Competition Law and Synthetic Economy Ecosystems and Antitrust Concerns

1. Introduction

A synthetic economy ecosystem is an economic environment in which production, exchange, pricing, allocation, and decision-making are increasingly coordinated through artificial intelligence, synthetic data, autonomous agents, algorithmic systems, digital platforms, robotics, cloud infrastructure, digital twins, automated supply chains, and machine-to-machine transactions.

Unlike a conventional digital economy, a synthetic economy may contain autonomous or semi-autonomous economic actors that can:

• generate and exchange synthetic data;
• negotiate contracts algorithmically;
• determine prices dynamically;
• allocate scarce resources;
• select suppliers and customers;
• operate digital marketplaces;
• interact with other AI systems;
• control access to computational infrastructure;
• optimize supply chains without direct human intervention; and
• continuously learn from market information.

Competition law therefore has to address not merely traditional firms competing with one another, but also ecosystem control, data advantages, algorithmic coordination, interoperability, access to computing infrastructure, and autonomous decision-making.

The central competition-law question is:

When does an integrated synthetic ecosystem generate legitimate efficiencies, and when does its architecture become a mechanism for exclusion, coordination, foreclosure, or durable market power?

2. Meaning of a Synthetic Economy Ecosystem

A synthetic economy ecosystem can be understood as a network consisting of several interconnected layers.

A. Computational layer

This includes:

• cloud computing;
• GPUs and AI accelerators;
• data centres;
• edge computing;
• quantum or advanced computing resources;
• model-training infrastructure.

B. Data layer

This includes:

• conventional datasets;
• synthetic datasets;
• behavioural data;
• machine-generated data;
• industrial data;
• real-time transaction data;
• proprietary training datasets.

C. Intelligence layer

This includes:

• foundation models;
• generative AI;
• autonomous agents;
• optimization algorithms;
• recommendation systems;
• predictive systems;
• autonomous pricing systems.

D. Platform layer

Platforms may connect:

• consumers;
• businesses;
• suppliers;
• autonomous agents;
• developers;
• advertisers;
• financial institutions;
• logistics providers.

E. Physical-economic layer

AI systems may ultimately control:

• manufacturing;
• logistics;
• autonomous vehicles;
• energy systems;
• warehouses;
• robotics;
• agriculture;
• healthcare;
• financial services.

This creates the possibility that a single undertaking can occupy several vertically connected layers.

3. Why Synthetic Ecosystems Create Competition Concerns

3.1 Ecosystem market power

Traditional competition analysis often asks whether a firm possesses dominance in a defined relevant market.

Synthetic ecosystems complicate this because market power may arise from control over an entire technological stack rather than one product.

For example:

Cloud infrastructure → AI model → operating system → application store → payment system → consumer data

A firm controlling several layers can potentially disadvantage competitors at one layer by exploiting advantages obtained at another.

4. Relevant Market Definition

Competition authorities may need to examine several overlapping markets.

Possible markets include:

1. cloud-computing services;
2. GPU/AI accelerator infrastructure;
3. foundation models;
4. synthetic-data generation;
5. AI-agent services;
6. AI application markets;
7. digital advertising;
8. digital payments;
9. enterprise software;
10. autonomous logistics;
11. AI-enabled search;
12. data-management services.

The traditional SSNIP test may be insufficient in some synthetic markets because many services have zero monetary prices.

Authorities may therefore examine:

• quality;
• privacy;
• innovation;
• switching costs;
• data access;
• computational resources;
• interoperability;
• user attention;
• ecosystem participation;
• technological dependence.

5. Network Effects

Synthetic ecosystems can generate powerful direct and indirect network effects.

Direct network effect

The value of a platform increases as more users participate.

Indirect network effect

More users attract developers, suppliers and complementary services, which in turn attract additional users.

AI-specific feedback loop

A particularly important phenomenon is:

More users → more data → better model → better service → more users → more data

This can create a self-reinforcing competitive advantage.

Competition law must distinguish between:

• legitimate innovation-based advantages; and
• advantages artificially protected through exclusionary conduct.

6. Data Concentration

Synthetic economies may depend upon enormous quantities of data.

A dominant undertaking might possess:

• proprietary consumer data;
• industrial data;
• synthetic training data;
• transaction histories;
• behavioural information;
• real-time market data.

