Competition Law and Future-Oriented Antitrust Theories for Intelligent Networks

1. Introduction

Intelligent networks are networks in which software, artificial intelligence, sensors, algorithms, cloud infrastructure, connected devices, data systems and automated decision-making interact continuously. Examples include:

• AI-agent networks;
• Internet-of-Things (IoT) ecosystems;
• connected-vehicle networks;
• smart grids;
• telecommunications networks;
• cloud-computing ecosystems;
• autonomous logistics networks;
• digital payment networks;
• smart-city infrastructure; and
• interconnected platform ecosystems.

Traditional competition law generally examines markets, firms, products and transactions. Intelligent networks require a broader analytical approach because competitive power may arise not merely from market share, but from control over network architecture, data, interoperability, protocols, standards, computational resources, switching pathways and ecosystem participation.

This is increasingly visible in modern enforcement. The EU's current Digital Markets Act framework, for example, expressly addresses interoperability and data portability in mobile ecosystems, while 2026 proceedings concerning Google's Android require interoperability for competing AI services.

2. Meaning of Intelligent Networks

An intelligent network can be understood as:

A technologically interconnected system in which multiple users, firms, devices, platforms or infrastructure components interact through data-driven and increasingly automated mechanisms.

Its principal characteristics are:

1. Network effects – the value of participation may increase as the number of participants increases.
2. Data feedback loops – more users generate more data, which can improve algorithms and attract further users.
3. Algorithmic coordination – software can influence prices, access, ranking and allocation.
4. Interoperability dependence – competing systems may need access to technical interfaces.
5. Switching costs – users may find it costly to migrate data, applications, devices or workflows.
6. Multi-sidedness – consumers, developers, suppliers, advertisers and complementors may participate simultaneously.
7. Ecosystem integration – several formally separate markets can become technologically interconnected.
8. Dynamic competition – today's small platform may become tomorrow's infrastructure layer.

Consequently, competition can occur within a market, between platforms, between ecosystems and between alternative technological architectures.

3. Traditional Antitrust Theory Versus Future-Oriented Theory

The purpose is not to abandon conventional antitrust principles. Rather, future-oriented theories attempt to apply them to environments where competitive constraints are generated by technological interdependence.

4. Major Future-Oriented Antitrust Theories

A. Network-Orchestrator Theory

The first theory treats the network orchestrator as a potentially important source of competitive power.

An orchestrator may control:

• operating systems;
• cloud infrastructure;
• APIs;
• app stores;
• payment rails;
• data access;
• identity systems;
• technical standards;
• AI models; or
• network certification.

The central question becomes:

Does control over the network allow one undertaking to influence competition in connected markets?

Google Android provides an important example. The EU General Court described the case in terms of a multi-sided platform and ecosystem involving Android, Play Store, Google Search, Chrome, device manufacturers and mobile-network operators.

Future implication

Competition authorities may increasingly investigate architectural control, rather than merely asking whether an undertaking has a high market share in one conventional product market.

5. Network Effects and Feedback-Loop Theory

Intelligent networks frequently produce positive feedback loops:

Users → Data → Better algorithms → Better service → More users → More data

This can create rapid concentration.

A conventional market-share analysis may underestimate this phenomenon because the relevant competitive advantage is not simply present market share but the rate at which the network reinforces itself.

Competition concerns

• self-reinforcing dominance;
• data accumulation;
• exclusion of emerging competitors;
• reduced contestability;
• acquisition of nascent competitors;
• discriminatory access;
• preferential algorithmic treatment.

The Google Android litigation is particularly relevant because the Commission's theory involved network effects and restrictions concerning device manufacturers and application developers. The General Court's judgment specifically addressed product bundles, exclusivity payments and anti-fragmentation obligations.

6. Interoperability Theory

In intelligent networks, interoperability can be a condition of effective competition.

A dominant network may technically permit third-party participation while withholding:

• APIs;
• technical documentation;
• authentication;
• device functionality;
• interoperability protocols;
• data interfaces;
• messaging capabilities;
• operating-system functionality.

The resulting system can be formally "open" but practically closed.

Important development: Google/Enel X

In Alphabet/Google – Android Auto / Enel X, the Court of Justice considered whether refusal by a dominant undertaking to make a digital platform interoperable with a third-party application could constitute abusive conduct.

The dispute concerned Enel X's request for interoperability between its JuicePass electric-vehicle charging application and Google's Android Auto. The Court addressed the circumstances in which refusal of interoperability can have anticompetitive effects and when objective justification may exist.

Future significance

This principle can extend conceptually to:

• autonomous vehicles;
• smart-home systems;
• AI assistants;
• wearable devices;
• healthcare platforms;
• industrial IoT;
• smart-grid systems.

