Competition Law and Future Competition Frameworks for Autonomous Global Networks

Introduction

Autonomous global networks are interconnected digital, physical, financial, logistics, energy, communications, AI, and platform systems capable of making or implementing decisions with limited direct human intervention. Examples include AI-driven marketplaces, autonomous logistics networks, algorithmic trading systems, cloud ecosystems, smart grids, autonomous vehicles, digital identity networks, blockchain-based markets, and machine-to-machine commerce.

Competition law was traditionally designed around human-controlled firms operating within identifiable geographic and product markets. Autonomous global networks challenge these assumptions because decision-making may be distributed among algorithms, AI agents, platforms, data providers, cloud infrastructure, and automated protocols operating across multiple jurisdictions.

Existing digital-market enforcement already demonstrates the direction of travel. The EU's Digital Markets Act (DMA), for example, regulates designated gatekeepers and currently covers major services supplied by Alphabet, Amazon, Apple, Booking, ByteDance, Meta and Microsoft. The Commission has also begun examining cloud computing as a possible gatekeeper environment, emphasizing interoperability, contractual conditions, switching costs and ecosystem effects.

The future framework therefore needs to combine traditional antitrust principles with ex-ante digital regulation, algorithmic accountability, interoperability, data access and international cooperation.

I. Meaning and Characteristics of Autonomous Global Networks

An autonomous global network has several distinctive characteristics.

1. Machine-mediated decision-making

Prices, rankings, allocation of resources, routing, credit decisions, advertising and matching may be determined automatically.

2. Distributed control

No single company may control the entire economic process. A transaction might simultaneously involve:

• an AI agent;
• cloud infrastructure;
• an operating system;
• an API;
• a payment network;
• a data provider;
• a marketplace; and
• an automated logistics provider.

3. Network effects

The value of the network may increase as more users and businesses join it.

4. Data feedback loops

More users generate more data, which improves algorithms, which attracts more users, creating a reinforcing competitive advantage.

The FTC has specifically recognized that digital-platform remedies may need to account for network effects and data feedback loops, rather than merely restoring the market conditions existing before the infringement.

5. Cross-border operation

An autonomous network may make decisions simultaneously affecting users in dozens of jurisdictions.

6. Continuous optimization

Unlike traditional businesses, autonomous systems can change prices, rankings, supply allocation and contractual conditions continuously.

II. Why Traditional Competition Law May Become Insufficient

Traditional competition law generally asks:

1. What is the relevant market?
2. Who are the competitors?
3. Does an undertaking possess market power?
4. What conduct has occurred?
5. What is its effect on competition?

Autonomous networks complicate each question.

A. Market definition

A network may operate simultaneously across:

• search;
• advertising;
• cloud computing;
• payments;
• AI;
• data;
• logistics; and
• marketplace services.

Consequently, defining the market exclusively according to a conventional product category may underestimate competitive constraints.

B. Identification of the undertaking

The economically relevant actor may be:

• the platform owner;
• the AI developer;
• the infrastructure provider;
• the network participants collectively;
• the protocol developer; or
• multiple entities operating through contractual or technical arrangements.

C. Attribution of conduct

If an AI system autonomously changes prices, the legal question becomes whether the conduct is attributable to the firm operating the system.

D. Temporal difficulty

Competition authorities traditionally investigate past conduct. Autonomous markets may require intervention before a harmful structure becomes entrenched.

III. Core Future Competition Framework

A future framework should contain at least ten interconnected components.

1. Algorithmic Competition Regulation

Competition authorities should examine algorithms as potential instruments of:

• exclusion;
• discrimination;
• coordination;
• self-preferencing;
• personalized pricing;
• market allocation;
• foreclosure; and
• entry deterrence.

The U.S. FTC and DOJ have already recognized competition risks arising from AI systems and the AI ecosystem. An international joint statement emphasized maintaining competition throughout the AI ecosystem.

Future rule

Large autonomous systems could be required to maintain:

• algorithmic audit trails;
• decision logs;
• model-change records;
• pricing histories;
• access records; and
• records showing the basis for material competitive decisions.

IV. Autonomous Algorithmic Collusion

One of the most important future issues is machine-to-machine coordination.

Suppose several competing AI agents independently learn that maintaining high prices maximizes their respective profits.

Even without a traditional human agreement, parallel algorithmic conduct could produce coordinated outcomes.

