A09 Understanding Low-Frequency Noise from Data Centres

ByMalaysia Data Centre Observatory

Why One Technical Issue Has Become a Global Community Concern

Overview and Reading Guide to the MDCO Low-Frequency Noise Series

Key Takeaways

  • Low-frequency noise (LFN) has become one of the most frequently discussed environmental issues associated with modern hyperscale data centres.
  • The issue extends beyond acoustics and involves engineering, environmental science, public health research, planning, governance, regulation and community engagement.
  • Different stakeholders often reach different conclusions because they examine the issue through different professional disciplines, responsibilities and objectives.
  • Understanding low-frequency noise requires consideration of both technical evidence and the broader governance ecosystem within which modern data centres are designed, approved and operated.
  • This article introduces the issue and serves as the central guide to MDCO’s continuing series examining low-frequency noise from multiple perspectives.

Why This Topic Deserves Its Own Series

Modern data centres are often discussed in terms of artificial intelligence, cloud computing, economic investment, electricity demand and water consumption. These topics frequently dominate policy announcements, media coverage and public debate because they reflect the rapidly expanding role of digital infrastructure in modern society.

Yet for many communities living near large data centres, a different issue has become increasingly prominent: the continuous low-frequency humming or droning sound associated with mechanical cooling equipment.

Although this may initially appear to be a relatively narrow engineering issue, it quickly expands into questions that extend well beyond engineering. Discussions about low-frequency noise frequently involve planning authorities, environmental regulators, acoustic consultants, medical researchers, equipment manufacturers, local governments, facility operators and nearby residents. The same issue may be discussed simultaneously in terms of physics, environmental compliance, public health, planning policy, community wellbeing and infrastructure design.

This diversity of perspectives explains why low-frequency noise has become one of the more challenging aspects of modern data centre development. Different stakeholders often ask different questions, apply different standards of evidence and pursue different objectives. Engineers may focus on equipment performance and acoustic measurements. Regulators may consider compliance with environmental standards. Researchers may examine the scientific evidence relating to human health. Communities may simply ask why a persistent sound has become part of their everyday environment.

None of these perspectives is necessarily incorrect. Rather, each represents one part of a much larger system.

This reflects a recurring theme throughout the Malaysia Data Centre Observatory (MDCO). Complex infrastructure issues rarely belong to a single discipline. They emerge from the interaction of technology, governance, regulation, economics and society.

For this reason, MDCO is undertaking this dedicated series on low-frequency noise. Rather than approaching the issue from only one perspective, the series examines how engineering systems generate sound, how sound is measured, what scientific research currently indicates, how different jurisdictions regulate the issue, how different stakeholders understand their responsibilities, and what technologies are available to reduce potential impacts.

The objective is not to determine whether one stakeholder is right and another is wrong. Instead, it is to improve structured understanding of an issue that increasingly sits at the intersection of digital infrastructure and community life.

MDCO Insight: Low-frequency noise is one of the few data centre issues that simultaneously involves engineering, environmental science, health research, planning, governance and community trust.

What Is Low-Frequency Noise?

Sound is commonly described according to its frequency, measured in hertz (Hz). Human hearing generally extends from approximately 20 Hz to 20,000 Hz, although sensitivity varies considerably across individuals and tends to decrease with age.

Low-frequency noise typically refers to sound within the lower portion of the audible spectrum, often below around 200 Hz. Unlike higher-frequency sounds, low-frequency noise is frequently described using terms such as humming, droning, rumbling or throbbing. Some individuals report not only hearing the sound but also experiencing sensations of vibration or pressure, particularly within enclosed spaces.

One characteristic that distinguishes low-frequency noise from many other forms of environmental noise is its physical behaviour. Because of its relatively long wavelength, low-frequency sound can travel over considerable distances under certain conditions, penetrate buildings more readily than higher-frequency sound, and become more noticeable during quiet periods, particularly at night when background environmental noise is lower.

Importantly, not all low-frequency sound is problematic. Many aspects of modern life generate low-frequency sound, including road traffic, aircraft, industrial facilities, heating, ventilation and air-conditioning (HVAC) systems, electrical transformers, ships and wind turbines. Whether such sound becomes a community concern depends on many interacting factors, including sound level, frequency characteristics, duration, operating patterns, local environmental conditions, building construction, background noise and individual perception.

