Competition Law and Data Infrastructure Monopolisation Risks

1. Introduction

The modern digital economy depends not merely on software and algorithms but on data infrastructure: data centres, cloud-computing capacity, storage systems, APIs, identity infrastructure, search indexes, datasets, data-processing platforms, AI-training infrastructure, and the networks through which data are collected and processed.

This creates a competition-law problem that is somewhat different from traditional monopoly analysis. A firm may acquire market power not simply because it sells a popular product, but because it controls an infrastructure layer that competitors need in order to operate.

For example:

Data → storage → cloud infrastructure → computing/GPU capacity → APIs → applications → consumers

If one undertaking controls a critical layer, it may potentially use that position to disadvantage firms operating at another layer.

The risk is particularly significant where infrastructure is characterised by:

very high fixed costs;

economies of scale;

network effects;

switching costs;

large accumulated datasets;

interoperability barriers;

technical standards;

proprietary APIs;

vertical integration;

scarce computing capacity; and

substantial barriers to duplicating infrastructure.

The European Commission's recent implementation of the Digital Markets Act illustrates the direction of regulatory thinking: gatekeepers can be subject to data-access obligations precisely because control over strategically important data can reinforce market power. (Digital Markets Act (DMA))

2. What Is "Data Infrastructure Monopolisation"?

Data infrastructure monopolisation occurs where an undertaking obtains or maintains substantial market power through control over infrastructure that is indispensable, difficult to replicate, or strategically important for collecting, processing, storing, transferring or accessing data.

It can occur at several levels.

A. Physical infrastructure

Examples include:

hyperscale data centres;

data-centre campuses;

fibre networks;

submarine cables;

internet exchange facilities;

specialised GPU clusters;

cloud infrastructure;

server capacity.

B. Digital infrastructure

Examples include:

cloud platforms;

operating systems;

application programming interfaces;

authentication systems;

app stores;

search indexes;

digital identity systems;

interoperability interfaces.

C. Data infrastructure

Examples include:

unique datasets;

real-time consumer data;

transaction databases;

mapping databases;

search-query databases;

behavioural datasets;

AI-training datasets.

D. Hybrid infrastructure

Increasingly, the most important monopolisation risks arise from combinations of physical and digital infrastructure:

Data centre + cloud + proprietary data + AI models + APIs + distribution platform.

This creates the possibility of vertical foreclosure, where an infrastructure owner uses control over an upstream layer to exclude or weaken competitors downstream.

3. Why Data Infrastructure Creates Competition Problems

3.1 High fixed costs

Data infrastructure frequently requires enormous capital expenditure.

A new entrant may need:

land;

electricity;

cooling systems;

servers;

networking;

GPUs;

cybersecurity;

backup systems;

fibre connectivity;

software;

technical personnel.

Consequently, an incumbent with large existing infrastructure may have a substantial cost advantage.

The competitive concern arises where the infrastructure cannot reasonably be replicated by smaller competitors.

3.2 Economies of scale

Data infrastructure frequently exhibits substantial economies of scale.

The average cost of processing or storing data may fall as capacity increases.

A dominant firm therefore has an incentive to continue expanding infrastructure because:

greater scale → lower unit costs → lower prices or higher margins → more customers → more data → further scale.

This can produce a self-reinforcing competitive advantage.

4. The Data-Feedback Loop

One of the most important concepts is the data feedback loop.

A simplified model is:

More users
↓
More data
↓
Better algorithms/models
↓
Better products
↓
More users
↓
Still more data

This creates what competition economists sometimes describe as a data-driven network effect.

The problem is not necessarily that possessing data is unlawful. Competition law generally protects successful accumulation of resources through competition on the merits.

The concern arises where an undertaking uses accumulated data or infrastructure to:

exclude competitors;

deny necessary access;

discriminate against rivals;

impose unfair interoperability conditions;

engage in self-preferencing;

tie products;

foreclose downstream markets; or

prevent effective entry.

