Competition Law and Data Centre Hyperscaler Dominance

1. Introduction

Data-centre hyperscalers—principally Amazon Web Services (AWS), Microsoft Azure and Google Cloud—occupy a structurally important position in modern digital markets. They do not merely provide server space. They increasingly control an integrated infrastructure stack consisting of:

data centres and physical computing capacity;

cloud infrastructure services;

storage and databases;

networking and content-delivery infrastructure;

AI computing capacity and GPU clusters;

operating systems and enterprise software;

cybersecurity services;

developer tools and APIs;

AI foundation models and applications; and

marketplaces and other digital ecosystems.

This creates a competition-law problem that is more sophisticated than simply asking whether a cloud provider has a large market share.

The central question is:

When does ownership and control of essential computing infrastructure allow a hyperscaler to convert infrastructure strength into durable market power in adjacent cloud, AI, software, data and digital-service markets?

The issue is becoming particularly significant because European competition authorities have identified lock-in, switching costs, interoperability barriers, tying/bundling, restricted data access and contractual conditions as potential sources of reduced cloud-market contestability. In June 2026, the European Commission preliminarily concluded that AWS and Microsoft Azure should be designated as gatekeepers under the DMA despite not meeting the quantitative thresholds, because of their important gateway role and entrenched positions. (Digital Markets Act (DMA))

The OECD has similarly observed that cloud computing is highly concentrated among a small number of hyperscalers, particularly AWS, Azure and Google Cloud. (OECD)

2. What Is a Data-Centre Hyperscaler?

A hyperscaler is a technology company capable of operating computing infrastructure at enormous geographic and economic scale.

The relevant distinction is:

Traditional data centre

→ provides physical space, power, cooling and connectivity.

Cloud provider

→ provides virtualised computing, storage, databases and networking.

Hyperscaler

→ combines massive physical infrastructure with cloud platforms, software ecosystems, AI infrastructure, data services and complementary digital products.

This vertical integration can generate significant competition advantages.

For example:

Data centres → compute → cloud platform → enterprise software → AI services → data → applications

The same corporate group may control several layers simultaneously.

That creates the possibility of vertical leveraging.

3. Why Hyperscaler Dominance Is Different from Ordinary Market Dominance

Traditional competition analysis often concentrates on:

market definition;

market share;

barriers to entry;

consumer harm; and

exclusionary conduct.

Hyperscaler markets require additional attention to:

A. Switching costs

A company may have thousands of applications, databases and workloads operating on one cloud.

Moving them may require:

rewriting applications;

transferring enormous datasets;

changing APIs;

retraining personnel;

redesigning cybersecurity architecture;

migrating databases; and

accepting operational downtime.

Therefore, a nominally contestable customer may be economically locked in.

B. Network effects

The more customers a cloud provider has, the more attractive its ecosystem becomes.

More customers encourage:

more developers;

more third-party applications;

more marketplace offerings;

more technical expertise;

more AI integrations.

This can reinforce incumbent advantages.

C. Economies of scale

Hyperscalers can spread enormous infrastructure costs across millions of customers.

A new entrant cannot easily replicate:

global data-centre networks;

electricity procurement;

fibre connectivity;

GPU capacity;

cybersecurity infrastructure;

specialised engineering personnel.

Consequently, scale can become an entry barrier.

D. Data advantages

Hyperscalers may obtain substantial information concerning:

workloads;

resource consumption;

application performance;

enterprise demand;

cybersecurity events;

data-storage patterns;

AI utilisation.

The competition concern becomes particularly serious where infrastructure information can be combined with information from adjacent businesses.

4. Relevant Competition-Law Framework

A. Article 102 TFEU

In the European Union, Article 102 TFEU prohibits abuse of a dominant position.

Potential hyperscaler abuses include:

exclusionary pricing;

discriminatory pricing;

tying;

bundling;

refusal of interoperability;

discriminatory access;

self-preferencing;

exploitative contractual terms;

foreclosure of competing cloud providers;

leveraging dominance into AI markets.

B. Sections 2 and 7 of the Sherman Act

In the United States:

Section 2

Addresses monopolisation and attempted monopolisation.

Section 7 of the Clayton Act

Can address mergers and acquisitions that substantially lessen competition.