These characteristics make low-frequency noise more complex than a simple measurement of loudness. Understanding the issue requires consideration of both acoustical physics and human experience.

Later articles in this series will examine these characteristics in considerably greater detail.

MDCO Insight: Low-frequency noise differs from many other forms of environmental noise because its physical characteristics, measurement and human perception are often more complex than simple loudness alone.

Why Data Centres Can Produce Low-Frequency Noise

To understand why data centres may generate low-frequency noise, it is helpful to return to one of the central concepts introduced in →E03 – Why Data Centres Are Built This Way.

The primary purpose of a data centre is to provide computing capacity. However, computing inevitably generates heat. Every watt of electricity consumed by servers is ultimately converted into heat that must be removed if equipment is to operate reliably.

Cooling therefore becomes one of the largest engineering systems within any modern data centre.

Maintaining appropriate operating temperatures requires an extensive network of mechanical and electrical equipment. Depending on the facility design, this may include computer room air handlers (CRAH), computer room air conditioners (CRAC), chillers, cooling towers, dry coolers, condenser units, pumps, heat exchangers and large ventilation fans. Many of these systems operate continuously throughout the day and night because the computing equipment they support also operates continuously.

The purpose of these systems is not to generate noise. Their purpose is to transfer heat safely and efficiently away from sensitive computing equipment.

Nevertheless, whenever air is moved at high volumes, fluids circulate through piping systems, or mechanical equipment operates continuously, sound is inevitably produced. Some components also generate vibrations that propagate through supporting structures and the surrounding air. Under certain operating conditions, these sounds may contain significant low-frequency components that extend beyond what is commonly associated with everyday environmental noise.

It is important to note that cooling equipment is not the only potential source of low-frequency noise within a data centre. Electrical transformers, standby generators during testing or operation, large pumps, and other mechanical or electrical systems can also produce low-frequency sound under particular conditions. However, during normal day-to-day operation, the continuously operating cooling system is generally regarded as the most significant and persistent source of environmental low-frequency noise. Accordingly, this series focuses primarily on cooling-system noise while recognising that the relative contribution of other equipment varies according to the design, layout and operating characteristics of each facility.

This distinction is important because the cooling system is not designed to produce sound; it is designed to remove heat so that servers can operate reliably around the clock. Any noise generated is therefore an unintended consequence of delivering the continuous digital services upon which businesses, governments and society increasingly depend.

Recognising this relationship helps explain why discussions about data centre low-frequency noise cannot be separated from broader questions of reliability, resilience, energy efficiency and infrastructure design. Measures that reduce noise may also influence cooling performance, energy consumption, capital cost or operational flexibility. Understanding these interdependencies is therefore essential for evaluating both the challenges and the potential solutions associated with low-frequency noise.

Subsequent articles in this series will examine each of these cooling systems individually, explain how they generate different types of sound, and discuss the engineering approaches available to reduce their acoustic impacts.

MDCO Insight: Low-frequency noise originates not from computing itself, but from the infrastructure required to keep computing operating safely and continuously.

Why This Issue Matters

If low-frequency noise were simply an engineering issue, it would likely be resolved entirely through technical design standards. In practice, however, the issue has become considerably more complex.

In many situations, facility operators may report that measured sound levels comply with applicable regulatory requirements, while nearby residents continue to report persistent disturbance. Acoustic consultants may present measurement data demonstrating compliance with established assessment methods, while communities describe experiences that they believe are not fully reflected by those measurements. Researchers may identify areas where scientific evidence remains uncertain, while policymakers must nevertheless make planning and regulatory decisions based on the information currently available.

These differing perspectives do not necessarily indicate that any stakeholder is acting in bad faith. More often, they reflect the reality that different professions examine different aspects of the same issue.

An engineer may ask whether equipment is operating as designed.

An acoustician may ask whether sound measurements comply with recognised methodologies.

A regulator may ask whether statutory requirements have been satisfied.

A medical researcher may ask what evidence exists regarding particular health outcomes.

A local resident may simply ask why a persistent humming sound can still be heard inside a home during the night.