5. Essential Facilities Doctrine

The most directly relevant competition-law concept is the essential facilities doctrine.

The basic idea is that a dominant undertaking controlling an indispensable facility may, in exceptional circumstances, have a competition-law obligation to provide access to competitors.

However, courts have deliberately made this doctrine difficult to satisfy because compulsory access can reduce incentives to invest.

The classic European test comes from Bronner.

Generally, the following considerations become important:

Is the facility genuinely indispensable?

Can competitors realistically duplicate it?

Would refusal eliminate effective competition?

Is access technically and economically feasible?

Is there an objective justification for the refusal?

The EU courts have repeatedly stressed the importance of indispensability and the risk of eliminating competition when applying essential-facilities reasoning. (EUR-Lex)

For data infrastructure, the central question becomes:

Is the relevant dataset, computing infrastructure, API, cloud service or physical infrastructure genuinely indispensable, or can competitors develop alternatives?

6. Major Case Law

Case 1 — United States v. Terminal Railroad Association, 224 U.S. 383 (1912)

Principle

This is one of the foundational American cases concerning essential facilities.

A group of railroad companies controlled essential terminal infrastructure in St. Louis. Competitors could not effectively compete without access to the terminal facilities.

The Supreme Court regarded the arrangement as presenting a serious competition problem because control over infrastructure could be used to exclude rivals.

Relevance to data infrastructure

The analogy is powerful.

A data-centre operator could potentially occupy a comparable bottleneck position if:

competitors cannot reasonably obtain equivalent capacity;

the infrastructure is necessary to reach customers;

access is technically feasible; and

exclusion substantially restricts competition.

The modern equivalent of the railway terminal could potentially be a critical digital or physical data infrastructure node.

Case 2 — MCI Communications Corp. v. AT&T, 708 F.2d 1081 (7th Cir. 1983)

Importance

MCI v. AT&T is one of the most frequently cited American essential-facilities cases.

The Seventh Circuit articulated factors concerning:

control of the facility by a monopolist;

inability of competitors reasonably or practically to duplicate the facility;

denial of use of the facility; and

feasibility of providing access.

Data-infrastructure application

Suppose a dominant cloud provider controls a specialised infrastructure layer that competitors cannot economically reproduce.

The analysis could ask:

However, merely being expensive does not necessarily make infrastructure essential.

That distinction is critical.

Case 3 — Commercial Solvents v Commission, Joined Cases 6/73 & 7/73

The European Court of Justice recognised that a dominant undertaking could abuse its position where it controlled an input necessary for competitors operating in a downstream market and attempted to exclude those competitors.

Competition principle

A dominant undertaking cannot necessarily use control over an upstream input to eliminate competition downstream.

Data infrastructure relevance

Consider a hypothetical company controlling:

a unique data-processing infrastructure;

an essential API;

a dominant identity system; or

a critical cloud-computing layer.

If it simultaneously competes downstream with companies dependent upon that infrastructure, the incentive to discriminate becomes particularly significant.

This creates a classic vertical foreclosure problem:

Upstream infrastructure monopoly
↓
Denial/discrimination
↓
Downstream competitor disadvantage
↓
Reduced competition

Case 4 — RTE & ITP v Commission (Magill), Joined Cases C-241/91 P & C-242/91 P

Facts

Television broadcasters controlled information concerning programme schedules. They refused to license comprehensive programme information to publishers seeking to produce competing television guides.

The European courts recognised that, under exceptional circumstances, refusal to license information could constitute abuse of dominance.

Importance

The case is extremely relevant to data monopolisation.

It demonstrates that information can become competition-law relevant where:

the information is indispensable for a downstream product;

refusal prevents the emergence of a new product;

refusal lacks sufficient justification; and

the dominant undertaking effectively controls the downstream market.

Data infrastructure analogy

Imagine a dominant platform possessing a unique real-time dataset that competitors require to create a competing service.

The question would not simply be:

"Does the company own the data?"