7. Data-Network Theory

Data can function as a competitive infrastructure asset.

An intelligent network may accumulate:

• behavioural data;
• location data;
• transaction data;
• sensor data;
• search data;
• vehicle data;
• energy-consumption data;
• training data;
• network-performance data.

The important question becomes:

Does exclusive control over a strategically important dataset materially restrict competitive entry or innovation?

The EU's 2026 DMA measures concerning Google Search are illustrative: the Commission required mechanisms giving third-party search engines access to search data that Google Search can collect at scale.

Future theory

Data bottleneck theory would examine whether a dataset functions similarly to an essential competitive input.

However, data should not automatically be treated as an essential facility. Authorities would need to consider:

• uniqueness;
• replicability;
• time sensitivity;
• quality;
• cost of collection;
• privacy restrictions;
• interoperability;
• competitive necessity.

8. Algorithmic Network Theory

In intelligent networks, algorithms can become the functional decision-makers of markets.

Algorithms may determine:

• prices;
• rankings;
• access;
• search results;
• advertising allocation;
• credit decisions;
• transportation routes;
• energy allocation;
• inventory;
• recommendations.

Competition law therefore increasingly has to examine algorithmic architecture, not simply human agreements.

Possible theories of harm

1. Algorithmic collusion.
2. Algorithmic discrimination against rivals.
3. Algorithmic self-preferencing.
4. Algorithmic exclusion.
5. Coordinated use of common pricing software.
6. Personalised foreclosure.
7. Dynamic exploitation of switching costs.

The future issue is particularly difficult where the algorithm continuously adapts without an explicit human agreement.

9. AI-Agent Network Theory

The next generation of intelligent networks may contain AI agents capable of negotiating with other AI agents.

For example:

Buyer AI → Supplier AI → Logistics AI → Payment AI → Insurance AI

The agents may independently:

• negotiate;
• compare prices;
• allocate supply;
• select suppliers;
• determine routes;
• make purchases.

This creates a new antitrust question:

When autonomous agents produce coordinated market outcomes, what constitutes legally relevant coordination?

Traditional concepts of agreement and concerted practice may need adaptation to situations involving:

• machine-to-machine negotiation;
• common optimisation objectives;
• shared algorithmic infrastructure;
• automated responses;
• predictive pricing.

The legal inquiry would still need to establish the relevant human, contractual or technological connection rather than assuming that parallel algorithmic behaviour itself proves unlawful coordination.

10. Ecosystem Competition Theory

An intelligent network may encompass several connected markets.

For example:

Smartphone → Operating System → App Store → Payment System → Cloud → AI Assistant → Advertising

Competition in one layer may affect competition throughout the ecosystem.

The European Commission's ecosystem analysis increasingly considers factors such as interconnection, technical integration, interoperability, network effects and data-driven interdependencies.

Core proposition

A firm can possess competitive power because it controls an ecosystem, even where no individual component market completely explains its strategic position.

11. Six Major Case Laws

1. United States v. Microsoft Corp. — 253 F.3d 34 (D.C. Cir. 2001)

Facts

Microsoft possessed substantial power in operating systems and was accused of using that position to restrict competing technologies, particularly Netscape and Java.

Principle

The case demonstrates the importance of:

• network effects;
• applications barriers to entry;
• platform control;
• exclusionary agreements;
• technological tying;
• protection of emerging competition.

Relevance to intelligent networks

An intelligent network can similarly become difficult to challenge once:

users + developers + applications + data + complementary products

reinforce one another.

Microsoft therefore provides an important conceptual foundation for analysing platform-based network power.

2. Google Android — European Commission, Case AT.40099

The Commission found concerns involving Android's ecosystem, including arrangements concerning:

• device manufacturers;
• mobile network operators;
• Google Search;
• Chrome;
• Play Store;
• anti-fragmentation requirements.

The General Court subsequently considered the case in Google LLC and Alphabet v European Commission, Case T-604/18.

Significance

The case demonstrates how competition analysis can extend beyond one product to an interconnected ecosystem.

Intelligent-network relevance

It provides a framework for analysing:

• ecosystem foreclosure;
• network effects;
• interoperability;
• platform governance;
• default positioning;
• technical fragmentation.

3. Google/Enel X — Case C-233/23

This is particularly significant for intelligent networks.

The dispute concerned Android Auto and Enel X's JuicePass electric-vehicle charging application.

The Court of Justice addressed refusal of interoperability between a dominant digital platform and a third-party application, including the conditions relevant to objective justification.

Intelligent-network significance

The case connects competition law with:

• EV infrastructure;
• connected vehicles;
• digital interfaces;
• platform interoperability;
• smart mobility.

