Competition Law and Competition Implications of Synthetic Economies

1. Introduction

Synthetic economies are economic systems in which goods, services, assets, content, data, experiences, or even market participants are substantially created, simulated, generated, or optimized through artificial or digital technologies rather than through conventional production alone.

The concept may cover:

AI-generated products and services;

synthetic data;

virtual goods and digital assets;

AI-generated content;

simulated markets and environments;

synthetic financial products;

automated agents;

digital twins;

virtual worlds;

AI-generated software;

algorithmically generated designs;

artificial datasets used for training AI; and

automated economic decision-making.

From a competition-law perspective, synthetic economies create a fundamental question:

What happens to competition when the inputs, products, competitors and even demand signals of a market can increasingly be artificially generated?

Traditional competition law generally assumes that firms compete using relatively identifiable resources such as:

labour;

capital;

physical assets;

intellectual property;

distribution;

customer relationships.

Synthetic economies can change this structure because an AI system may generate enormous quantities of content, products, code, designs or data at very low marginal cost.

This can produce greater competition and innovation, but it can also create new forms of concentration and market power.

2. Meaning of Synthetic Economy

A synthetic economy can be represented as:

Digital infrastructure + AI/automation + synthetic inputs + algorithmic production + digital distribution = Synthetic Economy

For example, an AI company may use:

computing infrastructure;

foundation models;

synthetic training data;

automated agents;

algorithmic production; and

digital distribution

to produce thousands of products or services without a corresponding increase in human labour.

The competitive implications are substantial because the traditional relationship between firm size and productive capacity may change.

3. Difference Between Traditional and Synthetic Economies

4. Competition-Law Importance

Synthetic economies can affect virtually every major area of competition law:

1. Market definition

What is the relevant market when AI-generated and human-generated products compete?

2. Market power

Can control over AI infrastructure create dominance?

3. Mergers

Should acquisitions of AI startups, datasets or computing capacity be treated as potentially strategic acquisitions?

4. Abuse of dominance

Can dominant AI firms discriminate against competing applications?

5. Data concentration

Can control over training data create competitive advantages?

6. Algorithms

Can competing AI systems coordinate prices or strategies?

7. Vertical integration

Can an AI infrastructure provider favour its own downstream AI products?

8. Innovation

Can concentration reduce future technological competition?

5. Synthetic Inputs and Competition

One of the most important features of synthetic economies is that inputs themselves can be artificially generated.

Examples include:

synthetic datasets;

synthetic images;

synthetic text;

synthetic voices;

synthetic video;

synthetic financial data;

simulated consumer behaviour.

This can potentially reduce traditional barriers to entry.

A new entrant might not need access to enormous quantities of naturally occurring data if sufficiently effective synthetic data can be generated.

However, the company controlling the technology used to generate that synthetic input may itself become a bottleneck.

Thus:

Synthetic production can reduce dependence on traditional inputs while increasing dependence on computational infrastructure and AI models.

6. Computing Power as a Competition Variable

Synthetic economies are heavily dependent upon:

GPUs;

specialised chips;

cloud computing;

data centres;

networking;

energy;

model-serving infrastructure.

If a small number of companies control these resources, competition may become concentrated upstream.

A downstream AI company may therefore face:

AI application → cloud provider → computing infrastructure → semiconductor supplier

The competitive problem is that dominance at one level can potentially be leveraged into another level.

7. AI Infrastructure Concentration

Suppose one company controls a large portion of:

AI chips;

cloud infrastructure;

foundation models;

application distribution.

It could potentially give its own downstream services:

preferential access;

lower prices;

better computing capacity;

technical integration;

early access to new hardware.

Competitors could consequently face higher costs.

This is a classic vertical foreclosure concern.

8. Synthetic Economies and Network Effects

Synthetic economies can exhibit strong network effects.

For example:

More users → more interactions → more data → better AI → more users

This can create a self-reinforcing competitive advantage.

The problem becomes more serious where the platform also controls:

distribution;

data;

computing;

application stores;

advertising;

payment systems.

A firm can therefore develop an ecosystem in which competitors find it increasingly difficult to enter.