Instead:

"Does its control over the data allow it to eliminate effective competition in a downstream market?"

Case 5 — Oscar Bronner GmbH & Co. KG v Mediaprint, Case C-7/97

This is arguably the most important European case for understanding the limits of mandatory access.

Facts

Bronner operated a newspaper and sought access to Mediaprint's newspaper distribution system.

Mediaprint had a widespread distribution network.

Bronner argued that access was necessary to compete effectively.

Decision

The Court applied a demanding standard.

A facility must be genuinely indispensable. The fact that access would make competition easier or cheaper is insufficient.

The Court was concerned that forcing companies to share infrastructure could undermine their incentives to invest.

Significance for data infrastructure

This principle is crucial.

A cloud service may be:

commercially important;

cheaper than building an alternative;

technically superior;

widely used;

without necessarily being an "essential facility."

Competition law should not automatically convert commercial dependence into a compulsory access right.

The EU jurisprudence continues to emphasise this distinction. (EUR-Lex)

Case 6 — IMS Health GmbH & Co. OHG v NDC Health GmbH, Case C-418/01

Facts

IMS Health controlled a pharmaceutical sales-data system based on a particular regional structure.

A competitor wanted access to that structure to compete in the market for pharmaceutical data services.

Importance

The Court applied the exceptional circumstances framework associated with refusal to license intellectual property.

Three important considerations were highlighted:

refusal prevents the emergence of a new product;

refusal is unjustified; and

refusal reserves a downstream market to the dominant undertaking.

Data infrastructure relevance

This is particularly significant for:

proprietary databases;

data formats;

data taxonomies;

industry-standard datasets;

API structures;

proprietary data architecture.

A dominant firm cannot necessarily use intellectual-property rights or control over data architecture as an absolute shield against competition-law scrutiny.

But, again, the conditions are exceptional.

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

The Microsoft case is one of the most important modern precedents for technology infrastructure.

The European Commission found that Microsoft had abused its dominant position by restricting interoperability information needed by rival work-group server operating systems.

The General Court largely upheld the Commission's findings.

Competition principle

A dominant technology company may face obligations concerning interoperability where refusal prevents competitors from competing effectively.

Data infrastructure significance

The Microsoft reasoning is highly relevant to:

APIs;

interoperability protocols;

cloud interfaces;

data portability;

operating systems;

technical standards.

The modern competition question may therefore become:

Can an infrastructure provider technically interoperate with competitors but deliberately design its architecture to make interoperability commercially or technically ineffective?

That is potentially more sophisticated than a straightforward refusal to supply.

Case 8 — Google Shopping, Case T-612/17

The Google Shopping litigation is particularly important because it demonstrates that digital-market abuse does not always require a classic refusal to deal.

The EU courts upheld the finding that Google had favoured its own comparison-shopping service in general search results while disadvantaging competing comparison-shopping services. (EUR-Lex)

The case is important for data infrastructure because the dominant search engine simultaneously functioned as:

infrastructure;

gateway;

data collector;

ranking mechanism;

distribution channel.

The competitive advantage derived from control over the search infrastructure could therefore be leveraged into an adjacent market.

The General Court specifically considered the relationship between the case and the stricter Bronner essential-facilities test. (EUR-Lex)

Key lesson

A dominant infrastructure provider may potentially violate competition law not only by saying:

"You cannot access my infrastructure."

It may also do so by saying:

"You can access my infrastructure, but I will systematically design it to favour my own downstream service."

That is the self-preferencing problem.

Case 9 — Shamsher Kataria v Honda Siel Cars India Ltd., CCI Case No. 03/2011

This is particularly relevant from an Indian competition-law perspective.

The Competition Commission of India considered competition issues involving access to information, spare parts and repair-related markets in the automobile sector.

The case is frequently discussed in connection with the Indian version of essential-facilities reasoning.

Indian statutory relevance

Section 4 of the Competition Act, 2002 prohibits abuse of dominant position.