Each question is legitimate within its own context.

The challenge arises because these questions do not always produce the same answer, nor are they intended to.

One of the recurring themes throughout MDCO has been the importance of reducing information asymmetry. Low-frequency noise illustrates this particularly well. Each stakeholder often possesses detailed knowledge within their own field, yet relatively limited understanding of the objectives, constraints and evidence considered by others.

Improving understanding therefore requires more than simply collecting additional measurements or publishing additional regulations. It requires bringing together multiple perspectives into a shared framework that enables constructive discussion about engineering, health, governance and community outcomes.

MDCO Insight: Different stakeholders often reach different conclusions about low-frequency noise because they are asking different, but equally legitimate, questions.

International Case Studies

Low-frequency noise (LFN) associated with data centres has emerged as a topic of discussion in several mature data centre markets around the world. As hyperscale facilities have become larger, more numerous and increasingly integrated into communities, questions have arisen regarding how continuous mechanical operation, acoustic characteristics and community experience should be understood and managed.

These international examples do not suggest that all data centres create unacceptable noise impacts, nor that every community will experience similar concerns. Instead, they illustrate why low-frequency noise requires consideration from multiple perspectives, including engineering design, acoustic science, operational management, regulatory frameworks and community expectations.

For Malaysia, which is still in the early stages of large-scale data centre expansion, these experiences provide an opportunity to understand potential challenges and solutions before they become widespread.

Amazon Data Centre, Manassas, United States

When Cooling Infrastructure Becomes a Community Noise Issue

Northern Virginia is one of the world’s largest data centre markets, and the rapid growth of facilities has brought increased attention to operational noise from mechanical infrastructure.

In the Manassas area, residents raised concerns regarding persistent low-frequency noise associated with data centre operations, describing continuous humming or droning sounds that were particularly noticeable during quieter periods.

Following engagement between stakeholders and acoustic investigations, mitigation measures were implemented to reduce the impact.

The Manassas experience demonstrates that low-frequency noise can emerge as an unintended consequence of essential cooling infrastructure and that effective management may require detailed acoustic assessment and targeted engineering solutions.

Northern Virginia, United States

When Multiple Data Centres Create Cumulative Acoustic Effects

As discussed in → A06 – Loudoun County Case Study – When Infrastructure Becomes Background Noise, Northern Virginia provides a broader example of how low-frequency noise discussions can evolve when multiple hyperscale facilities operate within a concentrated geographic area.

In a single-facility context, noise assessment generally focuses on the contribution of one project. However, large data centre clusters introduce additional questions regarding cumulative effects, where multiple cooling systems and mechanical installations operate continuously within the same region.

The challenge is not necessarily whether an individual facility exceeds a specific noise limit, but whether the combined acoustic environment created by many facilities changes the experience of nearby communities over time.

This raises broader planning questions:

  • How should cumulative noise effects be assessed?
  • Should future developments consider existing acoustic conditions?
  • How can expansion plans balance digital infrastructure growth with community expectations?

The Northern Virginia experience demonstrates that low-frequency noise can become a regional planning consideration rather than only an individual facility issue.

Mount Pleasant, Wisconsin, United States

When AI Infrastructure Raises New Acoustic Questions

The emergence of large artificial intelligence (AI) infrastructure has introduced a new phase of data centre development, with facilities requiring significantly increased computing capacity and supporting mechanical systems.

In Mount Pleasant, Wisconsin, community concerns surrounding a large data centre development have included operational noise from the infrastructure required to support high-performance computing. Discussions have highlighted questions regarding continuous mechanical sound, acoustic assessment methods and the ability of existing approaches to address the characteristics of modern AI-oriented facilities.

The case reflects a broader trend: as data centres become larger and more power-intensive, the supporting cooling and electrical systems required to maintain operation may also increase in scale.

Although individual projects and circumstances vary, the Mount Pleasant experience demonstrates why acoustic considerations may become increasingly important during early planning and design stages of future AI-scale infrastructure.

Compass Data Centres, United States

When Engineering Solutions Reduce Acoustic Impacts

While low-frequency noise is often discussed as a regulatory or community issue, it is also fundamentally an engineering challenge.