Data concentration can become an antitrust concern where competitors cannot realistically replicate the relevant dataset.

Possible theories of harm include:

1. Refusal to provide access

A dominant undertaking may deny competitors access to indispensable datasets.

2. Discriminatory access

The platform may provide favourable data access to affiliated companies.

3. Data tying

Users may be required to provide data to obtain another service.

4. Data combination

A firm may combine datasets from several markets to reinforce its position.

5. Data foreclosure

Competitors may be deprived of sufficiently important information to compete effectively.

7. Algorithmic Pricing and Autonomous Coordination

One of the most significant synthetic-economy concerns is algorithmic collusion.

Suppose competing AI pricing systems independently observe market prices and continuously adjust their own prices.

Even without an explicit human agreement, algorithms might:

• recognize competitors' pricing patterns;
• punish deviations;
• stabilize prices;
• reduce discounting;
• coordinate market behaviour.

Competition law therefore has to distinguish between:

Explicit collusion

Humans or firms communicate and agree to coordinate.

Algorithm-assisted collusion

Humans agree to use algorithms to implement the cartel.

Hub-and-spoke coordination

A common platform or algorithm facilitates coordination among competitors.

Autonomous algorithmic coordination

Algorithms independently learn strategies that produce coordinated outcomes.

The last category presents particularly difficult questions concerning intent, attribution, causation and liability.

8. Hub-and-Spoke Risks

A synthetic platform can become the hub connecting numerous competitors.

For example:

Platform AI

↓
Supplier A
Supplier B
Supplier C
Supplier D

If the platform collects competitors' commercially sensitive information and feeds that information into a common pricing or allocation system, the platform could potentially facilitate coordination.

The relevant evidence may include:

• algorithm design;
• data inputs;
• communications;
• contractual provisions;
• model instructions;
• pricing outputs;
• internal documents;
• system logs.

9. Self-Preferencing

A synthetic ecosystem operator may own both:

• the infrastructure/platform; and
• competing downstream services.

It may therefore use its ecosystem to favour its own products.

Examples include:

• preferential search ranking;
• preferential AI-agent recommendations;
• lower API latency;
• superior access to computational resources;
• preferential access to synthetic data;
• better visibility in an AI marketplace.

The concern is particularly acute where users cannot easily determine whether an AI system's recommendation is neutral or commercially influenced.

10. Tying and Bundling

Synthetic ecosystems naturally encourage bundling.

A company may combine:

Cloud + foundation model + AI agent + enterprise software + payment system

Bundling is not inherently unlawful.

Competition concerns arise when the bundle is used to:

• exclude competitors;
• prevent multi-homing;
• increase switching costs;
• foreclose complementary products;
• extend dominance from one market into another.

The analysis generally requires consideration of:

• market power;
• coercion;
• foreclosure;
• efficiencies;
• consumer harm;
• availability of alternatives.

11. Interoperability and API Access

Synthetic ecosystems frequently depend upon APIs.

A dominant ecosystem may restrict:

• API access;
• data portability;
• model interoperability;
• identity portability;
• agent interoperability;
• cross-platform functionality.

Such restrictions may make it difficult for customers to move to competing ecosystems.

Competition authorities may therefore examine:

• technical interoperability;
• open standards;
• API access;
• portability;
• switching costs;
• multi-homing.

12. Switching Costs and Lock-In

Synthetic ecosystems can create extremely high switching costs.

A business may depend upon:

• proprietary AI models;
• customized prompts;
• proprietary APIs;
• accumulated training data;
• cloud architecture;
• automated workflows;
• agent configurations;
• employee training.

Leaving the ecosystem could therefore require substantial expenditure.

Competition concerns arise where the ecosystem deliberately increases those costs to prevent users from moving to competitors.

13. Exclusive Dealing

An ecosystem operator could require:

• exclusive cloud use;
• exclusive AI-model deployment;
• exclusive distribution;
• exclusive payment processing;
• exclusive data-sharing arrangements.

Exclusive agreements can create foreclosure when a dominant firm controls an important proportion of distribution or infrastructure.

The duration and coverage of exclusivity are particularly important.

14. Essential Facilities

Some synthetic-economy infrastructure may become strategically indispensable.

