Competition Law and Competition Concerns in Memory Monopolies

1. Introduction

A memory monopoly arises where one undertaking obtains, controls, or accumulates an unusually large and strategically valuable stock of historical information and where that accumulated information becomes difficult for competitors to replicate or access. The “memory” may consist of:

• historical transaction data;
• consumer preferences and behavioural histories;
• search and browsing histories;
• purchase records;
• location and mobility histories;
• machine-learning training data;
• industrial and operational records;
• financial histories;
• customer interaction records;
• historical prices and demand patterns; and
• longitudinal datasets generated over many years.

Competition law does not ordinarily prohibit a firm merely because it possesses a large database. The competition concern arises where accumulated information becomes a competitive bottleneck and is used to exclude rivals, reinforce dominance, raise entry barriers, discriminate between downstream businesses, facilitate coordination, or prevent effective switching.

The concept therefore sits at the intersection of dominance, data advantages, essential facilities, exclusionary conduct, refusal to supply, tying, leveraging, mergers, privacy and digital-platform competition.

2. Meaning of a Memory Monopoly

A memory monopoly can be understood through five characteristics:

A. Historical accumulation

The undertaking possesses information accumulated over a long period.

B. Scale

The database contains information that competitors cannot easily reproduce.

C. Persistence

The information becomes more valuable because it contains historical observations rather than merely current information.

D. Feedback effects

More users generate more information, which improves the service, which attracts additional users, producing still more information.

E. Competitive dependency

Competitors may become dependent upon access to the accumulated information to compete effectively.

A simplified model is:

Users → Data → Better prediction/service → More users → More data → Stronger market position

This creates a potentially self-reinforcing competitive advantage.

3. Why Memory Can Have Monopoly Characteristics

Not every large database constitutes a monopoly. Competition concerns become stronger where the accumulated information has several of the following characteristics:

4. Competition-Law Framework

Memory monopolies can potentially engage several areas of competition law.

A. Relevant Market

The first question is whether the undertaking possesses market power in a properly defined relevant market.

Possible markets include:

• search services;
• online advertising;
• digital payments;
• cloud services;
• financial-information services;
• health-data services;
• consumer analytics;
• AI services;
• e-commerce marketplaces; or
• data brokerage.

The relevant market may also have zero monetary price, meaning that consumers provide attention, behavioural information, or data rather than money.

5. Dominance

Possession of a large historical database can contribute to dominance, but data ownership alone does not automatically establish dominance.

Authorities may examine:

1. market share;
2. quality and uniqueness of data;
3. barriers to collecting equivalent data;
4. network effects;
5. economies of scale and scope;
6. switching costs;
7. interoperability;
8. access to alternative sources;
9. ability of competitors to reproduce the dataset; and
10. duration of the informational advantage.

The important question is:

Does the accumulated memory provide market power that competitors cannot reasonably overcome?

6. Data as a Barrier to Entry

A memory monopoly can create a particularly strong dynamic barrier to entry.

A new entrant may technically be able to build the same service but lack the historical data necessary to achieve comparable performance.

For example, an established platform may possess:

• ten years of search histories;
• years of consumer purchases;
• extensive advertising-response data;
• historical pricing information; and
• behavioural profiles.

A new entrant starting with zero historical observations may have difficulty matching the incumbent's prediction capabilities.

This is sometimes described as a data-based learning advantage.

7. Network Effects and Memory Effects

Traditional network effects concern the increasing value of a service as more users join.

Memory monopolies add another mechanism:

More users → More historical data → Better product → More users

This may produce a data feedback loop.

Competition authorities therefore need to distinguish between:

• ordinary economies of scale;
• legitimate learning effects; and
• exclusionary accumulation of information.

The mere existence of a data advantage is not sufficient to establish an infringement.

8. Refusal to Provide Historical Data

A dominant undertaking may possess information that competitors need to provide competing services.

Potential concerns arise if the dominant undertaking:

• refuses access without objective justification;
• prevents interoperability;
• prevents data portability;
• imposes discriminatory access conditions;
• technically restricts extraction;
• provides incomplete historical records; or
• supplies data to its own downstream service while denying it to competitors.