9. Data as a Competitive Asset

Data may be particularly important in synthetic economies.

Competitive advantages may arise from:

proprietary datasets;

user interaction data;

behavioural information;

industrial datasets;

transaction data;

model-generated data.

Synthetic data can reduce scarcity, but not necessarily eliminate it.

Certain forms of real-world data remain difficult to reproduce because they reflect:

actual consumer behaviour;

unique transactions;

proprietary operations;

real-world outcomes.

Therefore:

Synthetic data may supplement real data without necessarily making real data competitively irrelevant.

10. Synthetic Data and Market Entry

Synthetic data can potentially lower entry barriers.

A startup could generate training data without obtaining millions of real-world observations.

This may:

reduce costs;

accelerate experimentation;

improve privacy;

enable niche applications.

But if the dominant company owns the best synthetic-data generation model, it may control the process of creating the substitute input.

Therefore, competition authorities should examine both:

data ownership and data-generation capability.

11. Artificial Scarcity

Synthetic economies can also create an unusual phenomenon:

Artificial abundance combined with artificial scarcity.

AI can generate unlimited digital content, but platforms may create scarcity through:

exclusive access;

licensing;

digital rights;

proprietary standards;

subscription systems;

platform restrictions.

A platform may therefore have enormous production capacity while restricting access to the infrastructure needed to use that capacity.

This can create competition concerns involving:

exclusion;

tying;

discriminatory access;

interoperability;

licensing.

12. Synthetic Economies and Zero Marginal Cost

AI-generated digital products can have extremely low marginal production costs.

For example, after developing an AI system, generating:

another image;

another piece of text;

another software component;

another virtual object

may cost very little relative to the initial investment.

This can encourage:

aggressive pricing;

rapid scaling;

bundling;

product expansion.

However, low marginal cost does not necessarily mean low prices.

A dominant firm may use its market power to maintain high prices despite low production costs.

13. Predatory Pricing Concerns

Synthetic economies may make it easier for large firms to sustain aggressive pricing.

A dominant AI firm could potentially:

provide services free of charge;

bundle AI with another product;

subsidise one market using revenue from another;

temporarily price below cost;

use enormous infrastructure advantages to outlast smaller competitors.

Competition authorities may therefore need to consider the relationship between:

marginal cost;

average cost;

infrastructure investment;

cross-subsidisation;

network effects.

14. Algorithms and Tacit Coordination

Synthetic economies may increase the possibility of algorithmic coordination.

If competing AI systems independently monitor:

prices;

inventory;

demand;

competitors;

market conditions;

they may adjust behaviour rapidly.

Even without a traditional human agreement, algorithms could potentially produce parallel outcomes.

Competition law therefore faces the question:

When does algorithmic adaptation become unlawful coordination?

The legal answer depends upon the relevant jurisdiction and evidence of concerted conduct.

15. AI and Algorithmic Collusion

Algorithms may:

observe competitors;

predict their responses;

modify prices;

learn from market reactions;

repeat the process.

This can create highly stable pricing patterns.

Traditional cartel law generally focuses on communication or agreement. Synthetic economies may require greater attention to:

algorithm design;

information exchange;

common software providers;

contractual instructions;

human involvement;

intentional coordination.

16. Synthetic Agents as Economic Actors

A particularly novel issue is the emergence of AI agents capable of:

negotiating;

purchasing;

selling;

bidding;

advertising;

trading;

selecting suppliers.

An AI agent may potentially conduct thousands of transactions faster than a human.

This raises questions about:

attribution;

agency;

responsibility;

coordination;

market manipulation;

competition liability.

The fundamental principle remains:

The use of an algorithm does not automatically remove the underlying firm's responsibility for anti-competitive conduct.

17. Autonomous Agents and Market Coordination

Imagine ten companies using AI purchasing agents.

Each agent:

monitors prices;

negotiates contracts;

switches suppliers;

predicts competitors.

The market could become highly responsive.

But if all agents use similar objectives and information, they might produce similar market outcomes.

Competition authorities may therefore examine whether firms:

intentionally designed algorithms to coordinate;

exchanged sensitive information;

used a common algorithm provider;

instructed algorithms to follow anti-competitive strategies.

