation of uninterrupted digital services. Cooling systems illustrate the physical realities of high-performance computing. Telecommunications networks connect Malaysia to the global digital economy. Security systems protect increasingly valuable digital assets. Sustainability initiatives demonstrate how industry continues to respond to evolving environmental and societal expectations.

Many of the impacts discussed in public debates—including electricity demand, water use, noise, heat and land requirements—are not independent issues. They are interconnected consequences of designing facilities that are expected to operate continuously, securely and reliably.

Understanding these relationships helps move discussions beyond simple narratives of either economic opportunity or environmental concern. Instead, they become discussions about balancing multiple legitimate objectives that may sometimes reinforce one another and at other times compete.

One role of the Malaysia Data Centre Observatory (MDCO) is to support this broader understanding by presenting data centre development through an elevated, balanced and systems-based perspective. Better understanding does not remove trade-offs, but it can help stakeholders identify more informed, transparent and sustainable pathways forward.

MDCO Insight: Better decisions become more achievable when stakeholders understand not only how data centres operate, but why they are designed the way they are.

Selected References

• International Energy Agency (IEA). Energy and AI (2025); Data Centres and Data Transmission Networks (2024). https://www.iea.org
• Uptime Institute. Tier Standard: Topology and global data centre reliability guidance. https://uptimeinstitute.com
• International Organization for Standardization (ISO). ISO/IEC 22237 – Information technology — Data centre facilities and infrastructures. https://www.iso.org
• Telecommunications Industry Association (TIA). ANSI/TIA-942 Telecommunications Infrastructure Standard for Data Centers. https://tiaonline.org
• ASHRAE. Thermal Guidelines for Data Processing Environments. https://www.ashrae.org
• The Green Grid. Guidance on Power Usage Effectiveness (PUE), Water Usage Effectiveness (WUE), and data centre sustainability metrics. https://www.thegreengrid.org
• National Fire Protection Association (NFPA). NFPA 75 and NFPA 76 standards for information technology equipment and telecommunications facilities. https://www.nfpa.org
• Malaysia Digital Economy Corporation (MDEC). Information on Malaysia’s digital economy and digital infrastructure initiatives. https://mdec.my
• Tenaga Nasional Berhad (TNB). Information on Malaysia’s electricity transmission and grid infrastructure. https://www.tnb.com.my

Citation

Malaysia Data Centre Observatory (MDCO). Why Data Centres Are Built This Way? MDCO Explain Series No. E03 (Version 1.0, July 2026).

MDCO Note

This article forms part of the Malaysia Data Centre Observatory (MDCO) Explain Series, which aims to improve public understanding of data centre development through evidence-based, accessible and balanced analysis. It is intended for educational and informational purposes only and does not constitute legal, engineering, planning, environmental or professional advice.

Malaysia’s rapidly evolving data centre ecosystem includes facilities developed, owned or operated by organisations such as AirTrunk, Amazon Web Services (AWS), Bridge Data Centres, DayOne, EdgeConneX, Google, K2 Data Centres, Microsoft, NTT Global Data Centers, Princeton Digital Group (PDG), ST Telemedia Global Data Centres (STT GDC), STACK Infrastructure, Vantage Data Centers, YTL Data Centre Park and many others. MDCO is independent of these organisations, as well as governments, regulators, utilities and advocacy groups. Its role is to facilitate transparency, structured understanding and equal access to information by presenting publicly verifiable evidence, relevant context and multiple stakeholder perspectives. MDCO does not endorse, oppose or advocate for any particular organisation, project or policy position.