Potential examples include:

• cloud infrastructure;
• critical AI computing capacity;
• specialized data;
• interoperability infrastructure;
• dominant AI marketplaces.

However, not every important resource is automatically an essential facility.

Competition law generally requires a demanding assessment involving factors such as:

1. indispensability;
2. absence of realistic alternatives;
3. ability to provide access;
4. possibility of foreclosure;
5. justification for refusal.

15. Merger Control

Synthetic ecosystems create significant merger concerns.

A major platform may acquire:

• an AI startup;
• synthetic-data provider;
• cloud company;
• robotics firm;
• autonomous-agent developer;
• cybersecurity company;
• model-training company.

Even where the target has low current revenue, it may possess substantial innovation potential.

Competition authorities may therefore examine:

• innovation competition;
• future markets;
• nascent competitors;
• data assets;
• computational resources;
• ecosystem effects;
• vertical integration;
• potential competition.

16. Killer Acquisitions

A dominant ecosystem might acquire a promising technology before it becomes a serious competitor.

The concern is particularly significant where:

• the target has rapidly growing technology;
• revenue is currently low;
• the target possesses unique data or intellectual property;
• the acquirer already controls complementary infrastructure.

The central issue is whether the acquisition removes a significant future competitive constraint.

17. Vertical Foreclosure

Synthetic ecosystems can create vertical chains such as:

Cloud provider → AI model → AI agent → marketplace → consumer

If one undertaking controls several stages, it might disadvantage rivals at a downstream or upstream level.

Potential practices include:

• discriminatory access;
• margin compression;
• refusal to interoperate;
• preferential allocation;
• exclusive contracts;
• technical degradation;
• tying.

18. Predatory Pricing

AI services can have unusual cost structures.

Once a model has been developed, the marginal cost of serving another customer may be comparatively low, while infrastructure costs remain substantial.

A dominant undertaking could potentially subsidize an AI service using profits from another market.

Competition authorities may therefore examine:

• below-cost pricing;
• cross-subsidization;
• recoupment;
• exclusionary purpose/effect;
• innovation effects.

19. Algorithmic Discrimination

AI systems can facilitate discriminatory treatment of:

• suppliers;
• customers;
• advertisers;
• developers;
• distributors.

For example, an ecosystem may automatically provide different:

• prices;
• rankings;
• commissions;
• access conditions;
• visibility;
• computational resources.

The competition-law issue becomes particularly important where discrimination is based on a firm's competitive relationship with the platform.

20. Synthetic Data and Competition

Synthetic data can reduce some traditional barriers to entry because companies may generate artificial datasets rather than obtain large quantities of real-world data.

However, synthetic data may also increase concentration where:

• only a few firms possess the models necessary to generate high-quality synthetic data;
• model outputs depend on proprietary real-world datasets;
• synthetic data is incompatible across ecosystems;
• dominant firms control validation systems.

Thus, synthetic data can simultaneously be:

a competitive substitute for scarce data and a potential source of new market power.

21. Competition Between Autonomous Agents

Future markets may involve AI agents negotiating directly.

For example:

Buyer Agent A ↔ Seller Agent B

Buyer Agent C ↔ Seller Agent D

The agents may independently:

• negotiate prices;
• compare suppliers;
• select delivery arrangements;
• negotiate contracts.

Competition law will need to determine whether conduct generated by autonomous agents can be attributed to their human or corporate operators.

The fact that an algorithm makes the immediate decision does not necessarily eliminate responsibility under competition law.

22. Six Important Case Laws

The following cases do not all concern a fully developed "synthetic economy" in the modern sense. They provide legal principles that can be applied to synthetic ecosystems, AI platforms, data markets, algorithmic coordination and digital infrastructure.

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

Facts

Microsoft was found liable for anticompetitive conduct involving its Windows operating-system monopoly and the distribution of Internet Explorer.

Principle

The case is important for understanding how a dominant firm can use control over one technological layer to disadvantage competitors in an adjacent market.

Relevance to synthetic ecosystems

A synthetic ecosystem operator controlling:

• cloud infrastructure;
• operating systems;
• AI models;
• application distribution;

could potentially use control at one layer to restrict competition at another.

The case illustrates the importance of examining ecosystem leverage and technological integration.