18. Synthetic Economies and Merger Control

Synthetic economies create new types of acquisitions.

A large company may acquire:

an AI startup;

a foundation-model developer;

a dataset company;

a cloud-AI business;

a chip-design company;

an AI distribution platform.

The acquisition may appear small in current revenue terms but could have significant future competitive importance.

This raises the issue of nascent competition.

A startup with little current revenue may nevertheless represent an important future competitive constraint.

19. Killer Acquisitions

A dominant company may acquire a small AI startup before it becomes a significant competitor.

Competition authorities may therefore examine:

innovation pipelines;

technology capabilities;

talent;

patents;

datasets;

developer ecosystems;

potential future products.

The relevant question is not simply:

"How much revenue does the startup currently generate?"

but also:

"What competitive constraint could the startup provide in the future?"

20. Case Law 1: United States v. Google LLC — Search and Search Advertising

The U.S. Department of Justice brought an antitrust case against Google concerning its search and search-advertising businesses.

The case examined Google's agreements and conduct concerning distribution of search services and whether those practices maintained monopoly power.

Relevance to synthetic economies

The case illustrates the importance of:

distribution;

default positions;

network effects;

scale;

data advantages;

barriers to entry.

These factors can become even more important in AI markets.

Principle

Control over distribution can reinforce market power even when alternative technologies technically exist.

21. Case Law 2: United States v. Microsoft Corp. (2001)

In United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001), the court considered Microsoft's conduct concerning the Windows operating-system monopoly and Internet Explorer.

The case is highly relevant to synthetic economies because it demonstrates how a dominant firm can use control over one platform to affect competition in an adjacent technological market.

Competition implications

The case addressed:

platform power;

exclusionary conduct;

technological integration;

barriers to entry;

leveraging.

Synthetic-economy lesson

A dominant AI platform could potentially use control over:

cloud infrastructure;

operating systems;

app stores;

foundation models

to disadvantage competing AI products.

22. Case Law 3: Google Android — European Commission

The European Commission's Android decision concerned Google's practices involving the Android mobile operating system.

The Commission found concerns relating to:

tying;

search and browser distribution;

restrictions affecting competing mobile operating systems.

Synthetic-economy relevance

The case demonstrates how control over a platform can be leveraged into related markets.

In a synthetic economy, similar questions could arise where an AI infrastructure provider controls:

cloud + model + operating environment + application distribution.

Principle

Platform integration can become a competition concern where it limits competitors' access to important distribution channels.

23. Case Law 4: Amazon Marketplace — European Commission

The European Commission investigated Amazon's use of marketplace seller data.

The investigation examined whether Amazon used non-public seller data to compete with independent sellers on its marketplace.

Amazon offered commitments addressing the Commission's concerns.

Synthetic-economy relevance

This case is highly relevant to AI markets because data generated by users or business customers can become a competitive resource.

An AI platform may potentially obtain information from:

developers;

API users;

customers;

sellers;

applications.

The competition question becomes:

Can the platform use commercially sensitive information generated by dependent businesses to compete against those same businesses?

24. Case Law 5: Google Shopping — European Commission

In Google Search (Shopping), Case AT.39740, the European Commission found that Google had abused a dominant position by favouring its own comparison-shopping service in search results.

The case concerned:

self-preferencing;

platform power;

visibility;

discrimination against rivals.

Synthetic-economy relevance

AI platforms may become important gateways for:

search;

recommendations;

shopping;

software;

information.

If an AI platform controls user recommendations, it could potentially favour its own products.

Principle

A dominant digital intermediary may face competition-law scrutiny where it uses control over an important gateway to favour its own competing service.

25. Case Law 6: Qualcomm — European Commission

The European Commission investigated Qualcomm's conduct in the baseband-chip market, including exclusivity-related arrangements.

The case illustrates how a company with significant technological market power can potentially strengthen its position through contractual arrangements.

Synthetic-economy relevance

AI infrastructure markets may similarly involve:

exclusive supply;

preferred access;

rebates;

capacity reservations;

long-term contracts.