It illustrates how competition problems can emerge when physical infrastructure and digital platforms become inseparable.

4. FTC v. Qualcomm Inc.

The FTC challenged Qualcomm's conduct concerning modem-chip technology and alleged that Qualcomm used anticompetitive tactics to maintain monopoly positions. The district court ruled for the FTC in 2019, but the Ninth Circuit reversed in 2020.

Significance

The litigation illustrates the complexity of analysing:

• technology platforms;
• patent licensing;
• component markets;
• network standards;
• vertical relationships;
• innovation incentives.

Intelligent-network relevance

Modern intelligent networks depend upon semiconductor, connectivity and standards layers. Control at an upstream technological layer may therefore affect competition downstream.

5. Microsoft/Activision Blizzard — European Commission, Case M.10646

The transaction involved Microsoft's acquisition of Activision Blizzard and raised issues concerning Microsoft's broader gaming ecosystem, including cloud gaming.

The case is significant because ecosystem analysis can extend beyond traditional horizontal overlaps and examine how control over complementary products may affect emerging markets.

Contemporary scholarship identifies Microsoft/Activision and Booking/eTraveli as important examples of the tension between static market analysis and forward-looking ecosystem theories.

Intelligent-network relevance

The case demonstrates the importance of analysing:

• nascent markets;
• cloud-based services;
• complementary products;
• ecosystem expansion;
• potential foreclosure.

6. Booking/eTraveli — European Commission, Case M.10615

The European Commission prohibited Booking's acquisition of eTraveli in 2023.

The transaction was examined in the context of Booking's broader connected travel ecosystem, where accommodation, flights and other travel services could reinforce one another. Contemporary analysis identifies this decision as a significant example of ecosystem-based merger theory.

Intelligent-network relevance

The case illustrates a future-oriented concern:

A transaction involving apparently complementary services may strengthen an ecosystem's ability to reinforce its position across interconnected markets.

This is highly relevant to intelligent networks where data and cross-service integration can make separate services strategically interdependent.

12. Additional Relevant Authority: Google Search / AdSense

The European Commission's Google Search (AdSense) decision provides another important example of the relationship between platform control and competition.

It demonstrates how a dominant digital intermediary can influence competitive opportunities in connected advertising markets.

For intelligent networks, the analogous issue would arise where a network operator controls both:

the infrastructure through which market participants interact

and

the algorithm determining which participants receive access or visibility.

13. New 2026 Development: AI Interoperability

The future-oriented dimension is particularly clear in current EU enforcement.

In July 2026, the European Commission adopted binding specifications requiring Google to facilitate effective interoperability between competing AI services and key Android functionalities. The Commission stated that competing AI services should receive access comparable to Google's own AI services to relevant Android capabilities.

This is important because the competition problem is no longer simply:

Google Search vs another search engine

or

Android vs another operating system.

It increasingly concerns:

AI agent → operating system → applications → devices → data → services.

That is a quintessential intelligent-network problem.

14. Intelligent Networks and Essential Facilities

Traditional essential-facility doctrine asks whether a dominant undertaking controls infrastructure that competitors need.

In intelligent networks, the potentially essential component might be:

• API access;
• authentication infrastructure;
• cloud infrastructure;
• interoperability protocols;
• data interfaces;
• payment rails;
• app-distribution systems;
• network standards;
• identity systems;
• AI-compute infrastructure.

However, mandatory access should not automatically follow merely because a facility is technologically important.

Authorities must consider:

1. indispensability;
2. replication possibilities;
3. technical feasibility;
4. investment incentives;
5. security;
6. privacy;
7. intellectual-property interests;
8. proportionality;
9. competitive necessity.

15. Switching Costs and Migration Barriers

Intelligent networks can create unusually high switching costs because users may have to migrate:

• data;
• devices;
• applications;
• subscriptions;
• identities;
• AI preferences;
• transaction histories;
• business workflows;
• trained models;
• connected-device configurations.

This creates a new form of migration-based market power.

Competition law may therefore need to consider whether an undertaking deliberately makes switching technically or economically difficult.

The EU's DMA explicitly addresses data portability, recognizing that users may hold substantial amounts of important information within mobile ecosystems and may be reluctant to switch if that information cannot move with them.

16. Self-Preferencing in Intelligent Networks

Self-preferencing becomes particularly powerful when the network operator controls both:

1. the infrastructure, and
2. the decision-making algorithm.

Examples include:

• a smart-home platform favouring its own devices;
• an AI marketplace ranking its own AI models;
• a cloud platform favouring its own applications;
• an EV network favouring its own charging services;
• a digital assistant favouring affiliated services.