A case involving Compass Data Centres demonstrated how acoustic investigation can identify specific mechanical sources and support targeted mitigation measures. Through detailed assessment, engineers were able to identify contributing equipment characteristics and implement modifications to reduce acoustic impacts.

The case illustrates an important principle: effective noise management begins with understanding the source.

Low-frequency noise is influenced by factors including equipment selection, operating conditions, mechanical design, structural transmission, installation methods and acoustic treatment. As a result, mitigation strategies often require cooperation between equipment manufacturers, engineers, acoustic specialists and operators.

International Perspective

Together, these cases illustrate why low-frequency noise has become an important consideration in modern data centre development.

The issue begins with engineering systems designed to provide reliability. It becomes a community concern when persistent sound interacts with surrounding environments. It becomes a governance challenge when measurement methods, standards and responsibilities must be considered. Finally, it becomes a design opportunity when engineering solutions can reduce impacts while maintaining essential infrastructure performance.

For Malaysia, the lessons from international experience are not that data centre development should be restricted, but that potential impacts should be understood early and addressed through transparent, evidence-based approaches.

MDCO Insight: The most effective approach to low-frequency noise is not to view it solely as a conflict, but as a shared engineering, governance and community challenge requiring better understanding from all stakeholders.

Why Opinions Differ

One of the reasons low-frequency noise remains a challenging public issue is that different groups are often trying to answer different questions. Discussions can therefore appear contradictory even when each participant is acting in good faith.

An engineer designing a cooling system is primarily concerned with whether the equipment can reliably remove heat while meeting specified acoustic performance targets. The questions are often practical and measurable: How much sound does the equipment produce? Which frequencies dominate? Can noise be reduced without compromising reliability, efficiency or maintainability?

Medical researchers approach the issue differently. Their interest lies in whether exposure to low-frequency noise is associated with measurable health outcomes. They examine epidemiological studies, laboratory research, sleep quality, physiological responses, stress indicators and the strength of available scientific evidence. Their conclusions are necessarily influenced by study design, exposure conditions and the quality of available research.

Acousticians focus on the behaviour of sound itself. They ask how low-frequency sound propagates through the environment, how it interacts with buildings and terrain, and how it should be measured. They also recognise that conventional environmental noise metrics, which were developed primarily for higher-frequency sounds, may not always capture every aspect of low-frequency sound experienced by nearby communities.

Regulators must translate scientific knowledge into practical policy. They face questions such as which standards should apply, what constitutes reasonable protection, how compliance should be assessed, and how different public interests should be balanced. Regulatory frameworks are often developed using available scientific evidence while also considering practicality, consistency, enforceability and wider policy objectives.

Communities often begin from a far more immediate perspective. Their question is not how sound is measured or regulated, but why they hear or feel a persistent hum, particularly during the evening or at night. For affected residents, the lived experience frequently becomes more important than technical terminology or measurement methodology.

Data centre operators face another set of considerations. They must maintain continuous operation for critical digital infrastructure while responding to legitimate community concerns. Reducing low-frequency noise may involve equipment selection, acoustic treatment, revised operating strategies or changes to site layout, each of which can affect cost, energy efficiency, resilience or future expansion.

None of these perspectives is inherently incorrect. Rather, each reflects a different responsibility within the wider ecosystem.

Many public disagreements therefore arise because different stakeholders are attempting to answer different questions using different forms of evidence.

MDCO Insight: Understanding low-frequency noise requires integrating knowledge from engineering, acoustics, medicine, environmental science, regulation and community experience rather than relying on any single discipline alone.

What This Series Will Explore

Low-frequency noise (LFN) is a multidisciplinary issue that cannot be fully understood from a single perspective.

Engineers examine how sound is generated and controlled. Acoustic specialists study how sound travels and how it should be measured. Researchers investigate human perception and potential health effects. Regulators consider appropriate governance frameworks. Communities experience the practical consequences of infrastructure development. Operators and designers seek solutions that maintain reliability while reducing impacts.

This series therefore examines low-frequency noise as a whole-system issue, exploring the interaction between engineering, science, governance, community experience and practical solutions.

Understanding the Engineering

This section explains why data centres produce low-frequency noise and how engineering systems influence acoustic outcomes.