The legal framework therefore needs to distinguish:

Legitimate parallel conduct

Independent algorithms independently respond to market conditions.

Potentially unlawful coordination

Algorithms are deliberately designed or configured to:

• communicate competitively sensitive information;
• coordinate prices;
• divide customers;
• stabilize market shares; or
• punish deviation.

The FTC and DOJ have previously examined how algorithms can facilitate new forms of collusion, including pricing-algorithm risks.

V. Interoperability as a Competition Remedy

Autonomous networks may become extremely difficult to challenge because users become dependent upon an ecosystem.

Future competition law may therefore require:

• API interoperability;
• data portability;
• identity portability;
• messaging interoperability;
• cloud portability;
• payment interoperability;
• switching tools; and
• machine-readable access standards.

The DMA already illustrates this movement from purely ex-post antitrust toward ex-ante obligations for powerful gatekeepers.

VI. Data as Competitive Infrastructure

Data may become comparable to essential competitive infrastructure in some autonomous markets.

A dominant network may possess:

• transaction data;
• behavioral data;
• search data;
• training data;
• location data;
• industrial data; and
• real-time market information.

Future competition frameworks should therefore consider:

Data access

Whether competitors can obtain necessary data under fair conditions.

Data portability

Whether users can transfer their data.

Data interoperability

Whether competing systems can meaningfully use transferred information.

Data discrimination

Whether the dominant network provides better data access to itself than to competitors.

The EU's current DMA work on Google Search illustrates this trend: the Commission has adopted measures concerning sharing anonymised search data with eligible competing search services, including AI chatbots with search functionality.

VII. Autonomous Self-Preferencing

An autonomous network could continuously rank its own products or services above competitors.

For example, an AI marketplace might automatically determine:

"My platform's product is the most relevant."

If the algorithm systematically produces this outcome because the platform owns the underlying infrastructure, competition concerns arise.

The problem is particularly significant where the platform controls both:

infrastructure + marketplace + ranking algorithm.

The European Commission's 2026 DMA enforcement against Google illustrates this emerging approach. The Commission found Google in breach concerning self-preferencing in Search and restrictions on steering users toward alternative purchasing channels.

VIII. Autonomous Gatekeepers

Traditional dominance analysis may not adequately address networks that become unavoidable gateways.

A future framework could therefore classify certain networks as systemic digital gatekeepers when they possess combinations of:

• enormous user bases;
• durable network effects;
• significant data advantages;
• high switching costs;
• ecosystem control;
• interoperability control;
• strategic infrastructure;
• AI capabilities; and
• cross-border reach.

The EU DMA already adopts a gatekeeper-oriented model, with designation based on substantial internal-market impact, an important gateway role and a strong, entrenched position.

IX. Autonomous Mergers and Acquisitions

Future merger control must look beyond conventional market shares.

An acquisition of a seemingly small company may give an autonomous network:

• unique training data;
• a critical API;
• an AI model;
• a user community;
• an interoperability protocol;
• a cloud technology; or
• a strategic algorithm.

Therefore, merger analysis should consider innovation assets and future competitive potential, not merely present revenues.

This is particularly important for acquisitions of start-ups whose current turnover is low but whose technology may become an important competitive constraint.

X. Algorithmic Merger Simulation

Future merger review could employ computational tools to model:

• network effects;
• user switching;
• AI learning;
• pricing responses;
• entry barriers;
• interoperability;
• data accumulation; and
• innovation incentives.

Instead of asking only:

"What is the present market share?"

the authority may also ask:

"How does the network evolve if the transaction occurs?"

This would make merger control more dynamic.

XI. Autonomous Networks and Essential Facilities

A network can become economically indispensable when competitors cannot realistically operate without access to it.

Potential examples include:

• cloud infrastructure;
• payment rails;
• app stores;
• dominant AI infrastructure;
• telecom networks;
• identity infrastructure;
• logistics platforms;
• digital advertising exchanges.

Future rules may therefore develop an autonomous-network essential-facilities doctrine, requiring access where:

1. the facility is strategically indispensable;
2. duplication is economically or technically impracticable;
3. refusal substantially impairs competition; and
4. access can be provided without disproportionate harm to legitimate security or innovation interests.

XII. Case Laws

The following cases provide important foundations for developing future competition rules for autonomous networks.