Hyperscaler acquisitions therefore deserve particular scrutiny where they eliminate:

potential competitors;

AI startups;

cloud-management companies;

specialised data infrastructure firms; or

technologies capable of reducing switching costs.

C. UK Competition Act 1998

The Chapter II prohibition addresses abuse of dominance.

The UK has also conducted extensive analysis of public-cloud competition.

The CMA's cloud-services market investigation concluded in 2025 with an adverse effect on competition and recommended consideration of strategic market-status investigations concerning Microsoft and AWS. (GOV.UK)

D. Indian Competition Act, 2002

For India, the principal provisions are:

Section 4

Prohibits abuse of dominant position.

Relevant conduct may include:

unfair or discriminatory conditions;

unfair or discriminatory pricing;

limiting technical development;

denial of market access;

leveraging dominance from one relevant market into another.

Sections 5 and 6

Govern combinations and therefore become relevant to hyperscaler acquisitions.

Section 19

Provides the framework for determining relevant market, dominance and abuse.

5. Relevant Market Definition

Market definition becomes extremely difficult in hyperscaler cases.

A regulator could potentially examine:

Market 1

Infrastructure-as-a-Service (IaaS)

Market 2

Platform-as-a-Service (PaaS)

Market 3

Cloud storage

Market 4

Cloud databases

Market 5

AI computing infrastructure

Market 6

GPU cloud services

Market 7

Enterprise cloud services

Market 8

Data-centre capacity

Market 9

Interconnection services

Market 10

Cloud cybersecurity

The relevant market might also be geographically defined as:

national;

regional;

EU-wide;

global.

The correct market depends upon substitutability.

6. Case Law

Case 1 — Microsoft Corp. v Commission (Microsoft I), 2007

This is one of the most important precedents for hyperscaler dominance.

The European Commission found that Microsoft had abused its dominant position through conduct involving interoperability information and tying.

The case demonstrates that a dominant technology company cannot necessarily use control over one technological layer to disadvantage competitors operating at another layer.

Relevance to hyperscalers

A hyperscaler could theoretically possess:

dominance in cloud infrastructure

and use that position to favour:

its database products;

operating systems;

AI models;

cybersecurity tools;

analytics;

developer platforms.

The Microsoft precedent therefore supports scrutiny of technological leveraging.

7. Case 2 — Microsoft v Commission, Case T-201/04

The General Court upheld the essential substance of the Commission's finding that Microsoft had abused its dominant position.

The important principle is that interoperability can become a competition concern where access to technical information is indispensable for competitors to operate effectively.

Hyperscaler application

Consider a cloud environment where a hyperscaler controls:

proprietary APIs;

migration interfaces;

technical protocols;

authentication infrastructure.

If those interfaces are deliberately designed to make migration difficult, the conduct could raise abuse-of-dominance concerns.

The case therefore provides a conceptual foundation for analysing cloud interoperability.

8. Case 3 — Bronner v Mediaprint, C-7/97

The Court of Justice adopted a restrictive approach to refusal-to-deal claims.

Generally, a dominant undertaking is not automatically required to supply competitors.

However, exceptional circumstances can justify intervention where access to an infrastructure is indispensable.

Importance for data centres

The essential-facilities doctrine becomes relevant when a hyperscaler controls infrastructure that competitors cannot realistically reproduce.

The legal question would be:

Is access to the relevant cloud or infrastructure genuinely indispensable?

A regulator would consider:

technical alternatives;

economic feasibility;

duplication costs;

availability of substitute infrastructure;

capacity constraints.

Importantly, mere superiority is not the same as indispensability.

9. Case 4 — Magill, Joined Cases C-241/91 P and C-242/91 P

In Magill, the Court established exceptional circumstances under which refusal to license intellectual property could constitute abuse.

The case is relevant because hyperscaler dominance increasingly involves proprietary:

APIs;

software;

datasets;

technical interfaces;

AI models;

cloud-management technologies.

However, the threshold for compulsory access remains high.

Hyperscaler significance

A regulator cannot simply say:

"The infrastructure is important, therefore access must be granted."

It must establish the legal conditions for intervention.

This protects incentives for infrastructure investment while preventing strategic exclusion.

10. Case 5 — IMS Health v Commission, C-418/01

IMS Health refined the exceptional circumstances approach concerning refusal to license intellectual property.

The Court emphasised factors such as:

indispensability;

elimination of effective competition;

prevention of a new product or service;

absence of objective justification.