Building upon → E03 — Why Data Centres Are Built This Way, these articles examine the relationship between cooling infrastructure, reliability requirements and operational sound.

Topics include:

  • How data centre cooling systems operate
  • Why cooling infrastructure is essential for continuous operation
  • Major equipment that may generate low-frequency sound, including fans, compressors, pumps, chillers and cooling towers
  • How mechanical vibration becomes airborne and structural noise
  • Why low-frequency sound behaves differently from higher-frequency sound
  • How equipment selection, system design and operating conditions influence acoustic performance

Articles:

Understanding Human Experience and Scientific Evidence

Engineering explains how sound is produced. This section examines how humans perceive and experience sound.

Low-frequency noise is often discussed through personal experience, scientific research and technical measurement. Understanding the relationship between these perspectives is essential because a sound that is technically measurable may not be experienced in the same way by every individual.

Topics include:

  • What low-frequency noise means from an acoustic perspective
  • How low-frequency sound differs from ordinary environmental noise
  • Scientific research relating to annoyance, sleep disturbance and wellbeing
  • Why individuals may experience the same sound differently
  • The relationship between sound exposure, perception and reported impacts
  • How low-frequency noise is measured and why measurement can be challenging

Rather than advocating a particular position, these articles examine available evidence, areas of scientific understanding and areas where uncertainty remains.

Articles:

When Low-Frequency Noise Becomes a Community Conflict

Technical noise issues can become community conflicts when different stakeholders interpret the same situation differently.

A data centre may comply with applicable requirements, while nearby communities may continue to experience disturbance or concern. Understanding this gap requires examining not only sound levels, but also trust, communication, expectations and the limitations of existing assessment approaches.

Topics include:

  • Why technically compliant projects may still generate community concerns
  • The difference between regulatory compliance and community acceptance
  • How information gaps influence public perception and stakeholder trust
  • How noise concerns develop into wider planning and governance issues
  • Lessons from international data centre noise cases
  • The role of early engagement and transparent communication

Articles:

Governance and Accountability

Low-frequency noise raises important questions about how responsibilities are defined, managed and communicated.

Governance extends beyond legislation alone. It involves interactions between regulators, planning authorities, technical professionals, operators, customers, investors and communities.

Topics include:

  • International approaches to regulating environmental noise and low-frequency sound
  • Malaysian legislation, regulatory responsibilities and assessment frameworks
  • Planning conditions, environmental requirements and operational obligations
  • Roles of local authorities, environmental regulators, technical consultants and facility operators
  • How technical standards interact with statutory requirements
  • Corporate governance considerations, including operator policies, environmental commitments, risk management frameworks, customer expectations and stakeholder engagement practices
  • MDCO Dialogue with regulators, industry participants, researchers and other stakeholders

As part of this series, MDCO intends to invite relevant stakeholders to clarify how existing frameworks address low-frequency noise and how these frameworks are interpreted in practice.

These dialogues are intended to improve transparency and understanding rather than advocate any particular regulatory outcome.

Articles:

Solutions and Mitigation

Understanding the issue is only the first step. This section examines practical approaches to reducing low-frequency noise while maintaining reliability, efficiency and operational requirements.

Topics include:

  • Acoustic modelling and assessment
  • Equipment selection and system configuration
  • Mechanical isolation and vibration control
  • Acoustic enclosure and attenuation technologies
  • Cooling system optimisation
  • Site planning and building design approaches
  • Emerging technologies and future design considerations
  • Trade-offs between noise reduction, cost, energy efficiency and reliability

Articles:

  • (More Articles Under development — link will be included)

A Whole-System Perspective

The final articles bring together the engineering, scientific, governance and community perspectives examined throughout the series.

Low-frequency noise is not solely an acoustic issue. It reflects the interaction between infrastructure design, digital service demands, regulatory frameworks, community expectations and long-term development objectives.

Topics include:

  • Lessons from international case studies
  • Balancing reliability, sustainability and community outcomes
  • Future challenges from increasing data centre scale and AI infrastructure
  • Research gaps and future areas of improvement
  • MDCO’s integrated systems perspective

The objective is not to identify a single “correct” answer, but to provide a structured understanding of how different stakeholders view, assess and manage low-frequency noise.

