Competition Law and Strategic Governance of Machine Economies

Introduction

A machine economy is an economic environment in which important commercial decisions are increasingly made, assisted, or coordinated by software, artificial intelligence, algorithms, automated agents, platforms, sensors, smart contracts, and interconnected computational systems.

Examples include:

• algorithmic pricing and dynamic pricing;
• AI-driven marketplaces;
• autonomous trading and procurement;
• platform recommendation and ranking systems;
• automated advertising auctions;
• AI-based credit allocation;
• robotic logistics and supply chains;
• autonomous vehicles and mobility platforms;
• smart energy markets;
• machine-to-machine transactions;
• cloud and AI infrastructure;
• algorithmic merger strategies and investment decisions.

Competition law traditionally assumes that human firms make decisions. Machine economies complicate this assumption because commercially significant conduct may emerge from algorithms that learn, optimise, predict and react continuously.

The central competition-law question therefore becomes:

Who is responsible when machines perform the competitive function traditionally performed by human decision-makers?

The answer under existing competition law is generally that automation does not itself create immunity from antitrust liability. The legal analysis remains concerned with the underlying conduct, economic effects, agreements, market power and responsibility for deploying the technology.

I. Meaning of Strategic Governance of Machine Economies

Strategic governance goes beyond simply prosecuting individual antitrust violations.

It involves designing competition rules capable of dealing with:

1. algorithmic coordination;
2. machine-generated pricing;
3. AI-enabled market power;
4. control over data and computing infrastructure;
5. platform ecosystems;
6. automated exclusion and self-preferencing;
7. interoperability and access;
8. machine-to-machine contracting;
9. autonomous agents capable of making commercial decisions; and
10. future competitive risks that may arise before conventional market structures become entrenched.

The regulatory challenge is therefore partly ex post—detecting unlawful conduct—and partly ex ante—preventing technological architecture from becoming a mechanism for durable exclusion.

II. Why Machine Economies Create New Competition Problems

1. Algorithms can coordinate prices

Competitors no longer necessarily need to communicate through emails, meetings or telephone calls.

A common pricing algorithm may continuously observe market conditions and recommend or establish prices.

This creates several possible scenarios:

A. Explicit coordination

Competitors deliberately use technology to implement an agreement.

B. Algorithm-mediated coordination

Competitors provide sensitive information to a common algorithm which then produces coordinated pricing recommendations.

C. Tacit algorithmic coordination

Independent algorithms repeatedly observe one another and converge on prices without an express agreement.

The third category is particularly difficult because conventional competition law generally distinguishes between independent parallel conduct and an unlawful agreement or concerted practice.

III. Foundational Case Laws

1. Eturas UAB and Others v Lietuvos Respublikos konkurencijos taryba, Case C-74/14

This is one of the most important European cases concerning algorithmic coordination.

Travel agencies used a common computerised booking system. The system administrator introduced a restriction automatically limiting discounts available to customers. A message was circulated through the system concerning the restriction.

The Court of Justice considered whether the conduct could constitute a concerted practice under Article 101 TFEU.

Principle

The case demonstrates that competition law can apply where software architecture facilitates coordination between competitors.

However, the existence of a common technological system does not automatically establish liability for every participant. Evidence concerning knowledge, participation and the ability to distance oneself from the coordination remains important.

Importance for machine economies

Eturas establishes an important conceptual proposition:

A digital intermediary or automated system can become the mechanism through which competitive coordination is implemented.

Thus, competition authorities must examine not only communications between companies but also:

• system architecture;
• automated restrictions;
• system messages;
• software instructions;
• access logs;
• algorithmic parameters; and
• evidence showing knowledge or participation.

2. Meyer v Kalanick, 174 F. Supp. 3d 817 (S.D.N.Y. 2016)

This litigation concerned Uber's pricing algorithm.

The plaintiff alleged that Uber's algorithm effectively prevented drivers from competing independently on price and facilitated supra-competitive fares.

The district court refused to dismiss the antitrust allegations at the pleading stage, allowing the theory concerning an arrangement involving Uber and its drivers to proceed.

Principle

The case illustrates how an algorithm can potentially become the mechanism for implementing a vertical or horizontal pricing arrangement.

The legal issue is not whether a computer calculated the price.

The relevant question is:

What economic and contractual arrangement caused competitors or market participants to follow the algorithm?

Machine-economy significance

The case demonstrates the difference between:

“the algorithm independently determined the price”

and

“market participants agreed to use a system that determines the price for them.”

That distinction can be critical under antitrust law.

3. United States v David Topkins

In United States v David Topkins, the U.S. Department of Justice prosecuted an online poster seller for horizontal price fixing. The case involved competitors using algorithms to implement an agreement concerning prices for online posters. The defendant entered a guilty plea in 2015.