Hyperscaler application

Imagine a hyperscaler controls a proprietary technical interface essential for migration between cloud systems.

If competitors cannot realistically operate without it, and the refusal eliminates effective competition, IMS Health becomes highly relevant.

But again, the doctrine is exceptional.

11. Case 6 — Google Shopping, Case AT.39740

The European Commission found that Google had abused its dominant position in general search by favouring its own comparison-shopping service.

The broader principle is self-preferencing and leveraging.

Hyperscaler application

A hyperscaler may simultaneously operate:

cloud infrastructure;

cloud marketplace;

AI model platform;

enterprise software;

cybersecurity products.

Suppose its cloud marketplace systematically ranks its own products above rival products.

The competition issue would be:

Is the infrastructure provider using control over the gateway to disadvantage competing downstream products?

This is especially important where the hyperscaler determines:

ranking;

access;

visibility;

pricing;

technical compatibility.

12. Case 7 — Google Android, Case AT.40099

The European Commission's Android decision concerned Google's use of contractual arrangements involving Android and related services.

The case illustrates how dominance in one technological ecosystem can be leveraged through:

tying;

contractual restrictions;

default arrangements;

ecosystem incentives.

Hyperscaler application

A cloud provider might condition access to a valuable service upon adoption of another service.

For example:

"Discounted cloud infrastructure is available only if the customer also purchases our database, AI or security products."

The legal analysis would examine whether the arrangement produces foreclosure beyond legitimate efficiencies.

13. Case 8 — Qualcomm, Case AT.40220

The Qualcomm exclusivity case is useful for analysing payments and commercial incentives that may foreclose rivals.

The broader lesson is that competition law can examine contractual arrangements that make it economically difficult for customers to use competing suppliers.

Hyperscaler application

Hyperscalers increasingly use:

volume discounts;

committed-spend agreements;

credits;

rebates;

bundled services;

preferential financing.

These arrangements are not inherently unlawful.

But they become problematic where they effectively make switching to another cloud provider commercially irrational.

14. Case 9 — United Brands v Commission, 27/76

United Brands remains fundamental to the law of dominance.

The Court recognised that dominance concerns the ability of an undertaking to behave to an appreciable extent independently of competitors, customers and consumers.

Hyperscaler significance

A hyperscaler could potentially possess market power even where:

several competitors remain;

prices are not excessive;

customers technically can switch.

The real question is whether competitive constraints are sufficiently strong.

A cloud customer saying:

"We could technically leave AWS."

does not necessarily demonstrate effective competitive constraint if migration would cost millions and take years.

15. Case 10 — AKZO v Commission, C-62/86

AKZO is particularly relevant to predatory pricing and exclusionary pricing.

Hyperscalers have enormous financial resources and diversified revenue streams.

This creates the theoretical possibility of:

cross-subsidisation of cloud services from another profitable business.

For example, a conglomerate might temporarily price cloud infrastructure extremely aggressively to eliminate smaller rivals.

Competition law would examine:

cost benchmarks;

pricing duration;

recoupment;

exclusionary intent;

market structure;

effects on competitors.

16. The Central Problem: Vertical Leveraging

The most important competition-law risk may not be simple horizontal dominance.

It is vertical leveraging.

Consider:

Data centres

↓

Cloud infrastructure

↓

Operating systems

↓

Enterprise software

↓

AI infrastructure

↓

AI models

↓

Applications

A company controlling several layers can potentially discriminate against competitors at downstream levels.

This creates a vertical foreclosure problem.

17. Cloud Lock-In

Cloud lock-in is perhaps the most important structural concern.

Suppose a customer initially selects Cloud A.

Over time it develops:

proprietary databases;

machine-learning pipelines;

cloud-specific APIs;

identity-management systems;

monitoring tools;

security systems.

Migration becomes progressively more expensive.

The customer's theoretical ability to switch does not necessarily equal effective substitutability.

This distinction is critical in dominance analysis.

18. Egress Fees

One of the most controversial issues concerns data egress fees.

A customer may be allowed to download its data but have to pay substantial charges to move it elsewhere.

Competition analysis should ask:

Are the fees cost-based?

Are they necessary to recover genuine network costs?

Are they excessive?

Do they disproportionately penalise switching?

Are they combined with other contractual restrictions?

Do they discriminate between internal and external transfers?