Articles:

  • (Under development — link will be included)

Why MDCO Is Examining This Issue

The Malaysia Data Centre Observatory was established to improve transparency, structured understanding and equal access to information across Malaysia’s data centre ecosystem.

Low-frequency noise exemplifies the type of public-interest issue for which this approach is particularly valuable.

Different stakeholders often possess different pieces of the overall picture.

Engineers understand cooling technologies.

Acousticians understand sound propagation.

Medical researchers study potential health impacts.

Regulators interpret statutory responsibilities.

Operators understand operational constraints.

Communities experience the day-to-day consequences of nearby infrastructure.

Each perspective contributes valuable knowledge, yet no single perspective provides a complete understanding of the issue.

MDCO’s role is therefore not to determine who is right or wrong, nor to advocate for or against particular projects, technologies or policy outcomes.

Instead, MDCO seeks to reduce information asymmetry by bringing together publicly available information from multiple disciplines, organising it within a coherent framework, and presenting it in a manner that is accessible to all stakeholders.

This enables readers to better understand not only individual viewpoints, but also how those viewpoints relate to one another within the broader system.

MDCO Insight: Better outcomes are more likely when engineering, science, regulation and community experience are understood together rather than in isolation.

The Observatory Perspective

Throughout the Explain Series, MDCO has examined the data centre ecosystem from progressively broader perspectives.

→E03 — Why Data Centres Are Built This Way explained why complex cooling systems are essential for reliable digital infrastructure.

→A06 — Loudoun County Case Study demonstrated how large concentrations of data centres can generate community concerns, including those relating to environmental noise.

→E05 — The Governance Behind Modern Data Centres explored the overlapping frameworks that influence how facilities are designed, operated and managed.

→E06 — How Data Centres Are Approved in Malaysia explained the statutory institutions and approval mechanisms that oversee data centre development.

Low-frequency noise sits at the intersection of all these themes.

It is simultaneously an engineering challenge, an environmental issue, a regulatory consideration, a community concern and a governance question.

Understanding it therefore requires a systems perspective rather than a single-disciplinary approach.

As demand for digital infrastructure continues to grow, questions surrounding low-frequency noise are unlikely to disappear. The more important challenge is not whether the issue exists, but how societies can understand it more comprehensively, manage it more responsibly and balance the legitimate objectives of all stakeholders through transparent, evidence-based dialogue.

This series is intended as a contribution towards that broader understanding.

MDCO Insight: Complex public-interest issues are best understood by integrating multiple perspectives into a coherent whole rather than treating each discipline in isolation.

Selected References

  • World Health Organization (WHO) – Environmental Noise: Official resources on environmental noise, public health, and the Environmental Noise Guidelines for the European Region. https://www.who.int/
  • International Organization for Standardization (ISO) – Acoustics Standards: Official information on ISO standards relating to acoustics, environmental noise measurement and building acoustics, including the ISO 1996 series. https://www.iso.org/
  • ASHRAE – Heating, Ventilation, Air Conditioning and Refrigeration: Technical resources on HVAC systems, cooling technologies and building environmental control relevant to data centre cooling. https://www.ashrae.org/
  • Institute of Acoustics (IOA): Professional guidance and technical resources on environmental acoustics, noise assessment and acoustic engineering. https://www.ioa.org.uk/
  • World Green Building Council (WorldGBC): Resources on sustainable building design, operational performance, energy efficiency and the built environment. https://worldgbc.org/

Citation

Malaysia Data Centre Observatory (MDCO). Understanding Low-Frequency Noise from Data Centres. MDCO Analyse Series No. A09 (Version 1.0, July 2026).

MDCO Note

This article forms part of the Malaysia Data Centre Observatory (MDCO) Analyse 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, acoustic, environmental, medical or professional advice.

Low-frequency noise is a multidisciplinary subject spanning engineering, acoustics, environmental science, medicine, psychology, planning and public policy. MDCO does not advocate for or against any particular stakeholder, project, technology or regulatory position. Its role is to facilitate transparency, reduce information asymmetry and support informed discussion by explaining complex issues through publicly verifiable information and multiple complementary perspectives.

Similar Posts