Civil Law And Data Center Cooling System Failure Liability Claims In Europe

1. Introduction

Data-center cooling system failure liability arises when a cooling, HVAC, chilled-water, ventilation, environmental-control, or related mechanical/electrical system fails and causes damage to servers, stored data, business operations, customers, or the data-center operator.

Typical failures include:

chiller failure;

cooling-unit failure;

chilled-water instability;

inadequate cooling capacity;

pump failure;

defective HVAC design;

inadequate maintenance;

blocked heat exchangers;

incorrect airflow;

failure of redundancy;

inadequate temperature/humidity control;

defective controls or sensors;

poor commissioning;

failure to warn about a developing cooling problem.

A cooling failure can be particularly serious because data-center equipment generates substantial heat and generally requires continuous environmental control. EU data-center energy-efficiency guidance expressly recognizes that cooling is necessary to remove heat generated by IT equipment and maintain appropriate operating conditions. (EUR-Lex)

Important case-law qualification: there are relatively few reported European judgments specifically titled as “data-center cooling failure.” The strongest authorities therefore include several direct data-center cases, supplemented by closely analogous European HVAC, mechanical-installation, construction and maintenance cases. I identify that distinction below rather than treating every HVAC case as a data-center case.

2. Legal Basis of Liability

A cooling failure claim can arise from several legal relationships.

A. Contractual liability

The principal claim may be:

The cooling contractor failed to perform its contractual obligations.

Possible contractual obligations include:

installation;

design;

commissioning;

preventive maintenance;

emergency response;

24/7 monitoring;

temperature guarantees;

humidity control;

redundancy;

replacement of defective components.

B. Construction liability

Where the cooling system forms part of a newly constructed data center, liability may arise from:

defective design;

defective installation;

non-compliance with specifications;

unsuitable equipment;

inadequate capacity;

defective commissioning.

C. Professional negligence

Engineers, architects and specialist consultants may be liable where they negligently:

calculated cooling demand;

specified equipment;

designed chilled-water systems;

failed to account for heat loads;

failed to warn about foreseeable problems.

D. Tort/delict

Where the claimant and contractor are not contractually connected, national tort/delict law may become important.

E. Landlord/tenant liability

Where a data center operates from leased premises, the landlord may be responsible for building systems depending on:

the lease;

repair obligations;

service-charge provisions;

maintenance allocation;

statutory duties.

3. Why Cooling Failure Is Legally Different From Ordinary HVAC Failure

A normal office can sometimes tolerate temporary loss of air conditioning.

A data center generally cannot be treated in exactly the same way.

The Versailles Court of Appeal expressly observed in Azur Data Center v SCI Vendôme Bureaux that a data center's activity required a reinforced, continuously operational 24-hour cooling system and appropriately reliable electrical infrastructure. (Cour de Cassation)

Therefore, the contractual and factual context can be extremely important.

A claimant may argue:

“The contractor knew that the equipment was protecting continuously operating servers, so the required standard was higher than ordinary commercial air conditioning.”

But that does not automatically establish liability. The actual contract, technical specification, maintenance obligations and causation still have to be proved.

4. Case Law

Case 1 — Global Switch Amsterdam BV v Ingenieursbureau Wolter & Dros BV

Amsterdam Court of Appeal, ECLI:NL:GHAMS:2014:3037, 29 July 2014

Direct data-center cooling authority

This is one of the most important European cases for this subject.

Global Switch operated a data center in Amsterdam with a sophisticated climate-control system. It contracted with Wolter & Dros under a Service Maintenance Contract (SMC) concerning maintenance of equipment forming part of the climate-control system, including the drive system for the cooling units. (InView)

Black dust subsequently appeared on computer cabinets and equipment in a data-center hall.

Investigations attributed the dust to:

misalignment of pulleys;

increased wear of V-belts;

consequences for the climate-control equipment.

Global Switch claimed approximately €656,205.54 in damages and argued that the maintenance contractor had breached its obligations. (InView)

First-instance position

The trial court initially considered that Wolter & Dros had not acted as a proper contractor should have acted and considered it potentially liable.