Where computing resources are scarce, exclusive arrangements can potentially disadvantage emerging AI competitors.

Principle

Exclusivity involving an important technological input may become a competition concern when it forecloses rivals.

26. Case Law 7: AT&T v. United States

The historical AT&T antitrust litigation illustrates competition concerns surrounding control over telecommunications infrastructure.

The eventual structural separation of AT&T demonstrated the significance of infrastructure control for competition.

Synthetic-economy relevance

Modern AI infrastructure can similarly involve:

computing;

cloud infrastructure;

networks;

data centres.

If a vertically integrated firm controls both essential infrastructure and downstream services, competition authorities may consider whether competitors can obtain effective access.

27. Case Law 8: FTC v. Facebook (Meta)

The U.S. Federal Trade Commission's litigation involving Facebook concerned alleged maintenance of monopoly power through acquisitions and conduct involving emerging competitors.

The case is particularly relevant to synthetic economies because it concerns:

digital platforms;

network effects;

data;

acquisitions;

potential competition.

Synthetic-economy lesson

A dominant digital ecosystem may acquire firms that represent future competitive threats even when their current market position is relatively small.

This illustrates why competition analysis in synthetic economies should consider future innovation competition, not merely present market shares.

28. Synthetic Economies and Self-Preferencing

Self-preferencing can occur when a platform:

operates an infrastructure platform;

hosts competing applications;

controls ranking or recommendation;

favours its own products.

For example:

AI platform → AI marketplace → AI applications

If the platform's own application receives preferential treatment, competitors may be disadvantaged.

This resembles concerns considered in Google Shopping and other digital-platform cases.

29. Synthetic Economies and Vertical Foreclosure

A synthetic economy may have several levels:

Semiconductor → Cloud → Foundation Model → AI Application → Distribution

A company controlling multiple levels could potentially:

raise rivals' costs;

restrict access;

bundle services;

discriminate;

reserve scarce capacity;

impose exclusivity.

Competition authorities should therefore consider vertical market power rather than analysing each layer independently.

30. Synthetic Economies and Interoperability

Interoperability can be an important competitive safeguard.

Examples:

model interoperability;

data portability;

API access;

cloud portability;

application compatibility;

identity portability.

If customers can easily move between providers, market power may be constrained.

If switching is difficult, network effects may reinforce concentration.

31. Switching Costs

Synthetic economies can create significant switching costs because users may depend upon:

proprietary APIs;

customised models;

proprietary datasets;

software ecosystems;

cloud architecture;

AI workflows.

High switching costs can discourage customers from moving to competitors.

This may create durable market power even when several suppliers technically exist.

32. Lock-In

A synthetic economy may create:

Technological lock-in → reduced switching → reduced competitive pressure → greater market power.

For example, an enterprise may build its entire AI system around one provider's:

API;

database;

cloud infrastructure;

model;

security system.

Switching providers may then require significant expenditure.

Competition authorities may examine whether contractual or technical restrictions unnecessarily increase these costs.

33. Synthetic Economies and Intellectual Property

AI markets rely heavily on:

patents;

copyrights;

trade secrets;

model weights;

algorithms;

databases.

IP protection can promote innovation.

However, excessive control over essential technologies may create barriers to entry.

Competition law therefore has to balance:

innovation incentives

against

exclusionary effects.

34. Standard Essential Technologies

Synthetic economies may develop technical standards.

If a particular technology becomes essential to interoperability, control over the standard may provide significant market power.

Competition concerns may arise around:

discriminatory licensing;

excessive licensing fees;

refusal to license;

exclusionary standards.

35. Synthetic Economies and Consumer Welfare

Consumers may benefit from synthetic production through:

lower costs;

personalization;

greater variety;

faster services;

improved quality;

innovation.

But they may also face:

reduced privacy;

platform dependence;

algorithmic discrimination;

reduced choice;

higher switching costs;

hidden bundling.

Competition law therefore needs to consider both price and non-price dimensions of competition.

36. Quality Competition

Synthetic economies may compete on:

accuracy;

reliability;

speed;

privacy;

security;

explainability;

customization.