Principle

An algorithm does not transform conventional price fixing into lawful conduct.

If competitors agree upon prices and use software to implement that agreement, the software is simply the technological instrumentality of the cartel.

Strategic governance implication

Competition authorities should therefore investigate:

• source code;
• algorithmic instructions;
• pricing rules;
• communications concerning algorithm design;
• data inputs;
• common software providers; and
• changes in pricing behaviour following software implementation.

The machine is not the legal actor replacing the corporation; rather, the corporation's use of the machine may constitute the relevant conduct.

4. Google and Alphabet v European Commission — Google Shopping, Case T-612/17

The General Court upheld the central finding that Google had abused its dominant position by favouring its own comparison-shopping service over competing comparison-shopping services.

The case concerned automated search results and ranking mechanisms.

Principle

A machine economy may create competition problems through algorithmic allocation of visibility, rather than through prices alone.

A platform can influence competition by determining:

• which products appear;
• where they appear;
• how prominently they appear;
• which competitors receive traffic;
• which businesses receive data; and
• which services are integrated into the ecosystem.

Strategic significance

The Google Shopping litigation demonstrates that algorithmic ranking can constitute an important competitive parameter.

In machine economies, therefore, competition authorities must examine:

Who controls the algorithm that determines access to customers?

This can be as important as asking who controls the price.

5. Cornish-Adebiyi v Caesars Entertainment — Algorithmic Hotel Pricing

In 2024, the U.S. Department of Justice and Federal Trade Commission filed a statement of interest concerning allegations involving algorithmic hotel-room pricing.

The agencies stated that hotels cannot use an algorithm to engage in conduct that would violate antitrust law if performed directly by humans. They also highlighted concerns where a common algorithm provider receives competitively sensitive information from competing firms.

Principle

The crucial concept is:

Algorithmic implementation does not erase the underlying antitrust problem.

If competitors use a common pricing intermediary to exchange sensitive information and coordinate pricing, the technological form does not necessarily change the legal character of the conduct.

Machine-economy significance

This is particularly important for:

• hotel pricing;
• airline pricing;
• ride-hailing;
• rental housing;
• food delivery;
• advertising;
• logistics; and
• financial markets.

6. RealPage Algorithmic Pricing Litigation

The U.S. Department of Justice filed an antitrust lawsuit against RealPage in 2024, alleging that competing landlords supplied competitively sensitive information to RealPage and that its revenue-management software used that information to generate pricing recommendations. The complaint invoked Sections 1 and 2 of the Sherman Act.

Alleged mechanism

The government alleged a cycle in which:

Competitors' sensitive information

↓

Centralised software

↓

Algorithmic processing

↓

Pricing recommendations

↓

Landlord implementation

↓

Reduced independent price competition

Importance

The case represents a particularly important form of machine-economy governance because the alleged competitive problem is not simply a traditional cartel with human meetings.

The alleged coordination is embedded in:

• data collection;
• software architecture;
• algorithmic recommendations;
• information sharing; and
• repeated automated pricing.

The case therefore raises the question whether software infrastructure itself can facilitate market coordination.

7. Amazon Marketplace Competition Investigation

The UK Competition and Markets Authority investigated Amazon's marketplace concerning the use of third-party seller data, Buy Box selection and delivery-rate negotiations. The CMA accepted commitments from Amazon and closed the investigation in November 2023.

Importance

This demonstrates another dimension of machine economies:

Control over the algorithmic marketplace can create competition concerns even where the issue is not conventional price fixing.

A platform may simultaneously act as:

• infrastructure provider;
• marketplace operator;
• competitor to marketplace sellers;
• data collector; and
• algorithmic gatekeeper.

That creates a structural conflict.

IV. Major Competition-Law Problems in Machine Economies

1. Algorithmic Price Fixing

This is the most obvious issue.

A machine economy can facilitate:

• explicit price fixing;
• indirect coordination;
• common pricing algorithms;
• automated monitoring of competitors;
• real-time retaliation against discounting;
• elimination of price deviations.

Traditional cartel enforcement therefore needs to evolve from:

“Did executives communicate?”

towards:

“What information, instructions and technological mechanisms caused market participants to coordinate?”

V. Algorithmic Tacit Collusion

This is one of the hardest issues.

Imagine four competitors independently deploy reinforcement-learning algorithms.

Each algorithm learns:

“When I reduce price, competitors retaliate.”

Eventually the algorithms learn:

“Maintaining a high price maximises expected profits.”

No executive has instructed the algorithm to form a cartel.

The economic result may nevertheless resemble coordinated pricing.