However, Global Switch still had to prove causation — specifically, that timely intervention would have prevented the claimed damage.

It failed to establish that sufficiently.

Court of Appeal

The Amsterdam Court of Appeal ultimately concluded that there was no established breach of the contractor's maintenance obligation and rejected the claim. (InView)

Principle

Even in a highly critical data-center environment, the claimant must establish both the precise contractual maintenance obligation and the contractor's actual failure to perform it.

Importance

This case demonstrates three major principles:

data-center cooling maintenance can create specific contractual obligations;

technical expert evidence is central;

serious operational consequences do not eliminate the need to prove breach and causation.

5. Case 2 — Azur Data Center v SCI Vendôme Bureaux

Cour d'appel de Versailles, RG No. 18/02462, 18 February 2021

Direct data-center cooling/lease authority

Azur Data Center operated a data center in leased premises.

The case concerned repeated failures involving:

the cooling system;

electrical systems;

building infrastructure.

Azur alleged that the cooling failures caused:

interruptions lasting several hours;

loss of computer data;

operating losses;

commercial losses.

The tenant argued that the landlord had failed to satisfy its contractual obligations. (Cour de Cassation)

Court's analysis

The court found that the various cooling failures were principally attributable to fortuitous events and that the landlord had responded with the maintenance contractor within an undisputedly reasonable period.

The landlord had also agreed to renovation work.

The court therefore concluded that the landlord had acted with the diligence reasonably expected in the circumstances. (Cour de Cassation)

Importantly, the court recognized that the tenant's expectations were not those of an ordinary user because a data center required reinforced cooling operating continuously, 24 hours a day. (Cour de Cassation)

Principle

A data-center operator may have unusually demanding operational requirements, but the landlord's liability still depends on the contractual allocation of responsibility and proof of an actionable failure.

Practical significance

The case is especially relevant to:

data-center leases;

landlord obligations;

maintenance allocation;

service charges;

cooling-system breakdowns;

loss of data;

business interruption.

6. Case 3 — Hurley Palmer Flatt Ltd v Barclays Bank Plc

England and Wales High Court, Technology and Construction Court, [2014] EWHC 3042 (TCC)

Direct data-center cooling design dispute

This is a common-law English case, not a civil-law jurisdiction case, but it is an important European comparative authority.

Barclays engaged Hurley Palmer Flatt to provide mechanical and electrical engineering design services for a new data hall.

The dispute concerned the chilled-water system.

The system was alleged to be unstable, with fluctuations in the rate and direction of chilled-water flow.

The claim against the engineering consultant was valued at more than £4 million. (Bailii)

Legal significance

The case demonstrates how liability can arise at the engineering-design level, rather than only from physical failure of a cooling component.

A claimant can potentially argue:

defective design → unstable chilled-water system → inadequate cooling → data-center operational damage.

Principle

A professional engineering consultant may face contractual/professional liability where the cooling system it designs fails to perform the required engineering function.

Limitation

Because this is English common law, it should be used as comparative European authority, not as a civil-law precedent.

7. Case 4 — Global Switch (Property) Singapore Pte Ltd v Arup Singapore Pte Ltd

Singapore High Court, [2019] SGHC 122

This is outside Europe, so it is included only as a comparative data-center authority.

Global Switch operates data centers in several European cities, including London, Paris, Amsterdam and Frankfurt.

The dispute concerned engineering design for a data-center extension.

The court specifically recognized that cooling capacity is critical because sustained cooling failure can cause sensitive IT equipment to overheat. (eLitigation)

The court awarded only nominal damages of S$1,000 for Arup's breach concerning additional cooling, illustrating that:

breach and recoverable substantial loss are separate questions.

Principle

A proven engineering breach does not automatically establish substantial damages; the claimant must connect the breach to quantifiable loss.

Classification

Comparative authority — not European law.

It is useful because it directly analyses data-center cooling engineering in substantial technical detail.