A reasonable network charge is not automatically anticompetitive.

But excessive switching charges may reinforce lock-in.

19. Interoperability Restrictions

Interoperability is another major issue.

Hyperscalers may control proprietary:

APIs;

databases;

identity systems;

orchestration tools;

networking architecture.

Where interoperability is unnecessarily restricted, competitors may face substantial barriers.

The European Commission's ongoing cloud work specifically identifies interoperability, access to business-user data, tying/bundling and imbalanced contractual terms as areas of concern. (Digital Markets Act (DMA))

20. Tying and Bundling

Hyperscalers operate huge product portfolios.

For example:

Cloud infrastructure + AI model + cybersecurity + database + productivity software

Bundling can create efficiencies.

But competition concerns arise where:

dominance in Product A is used to force adoption of Product B.

The relevant questions include:

Are the products separate?

Does the dominant product possess market power?

Is the second product effectively compulsory?

Does the arrangement foreclose rivals?

Are there objective technical or economic justifications?

Microsoft's earlier cases provide particularly useful doctrinal guidance.

21. Self-Preferencing

A hyperscaler may operate both:

the infrastructure platform

and

competing downstream services.

For example:

Cloud provider → owns marketplace

and

Cloud provider → sells its own cybersecurity software.

If the platform:

ranks its own product first;

provides it superior API access;

gives it cheaper infrastructure;

restricts rival visibility;

competition concerns arise.

Google Shopping provides an important analytical reference point.

22. AI and GPU Infrastructure

The hyperscaler problem is becoming more serious because of AI.

AI requires:

enormous GPU clusters;

high-speed networking;

specialised cooling;

large electricity supplies;

massive storage;

sophisticated data pipelines.

Consequently:

AI competitiveness → compute availability → data-centre capacity

The result is that data-centre capacity itself can become strategically important.

Recent industry developments show extraordinary investment in AI data-centre capacity, while cloud providers continue to occupy central positions in supplying compute resources. (Financial Times)

23. GPU Scarcity and Competition

Suppose a hyperscaler controls a disproportionately large share of available AI compute.

It could potentially:

reserve capacity for its own AI subsidiaries;

give preferential access to affiliated AI models;

impose restrictive allocation conditions on independent AI companies;

use long-term capacity agreements to foreclose rivals.

This produces a new form of infrastructure foreclosure.

Competition authorities may therefore need to examine not merely cloud market shares but also:

who controls scarce computational capacity.

24. Data Advantage and Competitive Feedback Loops

Hyperscaler dominance can create a feedback loop:

More customers

↓

More infrastructure utilisation

↓

More revenue

↓

More investment

↓

More data centres and GPUs

↓

Better prices/performance

↓

More customers

This is a legitimate economy of scale in many circumstances.

But competition law becomes concerned where the feedback loop is strengthened artificially through exclusionary conduct.

25. Data Centre Capacity as an Entry Barrier

Building hyperscale infrastructure requires:

enormous capital expenditure;

land;

electricity;

water/cooling;

fibre;

permits;

construction expertise;

GPUs;

network equipment;

security.

Therefore, entry barriers are substantial.

However, high capital expenditure alone does not prove dominance.

Competition authorities must distinguish:

Natural scale economies

from

Artificially created barriers.

The latter could arise through:

exclusive capacity agreements;

foreclosure of electricity supply;

exclusive network arrangements;

discriminatory access;

restrictive cloud contracts.

26. Competition Between Hyperscalers and Independent Data Centres

Competition is not limited to AWS vs Azure vs Google Cloud.

There may also be competition between:

hyperscalers;

colocation providers;

regional cloud providers;

sovereign cloud providers;

specialised GPU providers;

enterprise-owned data centres;

edge-computing providers.

The relevant question is therefore whether these alternatives impose real competitive constraints.

27. The UK Approach

The UK provides an especially important example.

The CMA's public-cloud market investigation concluded that competition concerns were sufficiently serious to justify remedies and recommended consideration of strategic market-status investigations for Microsoft and AWS. (GOV.UK)

This demonstrates a shift from purely retrospective antitrust enforcement toward proactive digital-market regulation.

28. EU DMA Development

The EU's approach is particularly significant.