Case 2: United States v. Google LLC — Search and Search Advertising

The U.S. search-distribution litigation involving Google concerns alleged exclusionary agreements relating to search distribution and default positions.

Principle

Control over important distribution channels can reinforce market position even when consumers technically have alternative products.

Synthetic-economy relevance

An AI ecosystem could potentially become a major distribution channel for:

• AI agents;
• search;
• applications;
• recommendations;
• autonomous purchasing.

If the ecosystem determines which competing services users encounter, default placement and distribution arrangements can become important competition-law issues.

Case 3: European Commission v. Google — Google Shopping (2017)

Facts

The European Commission found that Google had abused a dominant position by systematically giving prominent placement to its own comparison-shopping service while applying demotion mechanisms to competing comparison-shopping services.

Principle

A dominant digital platform can potentially distort downstream competition through self-preferencing.

Synthetic-economy relevance

A dominant AI marketplace could potentially rank its own:

• AI applications;
• agents;
• shopping services;
• data services;

more favourably than competing services.

The case therefore provides an important analytical framework for AI recommendation and self-preferencing concerns.

Case 4: Google Android — European Commission (2018)

Facts

The European Commission addressed Google's practices concerning the Android ecosystem, including tying arrangements involving Google applications and restrictions affecting competing services.

Principle

A dominant undertaking may face competition-law scrutiny when it uses contractual or technological arrangements to extend market power across connected markets.

Synthetic-economy relevance

The same analytical issue can arise where a dominant ecosystem bundles:

cloud infrastructure + AI model + operating system + AI assistant + application marketplace.

The case demonstrates why competition analysis must consider ecosystem-wide effects rather than examining each product in isolation.

Case 5: European Commission v. Amazon — Marketplace / Seller Data

European competition authorities investigated Amazon's use of non-public marketplace seller data and examined whether the platform used information generated by independent sellers to compete with those sellers.

Principle

A platform occupying an intermediary role may create competition concerns when it simultaneously:

1. operates the marketplace; and
2. competes with businesses using that marketplace.

Synthetic-economy relevance

A synthetic ecosystem could receive enormous amounts of commercially sensitive information from:

• AI developers;
• suppliers;
• retailers;
• advertisers;
• autonomous agents.

If the ecosystem uses that information to compete against the firms supplying it, information asymmetry and platform self-preferencing become significant issues.

Case 6: FTC v. Facebook, Inc. (Meta)

The U.S. Federal Trade Commission's litigation concerning Facebook addresses allegations relating to maintenance of monopoly power through conduct involving acquisitions and platform policies.

Principle

Digital markets can exhibit:

• network effects;
• high switching costs;
• data advantages;
• entry barriers;
• ecosystem effects.

Synthetic-economy relevance

These characteristics can become even stronger in AI ecosystems because users may accumulate:

• customized models;
• workflows;
• agent histories;
• data;
• integrations;
• proprietary configurations.

Consequently, switching may become progressively more difficult as the ecosystem expands.

Case 7: United States v. Apple Inc. (2024)

The U.S. Department of Justice challenged Apple's alleged maintenance of monopoly power in smartphone markets through restrictions affecting developers and competing technologies.

Principle

Competition analysis may examine how restrictions imposed at one technological layer affect competition in complementary or adjacent markets.

Synthetic-economy relevance

An AI ecosystem operator could potentially impose restrictions affecting:

• interoperability;
• cloud access;
• payment systems;
• AI assistants;
• application distribution;
• competing AI models.

The case is therefore relevant to the broader issue of ecosystem governance and technological foreclosure.

Case 8: Epic Games, Inc. v. Apple Inc.

Facts

Epic challenged Apple's App Store practices, particularly restrictions concerning distribution and payment systems.

Principle

The case illustrates the importance of examining the relationship between:

• platform control;
• access conditions;
• payment systems;
• developers;
• alternative distribution mechanisms.

Synthetic-economy relevance

A future AI-agent marketplace could similarly control:

• access to users;
• payment processing;
• ranking;
• API access;
• distribution;
• commissions.

This makes platform governance a central competition-law issue.

23. Consolidated Case-Law Principles

24. Competition-Law Theories of Harm

Synthetic ecosystems can potentially generate several traditional theories of harm.

Horizontal concerns

• price fixing;
• algorithmic collusion;
• market allocation;
• coordinated conduct;
• information exchange.