The EU's 2026 DMA enforcement against Google included findings concerning preferential treatment of Google's own services in Search ranking.

17. Intelligent Networks and Merger Control

Traditional merger control often focuses on:

horizontal overlap + market shares + concentration.

Intelligent-network merger control increasingly requires analysis of:

A. Data acquisition

Will the transaction combine unique datasets?

B. Ecosystem expansion

Will a platform acquire a complementary service?

C. Nascent competition

Could the target become an important future competitor?

D. Infrastructure control

Will the merged entity control an essential technological layer?

E. Interoperability

Could the acquirer restrict compatibility after acquisition?

F. AI feedback loops

Could combined data improve AI systems in a manner that makes entry more difficult?

18. Future Theory: Contestability of Network Architecture

A particularly important future-oriented concept is architectural contestability.

Traditional competition asks:

Can another firm enter the market?

Intelligent-network analysis should additionally ask:

Can another firm construct a competing network architecture?

A market may appear open while its underlying architecture is extremely difficult to replicate because an incumbent controls:

• users;
• data;
• developers;
• standards;
• devices;
• cloud infrastructure;
• identity;
• payments;
• AI models.

Therefore, competition authorities may increasingly examine contestability of the network itself, rather than simply contestability of an individual product.

19. Remedies for Intelligent-Network Antitrust Problems

Possible remedies include:

1. Interoperability obligations

Require technically effective access to relevant network functionality.

2. Data portability

Allow users and businesses to transfer relevant data.

3. API access

Provide fair and non-discriminatory technical access.

4. Non-discrimination

Prevent discriminatory treatment between the operator's services and rivals.

5. Data-use restrictions

Prevent the strategic combination or exploitation of competitively sensitive data.

6. Structural separation

In exceptional circumstances, separate infrastructure from competitive downstream operations.

7. Merger remedies

Require licensing, interoperability or access commitments.

8. Algorithmic transparency

Where proportionate, provide information enabling regulatory assessment of discriminatory or exclusionary mechanisms.

9. Switching facilitation

Reduce unnecessary technical and contractual barriers to migration.

10. Continuous monitoring

Intelligent networks evolve rapidly, making one-time remedies potentially insufficient.

20. Key Legal Issues for Future Litigation

Future intelligent-network cases are likely to involve questions such as:

1. What constitutes the relevant market?
2. Is an ecosystem itself a relevant competitive unit?
3. When does network control constitute dominance?
4. When does interoperability become legally necessary?
5. Can data constitute an indispensable input?
6. When does self-preferencing become abusive?
7. How should algorithmic coordination be proved?
8. How should AI-agent conduct be attributed to firms?
9. How should potential competition be evaluated?
10. How should innovation competition be measured?
11. Can technical architecture itself constitute exclusionary conduct?
12. How should merger control address future ecosystem expansion?
13. How should privacy and competition objectives interact?
14. What remedies preserve innovation while preventing foreclosure?

21. Emerging Antitrust Framework

A future-oriented intelligent-network framework can be represented as:

Network Infrastructure
↓
Connectivity & Interoperability
↓
Data Generation
↓
AI / Algorithmic Processing
↓
User & Business Participation
↓
Network Effects
↓
Ecosystem Expansion
↓
Switching Costs
↓
Potential Entrant Foreclosure

Competition authorities therefore need to examine the entire competitive feedback loop, rather than one isolated commercial practice.

22. Core Principles

The emerging theory can be condensed into ten principles:

1. Network power can extend beyond product-market power.
2. Interoperability can be a competitive resource.
3. Data can reinforce network effects.
4. Algorithms can become instruments of exclusion or coordination.
5. Ecosystems can generate power across multiple connected markets.
6. Switching costs can preserve dominance even without contractual exclusivity.
7. Potential competition may be technologically rather than commercially defined.
8. Merger analysis must account for future ecosystem expansion.
9. Remedies may need to regulate technical architecture as well as contractual behaviour.
10. Competition policy must preserve both network innovation and contestability.

23. Conclusion

Future-oriented antitrust theory for intelligent networks represents an evolution from a predominantly market-centred model toward an analysis of networks, ecosystems, data, algorithms, interoperability and technological architecture.

The principal cases—Microsoft, Google Android, Google/Enel X, Qualcomm, Microsoft/Activision Blizzard, and Booking/eTraveli—illustrate different components of this evolution: platform control, network effects, interoperability, technological dependencies, ecosystem expansion and forward-looking foreclosure analysis.

The most significant future shift is that competition may no longer occur merely between firms selling products. It may occur between intelligent networks that control access to users, data, computing, applications, devices and other networks.

Accordingly, an effective future antitrust framework will have to evalu