Section 4(2)(c) specifically addresses:

denial of market access.

This provision can become important where a dominant infrastructure provider prevents competitors from accessing resources necessary to compete.

Data infrastructure application

Potential examples could include:

refusal to provide interoperable APIs;

denial of access to critical data;

discriminatory cloud interoperability;

discriminatory technical standards;

refusal to provide necessary infrastructure connectivity.

However, Indian law does not automatically classify all important data as an essential facility.

Case 10 — Matrimony.com v Google, CCI Case Nos. 07 & 30 of 2012

This is another important Indian digital-market precedent.

The CCI examined Google's conduct in relation to search services and preferential treatment.

The broader significance lies in recognising that digital platforms can occupy gateway positions and that discriminatory treatment can affect market access.

The case is particularly relevant to infrastructure monopolisation because search infrastructure is not merely a consumer-facing product.

It can also function as:

data infrastructure + discovery infrastructure + advertising infrastructure + distribution infrastructure.

That combination can give the platform considerable leverage over adjacent markets.

7. Data as an Essential Facility

One of the most controversial questions is:

Can data itself constitute an essential facility?

The answer should generally be:

Potentially, but only in exceptional circumstances.

A dataset becomes more competition-law significant where it possesses several characteristics.

1. Uniqueness

Competitors cannot obtain equivalent data elsewhere.

2. Scale

The dataset is sufficiently large to create a substantial competitive advantage.

3. Real-time character

Historical data may be replicable, whereas real-time information may be considerably more difficult to reproduce.

4. Network effects

The value of the data increases with the number of users.

5. High replication costs

Competitors cannot economically reproduce the dataset.

6. Downstream indispensability

The dataset is genuinely necessary to compete in a downstream market.

7. Absence of reasonable alternatives

Competitors cannot obtain comparable information through another legitimate source.

8. But "Data Is Valuable" Is Not Enough

This distinction is essential.

A company may possess billions of data points without possessing an antitrust-relevant essential facility.

Competition authorities should distinguish:

Valuable data

from

Indispensable data.

For example:

A proprietary customer dataset that gives a company a competitive advantage

is not necessarily equivalent to:

A dataset without which no competitor can realistically enter the relevant market.

This is why Bronner, IMS Health, Magill, and Microsoft remain important.

9. Data Centres as Physical Essential Facilities

Data centres present an even more interesting issue because they combine physical and digital infrastructure.

A hyperscale data centre may involve:

enormous electrical capacity;

high-speed fibre connections;

specialised cooling;

GPU clusters;

backup power;

physical security;

network interconnection;

geographic advantages.

A competitor might theoretically construct another data centre.

But the competition-law question is not simply theoretical.

The relevant question is:

Can a reasonably efficient competitor duplicate the infrastructure within a reasonable time and at a reasonable cost?

This requires examination of:

land availability;

electricity availability;

grid connection;

planning restrictions;

fibre access;

construction time;

GPU availability;

financing;

geographic constraints.

Current data-centre expansion demonstrates why these questions can become economically significant: in 2026, regulatory concerns in the United States have included electricity-grid constraints and infrastructure costs associated with rapidly expanding data-centre demand. (Reuters)

10. Cloud Computing and Monopolisation

Cloud infrastructure creates a particularly complex competition problem.

A dominant cloud provider may control:

Compute + storage + networking + databases + AI tools + security + identity + APIs.

A customer may therefore become deeply integrated into the provider's ecosystem.

This creates several potential competition risks.

A. Switching costs

Moving enormous quantities of data can be:

expensive;

time-consuming;

technically difficult.

B. Data egress costs

Charges associated with moving data away from the platform can discourage switching.

C. Proprietary APIs

Applications may become dependent on provider-specific interfaces.

D. Technical interoperability

Competitors may technically exist but be unable to reproduce the same integration.

E. Bundling

Cloud infrastructure may be bundled with:

cybersecurity;

databases;

AI models;

productivity software;

advertising;

identity services.