8. Case 5 — Orange Personal Communications Services Ltd v Hoare Lea

England and Wales High Court, [2008] EWHC 223 (TCC)

Direct data-center HVAC construction dispute

Orange operated the Bristol Data Centre.

The construction project included installation of an air-conditioning system.

A flood occurred and allegedly damaged Orange's equipment.

The alleged cause was failure of a compression joint connecting the chilled-water system to an air-conditioning unit. Orange alleged defective installation and alternatively design/project-management failures. (vLex)

The litigation involved:

the main contractor;

HVAC subcontractor;

engineering consultant;

project manager.

Importance

Although the reported judgment principally concerned procedural/pre-action issues rather than final cooling-liability determination, it demonstrates a recurring problem:

A cooling-system incident can generate overlapping claims against designer, main contractor, specialist subcontractor and project manager.

Principle

Where a data-center mechanical failure arises from construction, liability must be allocated across the contractual chain.

Classification

Comparative English authority.

9. Case 6 — Oberlandesgericht Düsseldorf, 23 U 27/14, 15 December 2015

German mechanical-maintenance authority with data-center consequences

The Düsseldorf Higher Regional Court considered claims involving:

construction of a heating system;

maintenance;

technical installations;

sprinkler systems;

documentation;

a power outage affecting a data center.

The claimant sought substantial damages for alleged defects and maintenance failures.

The court examined whether maintenance defects could actually be established through the evidence.

It ultimately rejected several alleged maintenance claims because the necessary causal connection could not be established with sufficient certainty. A limited amount had previously been awarded for a data-center power outage. (Fundstellensuche)

Importance for cooling claims

Although the principal system was heating rather than data-center cooling, the legal methodology is highly relevant.

The court considered:

contractual maintenance obligations;

technical expert evidence;

causation;

proof of defective maintenance;

documentation obligations;

damages arising from interruption of technical infrastructure.

Principle

Technical failure litigation requires proof that the particular contractor's maintenance failure caused the particular loss.

This is directly analogous to a cooling-system case.

10. Case 7 — Cour d'appel de Rouen, RG No. 22/01579, 8 February 2024

Direct data-center construction/CVC authority

This French case concerned construction of a data center involving specialist technical lots.

The relevant works included:

heating;

conditioning;

ventilation (CVC);

electrical systems;

generators;

UPS;

low-voltage systems;

other technical infrastructure. (Cour de Cassation)

The court examined the allocation of responsibilities among:

architect/design professionals;

project participants;

specialist contractors;

subcontractors.

Legal significance

The case demonstrates the importance of determining who was responsible for the CVC/HVAC package.

In a complex data-center project, responsibility may be divided between:

architect;

M&E engineer;

general contractor;

specialist HVAC contractor;

equipment supplier;

commissioning contractor;

operator.

Principle

A claimant must identify the contractual and technical responsibility for the defective HVAC/CVC component rather than treating every participant in the data-center project as jointly responsible.

11. Case 8 — Cour d'appel d'Orléans, RG No. 21/01980

HVAC design and maintenance — close French analogy

This case concerned defects associated with ventilation/heating installation.

The court examined:

design responsibility;

information supplied by the client;

contractor expertise;

subcontracting;

inadequate ventilation;

absence of maintenance;

technical measurements concerning temperature and humidity.

The court held that the professional responsible for the design had a duty to obtain sufficiently precise information about the user's needs for correct sizing of the ventilation installation.

It also rejected an attempt to rely on subcontracting as an automatic defence. (Cour de Cassation)

Principle

A specialist construction professional cannot necessarily escape liability merely because technical work was subcontracted where the professional itself had responsibility for obtaining the information necessary to design the system correctly.

Data-center application

For data centers, this is particularly relevant to:

rack-density calculations;

heat-load calculations;

airflow requirements;

redundancy requirements;

humidity specifications;

future expansion capacity.

12. Case 9 — BGH, X ZR 26/97, 19 October 1999

German technical-installation liability

The German Federal Court of Justice considered installation of a specialized fire-protection system in a data center.