In November 2025, the Commission opened investigations into whether AWS and Azure should be designated as gatekeepers and separately investigated whether the DMA adequately addresses cloud-specific competition problems. (Digital Markets Act (DMA))

In June 2026, the Commission announced its preliminary position that AWS and Azure should qualify as gatekeepers despite not satisfying the normal quantitative thresholds. (Digital Markets Act (DMA))

This illustrates an important doctrinal development:

Digital infrastructure can possess strategic gateway characteristics even without satisfying traditional platform metrics.

29. Data Centre Hyperscaler Dominance and Indian Competition Law

Under India's Competition Act, 2002, the central provision would be Section 4.

The CCI could investigate whether a dominant cloud/data-centre undertaking:

(a) imposes unfair conditions

For example:

excessive minimum commitments;

unreasonable termination restrictions.

(b) imposes discriminatory conditions

For example:

one price for affiliated entities;

another price for independent customers.

(c) denies market access

For example:

restricting competitors' interoperability.

(d) limits technical development

For example:

intentionally preventing migration technologies.

(e) leverages dominance

For example:

dominance in cloud infrastructure → foreclosure of competing AI services.

This last category could become particularly important in India's rapidly expanding AI ecosystem.

30. Essential Facilities and Hyperscalers

The essential-facilities doctrine should be applied cautiously.

Not every large data centre is an essential facility.

A facility becomes more legally significant where:

access is indispensable;

duplication is practically or economically impossible;

refusal eliminates effective competition;

access can be provided without destroying legitimate investment incentives.

The Bronner–Magill–IMS Health line of authorities therefore provides an important framework.

31. Abuse Through Contractual Design

Competition law should examine not merely prices but the architecture of contracts.

Potentially problematic provisions include:

excessive minimum-spend commitments;

automatic renewal;

long lock-in periods;

restrictive termination clauses;

high migration costs;

data portability restrictions;

preferential pricing for affiliated services;

anti-switching incentives;

restrictions on multi-cloud arrangements.

A contract that appears commercially voluntary may still produce exclusionary effects.

32. Multi-Cloud Competition

Multi-cloud arrangements are an important competitive safeguard.

A customer may simultaneously use:

AWS + Azure + Google Cloud + private infrastructure.

This can reduce dependence on any single hyperscaler.

However, switching between clouds may remain expensive because of:

data transfer;

API incompatibility;

application redesign;

cybersecurity architecture;

employee expertise.

Therefore, regulators may examine whether multi-cloud is economically practical, rather than merely technically possible.

33. Data Portability

Data portability can reduce switching costs.

Competition policy may therefore favour:

standardised formats;

APIs;

interoperability;

reasonable transfer mechanisms;

transparent migration procedures.

But portability must be balanced against:

cybersecurity;

privacy;

intellectual property;

trade secrets.

34. Merger Control

Hyperscaler dominance also creates major merger-control questions.

Authorities should examine acquisitions involving:

AI startups;

GPU/cloud providers;

data-centre operators;

cloud-management software;

cybersecurity firms;

database providers;

AI infrastructure companies.

The traditional question:

"Will the merger reduce existing competition?"

may be insufficient.

Authorities should also ask:

"Does the acquisition eliminate a potential future competitor?"

35. Killer Acquisitions

Suppose a hyperscaler acquires a small AI infrastructure company.

Today:

Startup = tiny market share

Tomorrow:

Startup = possible challenger

If the startup has technology capable of reducing cloud dependence, the acquisition may have significant competitive implications despite its current revenue being small.

This is the potential-competition theory.

36. Efficiency Defences

Hyperscalers can legitimately argue that vertical integration produces efficiencies.

For example:

lower infrastructure costs;

improved cybersecurity;

better latency;

improved reliability;

faster AI deployment;

lower prices;

better resource utilisation.

Competition law should therefore not treat size or integration itself as unlawful.

The question is:

Does the conduct produce legitimate efficiencies that could not reasonably be achieved through less restrictive means?

37. Consumer Welfare and Business Users

Unlike conventional consumer markets, cloud competition primarily concerns business customers.

The affected users may include:

banks;

hospitals;

governments;

manufacturers;

universities;

startups;

AI developers.

Therefore, harm may manifest as:

higher cloud expenditure;

reduced innovation;

reduced technological choice;

higher migration costs;

slower AI development;

reduced resilience.

38. Competition and Digital Sovereignty

There is also a strategic dimension.