F. Self-preferencing

The cloud provider may potentially privilege its own downstream applications.

11. AI Infrastructure Intensifies the Problem

Artificial intelligence adds a new layer.

A leading AI company may need:

GPUs;

cloud computing;

training datasets;

inference infrastructure;

electricity;

specialised networking;

foundation models;

distribution channels.

This produces a potential AI infrastructure bottleneck.

The competitive chain can be represented as:

Electricity
↓
Data centre
↓
GPU capacity
↓
Cloud platform
↓
Training data
↓
Foundation model
↓
API
↓
AI applications

If one undertaking controls several successive layers, it may possess opportunities for multi-level foreclosure.

12. Vertical Foreclosure

Vertical foreclosure occurs when a firm with market power at one level uses that power to restrict competition at another level.

For example:

Dominant cloud infrastructure
↓
preferential pricing
↓
own AI model becomes cheaper
↓
independent AI developers disadvantaged
↓
downstream competition reduced.

Or:

Dominant search engine
↓
control over search data
↓
preferential ranking
↓
own downstream service receives greater traffic
↓
competitors lose scale.

This is closely related to the reasoning examined in Google Shopping. (EUR-Lex)

13. Refusal to Deal vs Discriminatory Access

Competition law should distinguish two situations.

Scenario A — Complete refusal

"You cannot use the infrastructure."

This may trigger essential-facilities/refusal-to-deal analysis.

Scenario B — Discriminatory access

"You may use the infrastructure, but my own subsidiary receives substantially better access."

This may involve:

discriminatory terms;

self-preferencing;

margin issues;

leveraging;

exclusionary conduct.

The second situation can sometimes be legally easier to establish than a pure refusal-to-deal case because the authority may not need to establish the same level of indispensability required under the strictest essential-facilities doctrine.

The Google Shopping litigation illustrates this distinction. (EUR-Lex)

14. Network Effects and Data Monopolisation

Network effects can transform infrastructure into a competitive bottleneck.

Suppose:

More users → more data

and

More data → better service

and

Better service → more users.

The incumbent therefore obtains an advantage that a new entrant may find difficult to overcome.

This is sometimes described as a data-network effect.

Competition authorities should therefore examine not only current market share but also:

user growth;

data accumulation;

switching rates;

interoperability;

multi-homing;

access to alternative datasets;

entry costs;

data portability.

15. Lock-In as a Monopolisation Risk

Data infrastructure can create technical lock-in.

A customer may remain with a provider because moving requires:

transferring terabytes or petabytes of data;

rewriting applications;

changing APIs;

retraining models;

reconstructing databases;

changing security architecture;

renegotiating contracts.

This can produce an important distinction:

market share ≠ market power

but:

market share + high switching costs + infrastructure dependence + network effects

can create much stronger evidence of durable market power.

16. Interoperability as a Competition Remedy

Where infrastructure creates competitive bottlenecks, regulators may consider interoperability remedies.

Possible remedies include:

1. API access

Requiring a dominant provider to make relevant interfaces available.

2. Data portability

Allowing customers to move data to competing providers.

3. Interoperability

Ensuring competing systems can communicate.

4. Non-discrimination

Preventing a dominant provider from giving itself preferential access.

5. Functional separation

Separating infrastructure operations from downstream commercial operations.

6. Access obligations

Requiring access on fair and reasonable terms.

7. Transparency

Requiring disclosure of technical or contractual restrictions.

The EU's DMA illustrates the increasing importance of mandatory data-access mechanisms. In 2026, the Commission adopted measures concerning Google's sharing of anonymised search data with eligible competing services under Article 6(11). (Digital Markets Act (DMA))

17. Competition Law and Privacy Conflict

A particularly difficult issue is that competition law may favour data access while privacy law may restrict data disclosure.

For example:

Competition law:
"Give competitors access to the dataset."

Privacy law:
"You cannot disclose personal information without a lawful basis."

Therefore, compulsory access cannot simply mean:

"Give the competitor the raw personal data."