The specialist contractor had made representations about the safety characteristics of the system.

The case concerned:

specialist engineering;

representations by the contractor;

suitability of the system;

professional obligations;

contractual performance.

The BGH set aside the appellate judgment and remitted the matter for further consideration. (Anwalt24)

Relevance to cooling systems

The case supports a broader principle:

Where a specialist contractor undertakes to design/install a technically sophisticated protection system for a data center, contractual liability may depend on whether the promised technical performance was actually achieved.

Classification

This is a direct data-center technical-installation case, although the equipment was fire protection rather than cooling.

13. CJEU — Veronsaajien oikeudenvalvontayksikkö, C-215/19, 2 July 2020

This is not a liability case, but it is useful for understanding the legal character of data-center cooling.

The CJEU considered Finnish data-center hosting services.

The servers were housed in premises where:

data connections were available;

humidity was controlled;

heat was precisely controlled;

the environment was cooled to allow proper server operation. (EUR-Lex)

Importance

The case demonstrates that cooling is not merely an incidental building service in a data-center environment.

It can be an integral component of the hosting service itself.

That distinction can matter in contract litigation.

14. Legal Elements of a Cooling Failure Claim

A claimant should generally establish:

1. Contract

What agreement governed?

construction contract;

maintenance contract;

lease;

SLA;

engineering appointment;

equipment supply contract.

2. Duty

What exactly was promised?

For example:

“Maintain cooling capacity at X MW.”

or:

“Maintain temperature within specified parameters.”

or:

“Provide N+1 cooling redundancy.”

3. Breach

Examples:

failure to maintain;

incorrect installation;

inadequate capacity;

defective pump;

failed chiller;

blocked heat exchanger;

improper airflow;

failure to repair;

failure to warn.

4. Failure event

The claimant must establish:

date;

duration;

affected cooling zone;

temperatures;

humidity;

alarm records;

equipment status.

5. Damage

Possible losses include:

server damage;

replacement equipment;

data restoration;

emergency cooling;

customer compensation;

business interruption;

lost revenue;

migration costs;

repair costs.

6. Causation

The central chain is:

Cooling failure → temperature/humidity excursion → equipment/data/service damage → financial loss.

15. Causation Is Often the Most Difficult Issue

The Global Switch v Wolter & Dros case is particularly instructive.

Even though the claimant alleged defective maintenance and initially obtained a favorable interlocutory assessment, it ultimately failed to establish the necessary causal connection between the alleged maintenance failure and the damage. (InView)

A data-center claimant therefore needs more than:

“The cooling system had a problem.”

It should establish:

“The particular contractual failure caused the particular physical/financial damage.”

16. Technical Evidence

Cooling disputes usually require expert evidence concerning:

Thermal performance

inlet temperature;

outlet temperature;

rack temperature;

humidity;

thermal load.

Mechanical performance

chiller efficiency;

pump performance;

flow rate;

pressure;

valve operation;

heat-exchanger condition.

Airflow

hot-aisle/cold-aisle arrangement;

air recirculation;

containment;

airflow obstruction.

Capacity

installed cooling capacity;

actual IT load;

redundancy;

peak load;

future capacity.

Controls

sensors;

alarms;

BMS;

automatic switching;

emergency response.

17. N+1 and Redundancy

Modern data centers frequently use redundancy.

For example:

N+1 cooling

means that sufficient additional capacity exists so that failure of one component does not necessarily cause loss of cooling capacity.

If a contractor expressly promised N+1 redundancy and installed only N capacity, this can be strong evidence of contractual non-performance.

But the claimant still needs to establish:

the obligation existed;

the system was non-compliant;

the failure caused the claimed damage.

18. Design Defect vs Maintenance Defect

This distinction is essential.

Design defect

The system was incorrectly designed from the beginning.

Example:

Chiller capacity = 8 MW
Actual required capacity = 12 MW.

Installation defect

The design was adequate but the equipment was installed incorrectly.

Example:

Incorrect chilled-water piping.