This can bring memory monopolies close to the essential-facilities/refusal-to-deal doctrine.

However, competition law generally does not create an automatic obligation for dominant firms to share every database they possess.

9. Data Portability

Data portability can reduce the competitive importance of accumulated memory.

Suppose a consumer has used one platform for ten years and accumulated:

• purchase histories;
• playlists;
• contacts;
• preferences;
• transaction histories;
• reviews; and
• personalised recommendations.

If leaving the platform means losing all historical information, switching costs increase.

Consequently:

Historical data + switching costs = stronger customer lock-in

Competition authorities may therefore examine whether restrictions on portability reinforce dominance.

10. Self-Preferencing

A dominant platform may use its historical information to favour its own downstream services.

For example, a marketplace could use years of seller and consumer information to:

1. identify profitable product categories;
2. launch its own competing products;
3. use historical sales data to optimise pricing;
4. rank its own products more prominently; and
5. restrict competing sellers' access to equivalent information.

The competition concern is not simply that the platform has information.

The concern is that the platform may use information obtained through its intermediary position to disadvantage competitors operating on that intermediary.

11. Leveraging

Memory monopolies can also facilitate leveraging.

A firm dominant in one market may use historical information obtained there to enter an adjacent market.

For example:

Dominant consumer platform → exclusive behavioural data → financial scoring → financial-services market

or:

Dominant marketplace → seller data → private-label products → retail market

The legal question is whether the informational advantage is being used in an exclusionary manner rather than merely competing vigorously.

12. Discriminatory Access

A dominant information provider could provide:

• superior data to affiliated businesses;
• delayed data to competitors;
• aggregated data to outsiders;
• granular data to its own subsidiary;
• cheaper access to preferred partners; or
• restricted APIs to competing businesses.

This creates potential discriminatory-access concerns.

The relevant comparison is often:

What information does the dominant undertaking possess, what information does it provide to competitors, and what information does it provide to itself or affiliated entities?

13. Tying and Bundling

Memory can also be connected to another product through tying.

For example, a dominant platform could require businesses to purchase:

historical analytics + advertising services

or:

data-storage service + proprietary analytics platform

The concern becomes stronger where access to the historical information is conditional upon purchasing another product.

14. Exclusive Data Accumulation

Exclusive agreements can transform ordinary data collection into an exclusionary strategy.

Examples include agreements requiring:

• merchants to provide transaction data exclusively;
• suppliers to share customer information only with one platform;
• advertisers to use a particular measurement system;
• hospitals to provide historical data exclusively to one analytics provider.

Such arrangements may foreclose competitors from obtaining equivalent information.

15. Algorithmic Competition

Historical information is particularly important for artificial intelligence and algorithmic systems.

A firm possessing extensive historical data may train algorithms that competitors cannot replicate.

This can affect:

• recommendation systems;
• fraud detection;
• credit scoring;
• advertising;
• search;
• autonomous systems;
• demand forecasting;
• healthcare analytics; and
• generative AI.

Thus, memory can become an input into algorithmic market power.

16. Merger Control

Memory monopolies raise significant issues in mergers.

A transaction may combine:

• a large consumer database with another large database;
• complementary datasets;
• customer information with transaction information;
• historical search data with advertising data; or
• proprietary industrial data with AI technology.

Even where the parties' existing market shares are moderate, the merger may create a powerful combined informational asset.

Authorities can therefore examine:

Horizontal effects

Will the merger eliminate an important competitor?

Vertical effects

Will the combined entity restrict rivals' access to information?

Conglomerate effects

Will information from one market strengthen the firm's position in another?

Innovation effects

Will the merger eliminate a potential competitor that could develop an alternative data ecosystem?

17. Privacy and Competition

Privacy and competition can overlap in memory monopolies.

A dominant platform may offer apparently free services while collecting extensive information.

If users cannot realistically switch because their historical information is trapped, competitive pressure concerning privacy quality may weaken.

Thus, competition may occur not only over:

• price,

but also over:

• privacy;
• data minimisation;
• transparency;
• security;
• interoperability; and
• user control.

18. Relevant Case Laws

The following cases provide important foundations for analysing competition concerns associated with memory monopolies.