1. United States v. Microsoft Corp. (2001)

The Microsoft litigation demonstrated the importance of:

• platform power;
• network effects;
• interoperability;
• control over interfaces;
• exclusion of competing technologies; and
• leveraging dominance from one technological layer into another.

Relevance

Autonomous networks may similarly control an underlying technological layer and use that control to disadvantage competing systems.

Principle: Control of a technological platform can have consequences beyond the immediate product market.

2. Google Search (Shopping) — European Commission / Google (2017)

The European Commission found Google had abused its dominant position by systematically giving prominent placement to its comparison-shopping service while demoting competing services.

Relevance

This is highly relevant to autonomous networks because AI-driven ranking systems may automatically determine which competitors are visible.

The future question becomes whether an autonomous ranking system can systematically privilege its owner's ecosystem.

Principle: Algorithmic ranking can become a competition concern when used to disadvantage competing services.

3. Google Android — European Commission (2018)

The Android case concerned restrictions involving mobile-device manufacturers and mobile operating systems, including practices related to search and app ecosystems.

Relevance

Autonomous ecosystems can combine:

• operating systems;
• app stores;
• search;
• advertising;
• data; and
• hardware.

The case therefore demonstrates the importance of analysing ecosystem leverage rather than isolated products.

4. United States v. Google LLC — Search and Advertising Litigation

The U.S. Google proceedings illustrate modern concerns concerning the acquisition and maintenance of market power through control over digital infrastructure and distribution.

Relevance

Autonomous global networks may similarly create durable competitive advantages through:

• default arrangements;
• distribution;
• data;
• scale;
• technological integration; and
• switching costs.

The broader lesson is that competition analysis increasingly has to examine the architecture through which markets operate.

5. FTC v. Facebook, Inc. / Meta

The FTC's case against Facebook alleges that Meta maintained its personal-social-networking monopoly through a course of conduct including acquisitions of Instagram and WhatsApp and restrictive conditions imposed on developers. The case remained pending according to the FTC's December 2025 update.

Relevance

Autonomous networks can use acquisitions to eliminate potential future competitors before they become significant constraints.

This supports closer examination of:

• nascent competition;
• innovation competition;
• ecosystem expansion; and
• acquisitions of strategically valuable technologies.

6. Epic Games v. Google

The Epic Games litigation concerning Google's app ecosystem illustrates competition concerns involving app distribution, payment systems and platform restrictions.

The FTC has emphasized that remedies in digital-platform cases may need to account for network effects and data advantages and restore conditions for future competition rather than simply reversing an isolated transaction.

Relevance

Autonomous global networks may similarly use technical architecture to make alternative suppliers less attractive or accessible.

7. Apple App Store / DMA Enforcement

The European Commission found Apple in breach of the DMA's anti-steering obligation in April 2025 and imposed a €500 million fine.

Relevance

This demonstrates the transition from traditional competition-law enforcement toward direct behavioural obligations on designated gatekeepers.

For autonomous networks, comparable rules could regulate:

• steering;
• interoperability;
• data access;
• ranking;
• switching;
• contractual restrictions; and
• self-preferencing.

8. Booking.com — Digital Markets Act

Booking was designated as a DMA gatekeeper for its online intermediation service. The Commission required compliance with obligations concerning, among other matters, business users' ability to offer better prices elsewhere and access relevant data.

Relevance

This is important for autonomous networks because intermediary platforms can control both:

access to consumers + information about transactions.

That combination can create significant competitive advantages.

XIII. Proposed Future Regulatory Architecture

A comprehensive autonomous-network competition framework could be organized into six layers.

This produces a shift from conventional firm-centred competition law toward network-centred competition governance.

XIV. Cross-Border Enforcement

Autonomous networks create jurisdictional problems because:

• the algorithm may be developed in one country;
• data may be stored in another;
• infrastructure may operate in a third;
• users may be located worldwide; and
• the competitive harm may arise elsewhere.

Future cooperation should therefore include:

1. International information sharing

Competition authorities should develop mechanisms for exchanging evidence about algorithmic systems.

2. Coordinated investigations

Multiple jurisdictions may investigate the same network simultaneously.

3. Common terminology

International authorities need consistent concepts concerning:

• algorithmic collusion;
• AI market power;
• data dominance;
• interoperability;
• network effects; and
• autonomous decision-making.

4. Coordinated remedies

Conflicting national remedies could otherwise fragment the network.