Maintenance defect

The system was correctly designed and installed but not properly maintained.

Example:

Heat exchanger became blocked because scheduled maintenance was not performed.

Operational defect

The system worked properly but was incorrectly operated.

Example:

Operator disabled automatic cooling redundancy.

External cause

Example:

Utility failure → cooling failure.

Each category can produce different liability.

19. Landlord Liability

A data-center lease can create a difficult allocation of responsibilities.

The lease may state:

Landlord: building-level cooling plant.

Tenant: internal CRAC units.

or:

Landlord: entire HVAC system.

Tenant: server-room environmental monitoring.

The Azur Data Center litigation illustrates the importance of examining the lease and actual maintenance conduct rather than assuming that every cooling failure is automatically a landlord breach. (Cour de Cassation)

20. Contractor vs Subcontractor Liability

A large data-center project can involve:

Owner → general contractor → M&E contractor → HVAC subcontractor → equipment manufacturer.

The claimant must determine:

who contracted with whom;

whether direct warranty exists;

whether third-party rights exist;

whether the main contractor remains responsible for subcontractor performance;

whether contribution claims exist.

The Orange v Hoare Lea litigation illustrates exactly this multi-party structure. (vLex)

21. Manufacturer Liability

If the cooling failure results from a defective:

compressor;

chiller;

pump;

control board;

sensor;

valve;

the claimant may have claims against the manufacturer under:

contractual warranty;

product liability;

sale-of-goods law;

tort/delict;

statutory product-defect regimes.

However, product liability is different from contractual service liability.

22. Data Loss as Damage

Cooling failure can damage:

Physical equipment

servers;

storage arrays;

network equipment;

UPS systems.

Digital assets

corrupted data;

unavailable databases;

lost transactions.

Business operations

downtime;

SLA penalties;

customer departures;

lost revenue.

A claimant must prove that each category of loss is legally recoverable and causally connected.

23. Business Interruption

Business interruption is often much larger than the repair cost.

Example:

Cooling failure lasts 8 hours.

The claimant may seek:

emergency cooling;

equipment replacement;

lost revenue;

customer refunds;

service credits;

employee overtime;

migration expenses.

But courts may scrutinize:

foreseeability;

contractual exclusions;

remoteness;

mitigation;

proof of actual loss.

24. Mitigation of Loss

The operator has an important duty to mitigate where recognized by the applicable law.

Possible mitigation measures include:

shutting down non-critical racks;

transferring workloads;

activating redundant chillers;

emergency generators;

temporary cooling;

moving equipment;

reducing IT load.

Failure to mitigate can reduce damages.

25. Force Majeure

The defendant may argue:

“The cooling failure resulted from an extraordinary external event.”

Possible examples:

extreme weather;

utility failure;

natural disaster;

government action;

extraordinary grid event.

But force majeure generally depends on the governing national law and contract.

A mechanical failure caused by:

lack of maintenance

would ordinarily be much harder to characterize as an external force majeure event.

26. Contractual Liability Caps

Data-center contracts often contain:

liability caps;

exclusions of consequential loss;

business-interruption exclusions;

service-credit regimes;

liquidated damages;

insurance requirements.

Suppose:

Actual loss = €20 million
Contractual cap = €2 million.

The claimant may face a major dispute over the enforceability and scope of the cap.

The court must examine the applicable national law, wording and circumstances.

27. SLA and Service Credits

A data-center SLA may provide:

Availability = 99.999%.

or:

Maximum temperature deviation = specified range.

or:

Emergency response = 15 minutes.

A dispute can arise over whether:

SLA service credits

are:

the exclusive remedy; or

merely a minimum remedy while larger damages remain available.

That depends on the governing contract and national law.

28. Environmental Standards and Technical Benchmarks

Cooling-system liability can be informed by:

contractual specifications;

manufacturer specifications;

engineering standards;

data-center design standards;

industry practice;

temperature/humidity requirements;

applicable building regulations.