1. IMS Health GmbH & Co. OHG v NDC Health GmbH & Co. KG

Court: Court of Justice of the European Union
Citation: Case C-418/01

This is one of the most important cases for understanding information-related monopolisation.

IMS Health possessed a proprietary system for organising pharmaceutical sales information. Competitors sought access to the information structure.

The CJEU considered the circumstances in which refusal by a dominant undertaking to license an intellectual-property-related asset could constitute abuse.

The Court identified stringent conditions associated with compulsory access, including circumstances where access is indispensable, refusal prevents the emergence of a new product or service, and the refusal lacks objective justification.

Relevance to memory monopolies

Where historical information becomes indispensable to competing services, IMS Health provides an important framework for analysing:

• data access;
• interoperability;
• information infrastructure;
• indispensability; and
• refusal to license or provide access.

19. Microsoft Corp. v Commission

Court: General Court of the European Union
Citation: Case T-201/04

Microsoft was found to have abused its dominant position through, among other conduct, restrictions concerning interoperability information.

The case demonstrated that technological information controlled by a dominant firm can have significant competitive importance where rivals depend upon it to develop interoperable products.

Relevance

The case is useful for understanding:

• interoperability;
• technological information;
• refusal to provide information;
• leveraging;
• network effects; and
• exclusion of competitors.

A memory monopoly may similarly become problematic where historical information is necessary for interoperability or effective competition.

20. Google Search (Shopping)

Case: Google Search (Shopping)
EU Commission decision: 2017; General Court judgment, Case T-612/17

The case concerned Google's conduct in relation to comparison-shopping services and the placement of its own specialised search results.

The General Court upheld the essential finding that Google had abused its dominant position through conduct involving preferential positioning and display of its own comparison-shopping service.

Relevance to memory monopolies

The case is important where a platform:

• possesses extensive historical search information;
• operates as an intermediary;
• competes with businesses using the platform; and
• uses its platform position to favour its own downstream service.

It illustrates the broader platform-leverage/self-preferencing problem that can interact with accumulated informational advantages.

21. Google Android

Case: Google Android
EU Commission decision: 2018; General Court judgment, Case T-604/18

The case concerned several practices involving Google's Android ecosystem, including restrictions connected with search and mobile distribution.

The litigation examined how contractual arrangements within a digital ecosystem could reinforce Google's position in search.

Relevance

The case helps illustrate how a dominant digital ecosystem can use complementary products and contractual arrangements to preserve an established informational position.

For memory monopolies, this is relevant to:

• ecosystem lock-in;
• data accumulation;
• default arrangements;
• leveraging; and
• reinforcing dominance across adjacent markets.

22. Facebook/Meta Data and Terms Cases

European competition authorities have also examined the relationship between Facebook's data practices and competition.

A particularly significant case is:

Bundeskartellamt v Facebook/Meta

The German competition authority's proceedings concerned Facebook's combination of user data from different sources and the relationship between data practices and Facebook's dominant position.

The German Federal Court of Justice subsequently dealt with the competition-law assessment of Facebook's data practices.

Relevance

This line of litigation demonstrates that:

• data collection can be relevant to competition law;
• exploitation and exclusion can overlap;
• privacy-related conditions may have competition significance; and
• dominance can affect the competitive assessment of data terms.

This is particularly relevant to memory monopolies because extensive historical profiles can become a source of market power.

23. United States v Google LLC

Court: U.S. District Court for the District of Columbia
2024 judgment concerning search distribution practices

The U.S. government's case against Google examined Google's agreements and practices concerning distribution of its search engine.

The litigation is important because search dominance can generate extensive historical user and query information, which in turn can reinforce search quality, advertising capabilities and competitive advantages.

Relevance to memory monopolies

The case illustrates the relationship between:

distribution → users → queries/data → quality and monetisation → further distribution

This is closely related to the feedback mechanism underlying data-based competitive advantages.

24. United States v Microsoft Corp.

Court: U.S. District Court for the District of Columbia; D.C. Circuit

The Microsoft litigation examined Microsoft's conduct concerning operating systems and competing technologies, particularly its treatment of Netscape and Java.