A merger or exclusionary practice that reduces these dimensions can harm competition even where monetary prices remain unchanged.

37. Innovation Competition

Innovation is especially important in synthetic economies.

A dominant firm might have strong current market share but still face competition from an emerging technology.

For example:

Existing AI technology

versus

new architecture

versus

open-source model

versus

specialised AI system

Competition law must therefore account for innovation pipelines and potential technological substitution.

38. Open-Source AI and Competition

Open-source models can potentially reduce concentration by:

lowering entry barriers;

allowing experimentation;

reducing dependence on proprietary models;

encouraging interoperability.

However, open-source systems may still depend upon concentrated:

cloud providers;

chips;

data centres;

distribution channels.

Thus:

Open-source software does not automatically eliminate infrastructure concentration.

39. Synthetic Economies and Cloud Competition

Cloud infrastructure is particularly important.

AI firms may require:

enormous computing capacity;

storage;

networking;

specialised processors.

If a small number of cloud providers dominate, AI startups may become dependent upon them.

Competition concerns could arise if a cloud provider:

favours its own AI products;

limits portability;

imposes restrictive contracts;

ties AI services to cloud services;

uses customer data competitively.

40. Synthetic Economies and Acqui-Hiring

AI firms may acquire startups primarily to obtain:

engineers;

researchers;

technical teams;

intellectual property.

Such acquisitions can reduce competition even if the target has little revenue.

The competitive concern is that acquisition may remove an independent source of:

innovation;

talent;

technology;

future entry.

41. Synthetic Economies and Labour Substitution

Synthetic economies can dramatically reduce dependence upon labour for some tasks.

This can:

lower production costs;

enable small teams to compete globally;

increase productivity.

But it may also favour firms with access to:

computing;

capital;

proprietary models;

data.

Therefore, automation can simultaneously lower some barriers to entry and increase others.

42. Synthetic Economies and Market Definition

Traditional market definition may become difficult.

Suppose consumers can obtain a service from:

human professionals;

software;

AI systems;

automated agents.

Should all be included in the same relevant market?

The answer depends on:

substitutability;

functionality;

quality;

price;

consumer behaviour;

technological capability.

Competition authorities may need to analyse functional substitution, not merely product labels.

43. Multi-Sided Synthetic Markets

Many synthetic-economy platforms operate multiple sides:

Users ↔ AI platform ↔ Developers ↔ Advertisers ↔ Data providers

A competition problem affecting one side may affect the entire ecosystem.

Authorities therefore need to consider:

cross-side network effects;

data feedback loops;

platform governance;

access rules;

ranking;

pricing.

44. Synthetic Economies and Predatory Innovation

A powerful firm might deliberately release a product designed to make competing technologies economically unviable.

Examples could include:

free AI services;

bundled AI features;

below-cost computing;

exclusive APIs.

Competition law should distinguish:

vigorous innovation

from

strategic exclusion.

The mere fact that a dominant firm innovates rapidly is not itself anti-competitive.

45. Competition Remedies

Possible remedies include:

Structural remedies

divestiture;

separation of business units;

sale of assets.

Access remedies

API access;

cloud access;

licensing;

interoperability.

Data remedies

data portability;

restrictions on use of competitor data;

information firewalls.

Conduct remedies

non-discrimination;

prohibition of self-preferencing;

restrictions on exclusivity.

Merger remedies

divestiture of competing AI businesses;

licensing of technology;

preservation of independent R&D teams.

46. Indian Competition-Law Perspective

Synthetic economies can be analysed under the Competition Act, 2002.

Important provisions include:

Section 3

Prohibits agreements having or likely to have an appreciable adverse effect on competition.

This can become relevant to:

algorithmic collusion;

exclusivity;

information sharing;

technology agreements.

Section 4

Deals with abuse of dominant position.

Possible issues include:

unfair or discriminatory conditions;

denial of market access;

leveraging;

tying;

predatory pricing.

Sections 5 and 6

Concern combinations and merger control.

They are relevant to:

AI acquisitions;

data acquisitions;

technology mergers;

vertical integration;

potential-competitor acquisitions.