EU environmental guidance emphasizes proper airflow management, separation of hot and cold air, appropriate cooling capacity, maintenance and optimization of cooling systems. (EUR-Lex)

These materials can be useful evidence, but a technical best practice is not automatically a contractual obligation.

The legal question remains whether the applicable standard was incorporated into the contract or otherwise informs the applicable duty.

29. Important Distinction: Cooling Failure vs Power Failure

Cooling and electricity are interconnected but legally distinct.

Power failure

UPS/generator/grid problem.

Cooling failure

chiller/CRAC/pump/airflow/control problem.

Combined failure

power outage → cooling stops → servers overheat.

In a combined incident, several contractors may be responsible.

The German OLG Düsseldorf 23 U 27/14 case demonstrates how courts can separately analyze technical-system failures and the resulting data-center interruption loss. (Fundstellensuche)

30. Comparative Case-Law Table

The last Singapore case is intentionally identified as comparative rather than European.

31. Core Principles Emerging From European Authorities

Principle 1 — Contract determines the starting point

The court first asks what the contractor, landlord, engineer or operator actually promised.

Principle 2 — Data-center requirements can be unusually demanding

Azur Data Center is particularly useful because the court recognized the need for reinforced, continuously operational cooling. (Cour de Cassation)

Principle 3 — Maintenance obligations are enforceable

Global Switch v Wolter & Dros demonstrates that detailed maintenance obligations can become the basis of substantial litigation. (InView)

Principle 4 — Breach alone is insufficient

Causation and recoverable damage must also be established.

Principle 5 — Technical experts are extremely important

Cooling disputes often turn on engineering evidence rather than purely legal argument.

Principle 6 — Design responsibility matters

The Hurley Palmer Flatt litigation demonstrates that cooling-system instability can create claims against engineering designers. (Bailii)

Principle 7 — Subcontracting does not automatically eliminate responsibility

The French HVAC authority demonstrates the importance of the principal professional's own responsibilities concerning system requirements and design. (Cour de Cassation)

Principle 8 — External events can defeat liability

Azur Data Center illustrates that repeated failures do not necessarily establish landlord fault where the failures are fortuitous and the landlord responds diligently. (Cour de Cassation)

32. Typical Litigation Formula

A data-center cooling claim can be expressed as:

CONTRACT

↓

COOLING OBLIGATION

↓

DESIGN / INSTALLATION / MAINTENANCE FAILURE

↓

COOLING INTERRUPTION

↓

TEMPERATURE/HUMIDITY EXCURSION

↓

SERVER / DATA / SERVICE DAMAGE

↓

FINANCIAL LOSS

↓

CAUSATION

↓

DAMAGES

↓

CONTRACTUAL CAP / CONTRIBUTORY FAULT / FORCE MAJEURE

33. Exam-Oriented Revision Table

34. Conclusion

Data-center cooling system failure liability in Europe is principally governed through ordinary civil-law doctrines of contract, construction liability, professional negligence, lease obligations, product liability and damages, rather than through a single “data-center cooling” statute.

The most directly relevant authorities are Global Switch v Wolter & Dros, Azur Data Center v Vendôme Bureaux, Hurley Palmer Flatt v Barclays, Orange v Hoare Lea, the OLG Düsseldorf 23 U 27/14 decision, and the French Rouen data-center CVC litigation. Together, they show that courts focus on:

who assumed responsibility for cooling;

what technical standard was contractually required;

whether the system was properly designed, installed and maintained;

whether the failure was attributable to the defendant;

whether the failure actually caused the claimed loss;

whether the claimant mitigated its losses; and

whether contractual limitations or external events restrict recovery.

The central legal formula is:

COOLING OBLIGATION → DESIGN/MAINTENANCE FAILURE → COOLING INTERRUPTION → THERMAL DAMAGE → CAUSATION → FINANCIAL LOSS → LIABILITY → COMPENSATION

For examination purposes, the most important case is Global Switch v Wolter & Dros because it directly concerns a data-center climate-control maintenance contract and demonstrates the crucial distinction between alleged technical failure and legally proven breach/causation. (InView)