The case is a foundational authority on exclusionary conduct involving a dominant technological platform.

Relevance

Its principles are useful for memory monopolies involving:

• platform control;
• network effects;
• technological ecosystems;
• exclusionary agreements;
• leveraging; and
• barriers to entry.

The case demonstrates that competition law may address strategies that protect an entrenched technological position even where the protected position initially resulted from successful competition.

25. Bronner v Mediaprint

Court: CJEU
Citation: Case C-7/97

The Court considered when access to an infrastructure controlled by a dominant undertaking could be required under Article 102 TFEU.

The Court adopted a demanding approach to the concept of indispensability.

Relevance to memory monopolies

Bronner is important because a database should not automatically be classified as an essential facility simply because it is valuable.

A competition authority must carefully examine:

• whether alternatives exist;
• whether access is genuinely indispensable;
• whether duplication is economically or technically feasible; and
• whether refusal eliminates effective competition.

26. Magill

Cases: Joined Cases C-241/91 P and C-242/91 P

Magill concerned television programme information and the refusal to license certain information.

The case is a major authority on compulsory licensing and refusal to supply information protected by intellectual-property rights.

Relevance

Magill is particularly useful for analysing situations where:

• information is commercially valuable;
• an undertaking controls the relevant information;
• competitors require that information;
• refusal prevents a new product or service; and
• access may be necessary for effective competition.

It therefore provides an important conceptual foundation for information-based monopolies.

27. Competition Concerns Specific to Memory Monopolies

A. Historical-data foreclosure

An incumbent may accumulate information faster than rivals and thereby make market entry increasingly difficult.

B. Data hoarding

A dominant undertaking may collect information without making equivalent information available to dependent businesses.

C. Data portability restrictions

Users may remain with a platform because their accumulated history cannot be transferred effectively.

D. Interoperability restrictions

Competitors may be unable to use historical information in interoperable systems.

E. Self-preferencing

The dominant platform may exploit information gathered from third-party users to favour its own competing products.

F. Discriminatory access

Affiliated companies may receive better-quality information than independent competitors.

G. Leveraging

Historical information from one market may be used to establish or reinforce dominance in another.

H. Predatory acquisition of data

Acquisitions may be used to eliminate emerging competitors or consolidate complementary datasets.

I. Algorithmic advantages

Historical information may give a dominant firm superior machine-learning performance.

J. Coordinated effects

Shared historical market information can potentially make monitoring competitors easier and facilitate coordination.

28. Memory Monopolies and Essential Facilities

The essential-facilities doctrine requires particular caution.

A database may be economically important without being legally indispensable.

A useful analytical sequence is:

Is the undertaking dominant?

↓

Is the information genuinely unique?

↓

Can competitors reproduce or obtain equivalent information?

↓

Is access indispensable?

↓

Would refusal eliminate effective competition?

↓

Is there objective justification for refusal?

↓

What access remedy would be proportionate?

This prevents competition law from becoming a general requirement that businesses share all commercially valuable information.

29. Remedies

Where competition law establishes an infringement, possible remedies may include:

1. Data access

Controlled access to specified information.

2. API access

Technical interfaces enabling competitors to obtain data.

3. Data portability

Allowing consumers or businesses to transfer historical information.

4. Interoperability

Technical standards allowing competing services to interact.

5. Non-discrimination

Preventing discriminatory treatment between the dominant firm's own services and independent rivals.

6. Data separation

Separating datasets obtained from different business activities.

7. Behavioural restrictions

Preventing the dominant firm from using intermediary data to compete unfairly with businesses dependent upon the intermediary.

8. Structural remedies

In exceptional circumstances, competition authorities may consider structural measures where behavioural remedies are inadequate.

30. Indian Competition-Law Perspective

Under the Competition Act, 2002, memory-monopoly issues can potentially arise principally under:

Section 4 — Abuse of dominant position

Potential theories include:

• unfair or discriminatory conditions;
• unfair or discriminatory prices;
• denial of market access;
• limiting or restricting markets;
• leveraging dominance in one relevant market into another;
• discriminatory access to information; and
• exclusionary platform practices.