Possible solutions include:

anonymisation;

aggregation;

secure data environments;

query-based access;

privacy-preserving computation;

differential privacy;

controlled APIs.

This is especially important for India because the Competition Act must operate alongside India's data-protection framework.

18. Indian Competition Law Framework

The central provision is Section 4 of the Competition Act, 2002.

The relevant forms of abuse may include:

Section 4(2)(a)

Imposition of unfair or discriminatory conditions or prices.

Section 4(2)(b)

Limiting or restricting:

production;

technical development;

markets.

Section 4(2)(c)

Denial of market access.

Section 4(2)(e)

Using dominance in one relevant market to enter into or protect another relevant market.

The last two provisions are particularly important for data infrastructure.

19. Section 4(2)(c): Denial of Market Access

Suppose a dominant data infrastructure company controls a facility required by downstream competitors.

If it refuses access or provides discriminatory access, Section 4(2)(c) could become relevant.

The key question would be:

Has the conduct actually or potentially prevented competitors from accessing the market?

This does not mean every refusal constitutes abuse.

The CCI would need to examine:

dominance;

relevant market;

necessity of the infrastructure;

alternatives;

foreclosure effects;

objective justification;

proportionality.

20. Section 4(2)(e): Leveraging

Section 4(2)(e) is particularly important for infrastructure monopolisation.

Imagine:

Market A: cloud infrastructure
Market B: AI services.

If an enterprise possesses dominance in Market A and uses that dominance to protect its own position in Market B, Section 4(2)(e) may become relevant.

This captures the fundamental danger of infrastructure leverage.

21. Data Infrastructure and Merger Control

Monopolisation risks can also arise through acquisitions.

A dominant platform may acquire:

a major cloud provider;

a data-centre operator;

a unique dataset;

an AI infrastructure company;

a cybersecurity platform;

a data analytics company.

Even where the target has relatively low current revenues, the transaction may be competitively significant because of its strategic data or infrastructure value.

Therefore, traditional turnover-based merger analysis may sometimes fail to capture the full competitive significance of data assets.

India's competition authorities are already examining data-centre transactions; for example, the CCI approved CPPIB's acquisition of certain shareholding in Ctrl S Datacenters in May 2026, a business involved in data-centre, colocation and cloud-related services. (Press Information Bureau)

22. Key Competition Risks

The major risks can be summarised as follows:

23. Economic Justifications for Infrastructure Ownership

Competition law should not automatically punish infrastructure scale.

Large infrastructure can produce substantial benefits:

lower costs;

greater reliability;

improved security;

faster innovation;

economies of scale;

investment incentives;

better quality;

technological innovation.

The central legal challenge is therefore to distinguish:

Competition on the merits

from

exclusionary exploitation of infrastructure power.

This is why the Bronner principle remains so important: compulsory access can undermine incentives to build infrastructure in the first place. (EUR-Lex)

24. The "Build or Share" Problem

A fundamental policy dilemma exists.

If authorities say:

"Whenever infrastructure becomes successful, you must share it."

firms may reduce infrastructure investment.

But if authorities say:

"Infrastructure owners never have to provide access."

a dominant firm could potentially construct an unassailable bottleneck.

Competition law therefore seeks a middle ground.

The preferred approach

Exceptional access obligation + strong evidence of indispensability + safeguards against free-riding.

This preserves incentives to invest while preventing infrastructure from becoming an instrument of permanent exclusion.

25. A Proposed Legal Test for Data Infrastructure Monopolisation

A competition authority could analyse a case through eight stages.

Stage 1 — Define the relevant market

Is the relevant market:

cloud computing?

GPU capacity?

data-centre services?

data storage?

data analytics?

search?

AI infrastructure?

API services?

Stage 2 — Establish dominance

Consider:

market share;

entry barriers;

infrastructure scale;

network effects;

switching costs;

control of unique data;

financial resources.

Stage 3 — Identify the bottleneck

What exactly does the firm control?