If a small number of foreign hyperscalers control a country's:

cloud infrastructure;

AI computing;

public-sector workloads;

financial-sector infrastructure;

dependency may become economically and strategically significant.

However, digital sovereignty policy is not automatically competition law.

A government must distinguish:

national-security policy

from

competition-policy intervention.

39. Possible Competition Remedies

Structural remedies

In extreme circumstances:

divestiture;

separation of infrastructure and downstream businesses;

prohibition of certain acquisitions.

These should normally be considered only where behavioural remedies cannot work.

Behavioural remedies

More practical remedies may include:

1. Interoperability

Require reasonable technical interoperability.

2. Data portability

Allow customers to transfer data efficiently.

3. Limits on switching charges

Prevent unnecessary migration penalties.

4. Non-discrimination

Require equivalent treatment of competing services.

5. Prohibition of tying

Prevent forced bundling.

6. Marketplace neutrality

Require fair treatment of competing products.

7. Transparency

Require disclosure of relevant contractual restrictions.

40. Competition-Law Risk Matrix

41. Six Core Legal Principles Emerging from the Case Law

The case law collectively establishes six particularly important principles.

Principle 1 — Dominance is not unlawful by itself

United Brands makes clear that competition law targets abuse, not mere economic success.

Principle 2 — Infrastructure control can become legally significant

Bronner demonstrates that exceptional circumstances can justify access obligations.

Principle 3 — Interoperability matters

Microsoft shows that technological control can have competition consequences where interoperability is strategically restricted.

Principle 4 — Refusal to license can exceptionally be abusive

Magill and IMS Health establish a demanding framework.

Principle 5 — Dominance can be leveraged into neighbouring markets

Google Shopping and Android illustrate the danger of using control over one ecosystem to disadvantage competitors elsewhere.

Principle 6 — Pricing and contractual incentives can foreclose competitors

AKZO and Qualcomm demonstrate the importance of analysing exclusionary commercial strategies.

42. The Central Legal Test for Hyperscaler Dominance

A sophisticated competition-law investigation should therefore proceed through the following sequence:

Step 1 — Define the relevant market

↓

Step 2 — Determine hyperscaler's market power

↓

Step 3 — Identify infrastructure bottlenecks

↓

Step 4 — Examine switching costs

↓

Step 5 — Analyse interoperability

↓

Step 6 — Examine tying/bundling

↓

Step 7 — Examine self-preferencing

↓

Step 8 — Analyse pricing and rebates

↓

Step 9 — Examine vertical leveraging

↓

Step 10 — Assess actual or potential foreclosure

↓

Step 11 — Examine efficiencies and objective justification

↓

Step 12 — Select proportionate remedies

43. Overall Assessment

The competition-law problem presented by hyperscaler dominance is not simply that AWS, Azure or Google Cloud are large.

The deeper concern is the possibility of a self-reinforcing infrastructure ecosystem:

Data centres → compute capacity → cloud infrastructure → data → software → AI → customers → more scale

Once this ecosystem becomes sufficiently entrenched, the hyperscaler may possess the ability to make competitors' entry, expansion or switching increasingly difficult.

The most important risks therefore concern:

cloud lock-in;

high switching costs;

data portability barriers;

interoperability restrictions;

egress charges;

tying and bundling;

self-preferencing;

exclusive contracts;

preferential allocation of scarce AI/GPU capacity;

vertical integration;

leveraging into AI and software markets; and

acquisition of potential competitors.

The contemporary regulatory direction is significant. The EU is now expressly examining whether cloud infrastructure providers can function as important gateways, while the UK has identified structural competition problems in public cloud. (Digital Markets Act (DMA))

The emerging legal principle can therefore be stated as follows:

A hyperscaler's ownership of massive data-centre infrastructure is not itself an antitrust violation. The competition-law concern arises when infrastructure scale, technological integration, contractual restrictions, data advantages or ecosystem control are used to make adjacent markets less contestable or to exclude actual or potential competitors.

Principal authorities

United Brands v Commission, Case 27/76

AKZO v Commission, Case C-62/86

Magill, Joined Cases C-241/91 P and C-242/91 P

Bronner v Mediaprint, Case C-7/97

IMS Health v Commission, Case C-418/01

Microsoft v Commission, Case T-201/04 / Commission decision

Google Android, Case AT.40099

Google Shopping, Case AT.39740

Qualcomm, Case AT.40220