Section 3 — Anti-competitive agreements

Potential concerns include:

• exclusive data-sharing agreements;
• restrictive information-exchange arrangements;
• data-access restrictions;
• exclusive dealing; and
• coordinated use of commercially sensitive information.

Section 5 and Section 6 — Combinations

Transactions involving large databases may raise concerns where a merger combines complementary information assets and substantially strengthens market power.

31. China Perspective

China's Anti-Monopoly Law is also relevant to memory-monopoly problems involving digital platforms and data.

Potential concerns include:

• abuse of a dominant market position;
• discriminatory treatment;
• refusal to deal;
• tying;
• unreasonable trading conditions;
• exclusionary platform practices;
• concentration of data resources; and
• merger-related data advantages.

China's digital-platform enforcement experience demonstrates the increasing importance of examining data, algorithms and platform ecosystems as sources of competitive power, rather than looking only at traditional price and output measures.

32. Key Distinction: Data Advantage vs Memory Monopoly

This distinction is crucial.

Ordinary data advantage

A firm has more information because it has competed successfully and invested in collecting it.

Not necessarily an antitrust problem.

Memory monopoly

The accumulated information becomes a durable competitive bottleneck and the undertaking uses that position to exclude competitors, prevent switching, restrict access, or extend dominance.

Potential competition-law problem.

Therefore:

Competition law should generally protect competition in the use of information, rather than punish firms merely for possessing successful historical datasets.

33. Analytical Framework for Courts and Competition Authorities

A useful six-stage framework is:

Stage 1 — Identify the memory asset

What historical information is being controlled?

Stage 2 — Define the market

In which market does the information generate competitive significance?

Stage 3 — Establish market power

Does the database materially contribute to dominance?

Stage 4 — Examine conduct

Has the undertaking:

• denied access?
• restricted portability?
• discriminated?
• self-preferenced?
• tied products?
• leveraged information?
• entered exclusionary agreements?

Stage 5 — Examine competitive effects

Has the conduct:

• raised entry barriers?
• eliminated competitors?
• reduced innovation?
• increased switching costs?
• reduced consumer choice?
• weakened quality or privacy competition?

Stage 6 — Consider justification and remedy

Is the conduct objectively justified, and would a proposed remedy preserve incentives to innovate while restoring competitive access?

34. Key Principles Emerging from the Case Law

The case law collectively supports several principles:

1. Information can have competitive significance comparable to physical infrastructure.
2. Possession of unique information does not automatically create an antitrust duty to share it.
3. Indispensability is important in refusal-to-access cases.
4. Interoperability information can be competitively critical.
5. Network effects can reinforce informational advantages.
6. Platform dominance can make data practices more significant under competition law.
7. Historical data can create substantial entry barriers.
8. Self-preferencing can become more significant when the platform possesses extensive information about dependent businesses.
9. Data-related mergers can create competitive risks even where traditional market-share analysis appears less conclusive.
10. Privacy, portability and competition can interact where users are locked into a service by accumulated historical information.

35. Conclusion

Memory monopolies represent a modern form of potential market power in which historical information itself becomes a competitive asset. Unlike traditional monopolies based on physical infrastructure or exclusive ownership of natural resources, memory monopolies can arise from the accumulation of behavioural, commercial, technological and algorithmic information over time.

The principal competition concerns are data-based entry barriers, refusal of access, interoperability restrictions, data portability limitations, self-preferencing, discriminatory access, leveraging, exclusive data arrangements, algorithmic advantages and data-driven mergers.

The leading authorities—including Magill, IMS Health, Bronner, Microsoft, Google Shopping, Google Android, the Facebook/Meta proceedings and U.S. Microsoft and Google litigation—provide useful principles for evaluating when control over information crosses the line from a legitimate competitive advantage into potentially exclusionary market power.

The central legal question is therefore not:

“Does one firm possess most of the memory?”

but rather:

“Does control over accumulated memory create or reinforce market power, and is that power being used in a manner that restricts effective competition?”

That distinction is essential for applying competition law to AI systems, digital platforms, data markets, financial information, healthcare databases, e-commerce ecosystems and other information-intensive industries.