It might be:

physical infrastructure;

data;

API;

algorithm;

technical standard;

network;

cloud platform.

Stage 4 — Test replicability

Can competitors reasonably reproduce it?

Examine:

cost;

time;

technical feasibility;

legal restrictions;

access to electricity;

access to land;

access to data;

access to customers.

Stage 5 — Examine conduct

Is there:

refusal?

discrimination?

self-preferencing?

tying?

bundling?

exclusivity?

capacity withholding?

interoperability restriction?

Stage 6 — Demonstrate foreclosure

Would the conduct:

exclude competitors;

raise their costs;

prevent entry;

reduce innovation;

restrict consumer choice?

Stage 7 — Consider objective justification

The dominant firm should be able to demonstrate legitimate reasons such as:

cybersecurity;

capacity constraints;

privacy;

intellectual-property protection;

technical integrity;

legitimate investment protection.

Stage 8 — Apply proportionality

Even where intervention is justified, the remedy should be no broader than necessary.

26. Comparative Case-Law Lessons

The major cases collectively establish an important hierarchy:

Terminal Railroad
→ infrastructure control can exclude competitors.

MCI v AT&T
→ essential-facilities conditions must be carefully established.

Commercial Solvents
→ upstream dominance can be leveraged against downstream competitors.

Magill
→ control over indispensable information can create exceptional access obligations.

Bronner
→ indispensability is a demanding requirement.

IMS Health
→ intellectual-property/data structures can raise exceptional refusal-to-license concerns.

Microsoft
→ interoperability can become a competition-law obligation.

Google Shopping
→ digital infrastructure can facilitate self-preferencing and leveraging without a conventional refusal to supply. (EUR-Lex)

Shamsher Kataria
→ Indian competition law can address denial of access through the broader prohibition of abuse of dominance.

Matrimony.com v Google
→ dominant digital gateways can create market-access and discrimination concerns.

27. The Central Legal Distinction

The most important distinction in this area is:

"Important infrastructure" is not automatically "essential infrastructure."

A competition authority should intervene only when the evidence establishes that infrastructure control is being transformed into a mechanism of durable exclusion or exploitation.

Therefore:

Scale alone ≠ abuse

Data ownership alone ≠ monopoly

Infrastructure ownership alone ≠ essential facility

High switching costs alone ≠ dominance

But the combination of:

dominance + bottleneck control + lack of realistic alternatives + exclusionary conduct + foreclosure effects

creates a much stronger competition-law case.

28. Conclusion

Data infrastructure is increasingly becoming the structural foundation of digital competition.

The traditional competition-law question was:

"Who controls the product?"

The modern question is increasingly:

"Who controls the infrastructure on which competing products depend?"

This change is particularly important in cloud computing, AI, search, data analytics, digital advertising and data-centre markets.

The case law from Terminal Railroad, MCI, Commercial Solvents, Magill, Bronner, IMS Health, Microsoft, Google Shopping, Shamsher Kataria and Matrimony.com provides a doctrinal foundation for analysing these problems.

However, the law should avoid treating every valuable dataset or large infrastructure system as an essential facility. The strongest cases arise where infrastructure is genuinely indispensable, difficult to replicate, controlled by a dominant undertaking, and deliberately used to foreclose downstream competition.

The emerging regulatory model is therefore moving toward a combination of:

competition law + interoperability + data portability + non-discrimination + merger control + sector-specific digital regulation.

The European Union's current data-access regime under the DMA is an important example of this shift: regulators are increasingly treating access to strategically important data as a means of preventing entrenched digital bottlenecks, rather than relying exclusively on traditional ex-post abuse-of-dominance litigation. (Digital Markets Act (DMA))

Core takeaway

The greatest competition-law risk from data infrastructure is not simply that one company owns a large amount of data or computing capacity; it is that control over an infrastructure bottleneck enables that company to determine who can compete, on what terms, and with what access to the essential inputs of the digital economy.