Vertical concerns

• tying;
• bundling;
• exclusive dealing;
• discriminatory access;
• margin squeeze;
• foreclosure.

Platform concerns

• self-preferencing;
• discriminatory ranking;
• interoperability restrictions;
• access restrictions;
• excessive commissions.

Data concerns

• data foreclosure;
• discriminatory data access;
• data combination;
• exclusive data arrangements.

Merger concerns

• killer acquisitions;
• vertical integration;
• innovation suppression;
• acquisition of nascent competitors.

25. Role of Competition Authorities

Competition authorities may increasingly need to examine not merely the observable output of an AI ecosystem but its architecture.

Relevant evidence may include:

• source-code documentation;
• algorithmic instructions;
• model-training methodology;
• API rules;
• contractual restrictions;
• internal communications;
• system logs;
• pricing histories;
• ranking mechanisms;
• data-access protocols.

This creates a movement from traditional market investigation toward technological and algorithmic competition analysis.

26. Possible Regulatory Remedies

Where competition concerns are established, authorities may consider remedies such as:

Structural remedies

• divestiture;
• separation of business units;
• prohibition of certain acquisitions.

Behavioural remedies

• non-discrimination;
• interoperability;
• API access;
• data portability;
• transparency requirements.

Platform remedies

• restrictions on self-preferencing;
• fair ranking rules;
• independent dispute mechanisms.

Merger remedies

• licensing commitments;
• access commitments;
• divestiture;
• firewall arrangements.

Algorithmic remedies

• independent auditing;
• logging requirements;
• explainability obligations;
• monitoring of pricing algorithms;
• restrictions on sharing competitively sensitive information.

27. Challenges for Competition Law

27.1 Attribution

Who is responsible when an autonomous AI agent makes the decision?

Possible candidates include:

• developer;
• platform;
• deploying company;
• operator;
• human decision-maker.

27.2 Explainability

Competition authorities may need to understand why an algorithm:

• increased a price;
• excluded a supplier;
• ranked a competitor lower;
• refused access.

27.3 Speed

AI markets can change faster than conventional enforcement procedures.

27.4 Market definition

Synthetic markets may converge rapidly, making conventional market boundaries unstable.

27.5 Innovation

Aggressive intervention may potentially interfere with legitimate innovation, while insufficient intervention may allow ecosystems to become entrenched.

27.6 Cross-border enforcement

Synthetic ecosystems operate globally, requiring coordination between:

• EU authorities;
• U.S. authorities;
• Asian competition authorities;
• national regulators.

28. A Proposed Analytical Framework

Competition authorities examining a synthetic ecosystem can proceed through the following sequence:

Step 1 — Identify the ecosystem

↓

Step 2 — Map technological layers

↓

Step 3 — Identify relevant markets

↓

Step 4 — Determine market power

↓

Step 5 — Identify network effects and switching costs

↓

Step 6 — Examine data and computational advantages

↓

Step 7 — Examine exclusionary or coordinating conduct

↓

Step 8 — Assess effects on competitors, innovation and consumers

↓

Step 9 — Consider efficiency justifications

↓

Step 10 — Select proportionate remedies

29. Difference Between Legitimate Ecosystem Integration and Anticompetitive Conduct

The existence of an integrated ecosystem does not itself establish an antitrust violation. Competition law generally requires an assessment of market power, conduct, effects, and applicable legal standards.

30. Conclusion

The synthetic economy represents a development of competition law beyond conventional digital-platform analysis. Its distinctive feature is the convergence of AI, synthetic data, autonomous agents, computing infrastructure, platforms and physical economic activity.

The principal antitrust concerns are likely to involve:

1. ecosystem dominance;
2. data concentration;
3. computational-resource concentration;
4. self-preferencing;
5. algorithmic coordination;
6. tying and bundling;
7. interoperability restrictions;
8. switching costs and lock-in;
9. exclusive dealing;
10. vertical foreclosure;
11. killer acquisitions; and
12. control over AI distribution channels.

The Microsoft, Google, Amazon, Meta, Apple and Epic litigation demonstrates that many of these concerns have identifiable precedents in earlier technology markets. The novel challenge is that synthetic ecosystems may combine data, algorithms, infrastructure and autonomous decision-making into one continuously adapting competitive structure.