47. Synthetic Economies and Competition Commission of India

The CCI may potentially examine factors such as:

market concentration;

technological barriers;

data advantages;

network effects;

switching costs;

entry barriers;

innovation;

vertical integration;

consumer dependence.

The analysis should not assume that a company is dominant merely because it has a sophisticated AI system.

Dominance must be assessed according to the relevant statutory and economic framework.

48. Major Competition Risks

The principal risks of synthetic economies can be summarized as:

AI infrastructure concentration.

Cloud concentration.

Data concentration.

Algorithmic coordination.

Self-preferencing.

Vertical foreclosure.

Exclusive agreements.

Killer acquisitions.

High switching costs.

Network effects.

Platform lock-in.

Control over standards.

Reduced innovation.

Discriminatory access.

Bundling and tying.

49. Potential Competition Benefits

Synthetic economies can also strengthen competition.

They may:

reduce production costs;

lower entry barriers;

enable small businesses to scale;

increase product variety;

accelerate innovation;

generate new competitors;

reduce dependence on scarce physical inputs;

improve personalization;

facilitate international market entry;

create new business models.

Competition law should therefore avoid treating AI or synthetic production itself as inherently problematic.

50. Six Core Case-Law Principles

51. Future Competition-Law Challenges

Synthetic economies are likely to create several difficult legal questions.

1. Who is the competitor?

A human business, AI company and autonomous AI agent may perform similar functions.

2. What is the relevant market?

The boundary between human and machine production may become unclear.

3. What constitutes coordination?

Algorithms may adapt without traditional communication.

4. What is the relevant asset?

The key competitive asset may be:

model;

data;

compute;

talent;

distribution;

algorithm.

5. How should future competition be measured?

Current revenue may not accurately capture the importance of an AI startup.

52. Conclusion

Synthetic economies represent a significant evolution in the structure of competition. They can dramatically reduce the marginal cost of producing digital goods and services while shifting competitive importance toward AI models, data, computing power, cloud infrastructure, algorithms, talent and distribution.

Their competition implications are therefore two-sided.

Pro-competitive possibilities

Automation → lower costs → innovation → new entry → greater consumer choice

Anti-competitive possibilities

Data + compute + models + network effects → concentration → lock-in → exclusion → reduced competition

The principal challenge for competition law is therefore to ensure that the benefits of synthetic production remain open to competing firms.

The most important areas for examination are:

AI infrastructure;

cloud computing;

synthetic data;

foundation models;

algorithmic pricing;

autonomous agents;

platform self-preferencing;

vertical integration;

exclusive arrangements;

AI mergers;

killer acquisitions;

interoperability;

switching costs;

innovation competition.

The cases of Microsoft, Google Shopping, Google Android, Amazon Marketplace, Qualcomm, Facebook/Meta and AT&T provide useful doctrinal foundations for analysing these emerging problems.

Quick Revision

Synthetic economies rely heavily on AI, automation and digital production.

Synthetic production can drastically reduce marginal costs.

Computing power may become a critical competitive input.

Data can remain an important source of market power.

Synthetic data can reduce some entry barriers.

Control over synthetic-data technology can create new bottlenecks.

AI platforms can generate strong network effects.

Autonomous algorithms may create new coordination risks.

AI mergers can eliminate future competitors.

Self-preferencing can affect AI distribution.

Vertical integration can create foreclosure concerns.

Cloud concentration can influence downstream AI competition.

Switching costs can reinforce platform power.

Interoperability can reduce lock-in.

Innovation competition is especially important in rapidly developing AI markets.

Microsoft illustrates platform leveraging.

Google Shopping illustrates self-preferencing.

Google Android illustrates platform tying and integration.

Amazon Marketplace illustrates data-related platform concerns.

Qualcomm illustrates exclusivity and technological inputs.

Facebook/Meta illustrates the importance of potential competition.

Indian analysis can involve Sections 3, 4, 5 and 6 of the Competition Act, 2002.

Synthetic economies can create both greater competition and new forms of concentration.

Competition law should therefore focus on preserving contestability, innovation, access and independent competitive alternatives.