XV. Competition Audits of Autonomous Networks

A future regulatory system could require systemic networks to conduct periodic competition impact assessments.

An audit could examine:

1. market concentration;
2. algorithmic discrimination;
3. self-preferencing;
4. access restrictions;
5. switching costs;
6. data accumulation;
7. interoperability;
8. acquisition strategy;
9. algorithmic coordination risks; and
10. potential foreclosure of emerging competitors.

This would shift enforcement from a purely reactive model to a continuous monitoring model.

XVI. Human Oversight and Accountability

Autonomous decision-making cannot mean autonomous legal responsibility.

A fundamental rule should be:

Autonomy of operation should not eliminate accountability of the economic operator.

Businesses deploying autonomous systems should remain responsible for competition-law compliance where they design, control, deploy or materially benefit from the system.

This is particularly important where companies argue that an unlawful outcome was generated by an AI system without direct human instruction.

XVII. Competition and AI Foundation Models

AI foundation models may become critical infrastructure for autonomous markets.

Competition authorities may therefore examine:

• exclusive cloud arrangements;
• access to training data;
• model distribution;
• compute access;
• model interoperability;
• acquisitions of AI start-ups;
• exclusive partnerships;
• API restrictions;
• preferential treatment of affiliated models; and
• tying AI services to cloud infrastructure.

The EU's continuing work concerning cloud computing and AI-related ecosystem effects demonstrates the growing importance of these questions.

XVIII. Remedies for Autonomous Global Networks

Traditional fines may be insufficient where network effects make the competitive harm persistent.

Possible remedies include:

Structural remedies

• divestiture;
• separation of infrastructure and marketplace functions;
• functional separation.

Behavioural remedies

• non-discrimination;
• interoperability;
• access obligations;
• anti-self-preferencing requirements.

Technical remedies

• open APIs;
• data portability;
• interoperable protocols;
• algorithmic transparency;
• independent audits.

Dynamic remedies

Authorities could periodically reassess whether a remedy remains effective as the network evolves.

XIX. Key Challenges

1. Innovation versus regulation

Excessive regulation may discourage technological innovation.

2. Transparency versus trade secrets

Authorities need sufficient information without unnecessarily disclosing proprietary algorithms.

3. Attribution

Determining responsibility for autonomous decisions may be difficult.

4. Global inconsistency

Different jurisdictions may impose conflicting requirements.

5. Algorithmic complexity

Competition authorities may lack the technical resources to investigate advanced AI systems.

6. Rapid technological change

A regulatory definition can become obsolete quickly.

7. False positives

Authorities must distinguish legitimate efficiency from exclusionary conduct.

XX. Future Principles

The emerging framework can be summarized through ten principles:

1. Algorithmic neutrality — automated systems should not systematically distort competitive access.
2. Interoperability — dominant networks should not unnecessarily prevent technical compatibility.
3. Data fairness — strategic data advantages should be examined for foreclosure effects.
4. Contestability — users and businesses should have realistic alternatives.
5. Switchability — switching barriers should be monitored.
6. Auditability — significant autonomous systems should maintain sufficient records.
7. Accountability — autonomous operation should not eliminate legal responsibility.
8. Innovation protection — enforcement should preserve incentives for technological development.
9. Cross-border cooperation — global networks require coordinated enforcement.
10. Dynamic supervision — competition analysis must account for rapidly evolving network structures.

Conclusion

The future of competition law for autonomous global networks is likely to move from a firm-centric, market-share-oriented model toward a network-centric and technologically informed model.

The central competition questions will increasingly concern:

• who controls the network;
• who controls the data;
• who controls interoperability;
• who controls the algorithm;
• whether competitors can enter;
• whether users can switch;
• whether autonomous systems can coordinate; and
• whether infrastructure owners can leverage control across connected markets.

The development of the EU DMA, including its gatekeeper regime and recent enforcement against Google and Apple, demonstrates the movement toward ex-ante regulation alongside conventional antitrust. Meanwhile, U.S. digital-platform litigation and international AI competition initiatives illustrate the increasing importance of network effects, data feedback loops and AI-specific competition risks.

Accordingly, a mature future framework should combine antitrust law + ex-ante gatekeeper regulation + algorithmic auditing + data governance + interoperability + merger scrutiny + international cooperation. Such a framework would allow autonomous global networks to deliver efficiency and innovation while preserving the fundamental competitive conditions necessary for new firms, technologies and business models to emerge.