Legal difficulty

Competition law generally requires more than merely observing parallel pricing.

Therefore, regulators must distinguish:

Independent intelligent adaptation

from

facilitated or consciously implemented coordination.

This creates a major gap between economic theory and conventional legal doctrine.

VI. Machine Learning and Market Power

Machine economies create new sources of dominance.

1. Data advantage

A dominant company may possess:

• enormous transaction datasets;
• behavioural information;
• consumer histories;
• proprietary training data;
• real-time market information.

Competitors may be unable to reproduce the same dataset.

2. Computing advantage

AI competition increasingly depends upon:

• GPUs;
• cloud infrastructure;
• model-training capacity;
• specialised chips;
• energy;
• data centres.

Control over computational infrastructure can therefore become an important competitive input.

3. Feedback loops

A platform may develop:

More users → more data → better algorithm → better service → more users.

This produces a reinforcing cycle.

Such feedback loops may make market entry increasingly difficult.

VII. Algorithmic Self-Preferencing

Machine economies often use ranking algorithms.

A vertically integrated platform can potentially favour:

• its own products;
• its own payment system;
• its own logistics service;
• its own advertising products;
• its own AI model;
• its own marketplace sellers.

The Google Shopping case demonstrates the significance of algorithmic preference and ranking in abuse-of-dominance analysis.

The competitive concern is not necessarily that the algorithm is inaccurate.

It is that the platform controlling the algorithm may simultaneously control the rules of visibility for its competitors.

VIII. Data as a Strategic Competition Asset

Machine economies convert data into an economic resource.

Competition authorities therefore need to consider:

Data accumulation

Can a dominant platform continuously collect data unavailable to rivals?

Data combination

Can the firm combine information from multiple markets?

Data portability

Can users transfer their information to competitors?

Data interoperability

Can rival systems interact with the dominant platform?

Data foreclosure

Can refusal to provide access prevent effective competition?

IX. AI Infrastructure and Vertical Foreclosure

The machine economy increasingly has a layered structure:

Semiconductors

↓

Cloud computing

↓

Foundation models

↓

AI applications

↓

Platforms

↓

Consumers and businesses

A company controlling several layers can potentially disadvantage rivals at downstream levels.

For example:

Cloud provider → AI model → marketplace → advertising → consumer service.

Competition analysis may therefore need to examine ecosystem power, rather than looking at each market in complete isolation.

X. Autonomous Agents and Machine-to-Machine Commerce

A future machine economy may contain AI agents that independently:

• negotiate prices;
• purchase inventory;
• select suppliers;
• enter contracts;
• bid in auctions;
• trade securities;
• purchase advertising;
• negotiate logistics;
• allocate electricity;
• change prices.

This raises a fundamental question:

Who is responsible for an autonomous agent's decision?

Possible legal approaches include responsibility of:

1. the owner;
2. the developer;
3. the operator;
4. the platform;
5. the data provider;
6. the algorithmic intermediary; or
7. multiple participants.

Competition law will probably continue to focus on human and corporate responsibility for designing, deploying and controlling the system, rather than treating software as a completely autonomous legal person.

XI. Competition Law and Machine Governance

Strategic governance can be organised around five layers.

Layer 1 — Human governance

Companies should identify:

• who approved the algorithm;
• who controls it;
• who changes its parameters;
• who receives algorithmic reports;
• who monitors competitive effects.

Layer 2 — Data governance

Authorities should examine:

• data sources;
• data sharing;
• sensitive information;
• access conditions;
• data portability;
• information asymmetry.

Layer 3 — Algorithm governance

Relevant questions include:

• What objective does the algorithm optimise?
• What constraints does it use?
• Does it monitor competitors?
• Does it automatically react to competitor prices?
• Can employees override it?
• Is there audit logging?

Layer 4 — Market governance

Authorities must examine:

• entry barriers;
• network effects;
• switching costs;
• interoperability;
• multi-homing;
• ecosystem dependencies.

Layer 5 — Regulatory governance

Authorities may use:

• antitrust investigations;
• merger control;
• behavioural remedies;
• structural remedies;
• access obligations;
• interoperability requirements;
• data-access remedies;
• algorithmic auditing;
• monitoring trustees;
• digital-market regulation.

XII. Machine Economies and Merger Control

Traditional merger analysis asks whether:

Company A + Company B = increased market concentration.

Machine economies require additional questions.

1. Data acquisition

Does the merger combine unique datasets?

2. Algorithm acquisition

Does the transaction remove an emerging algorithmic competitor?

3. AI capability

Does the acquisition provide control over an important AI model?

4. Infrastructure

Does the transaction combine critical cloud, semiconductor or computing resources?

5. Ecosystem leverage

Can the merged company extend power from one market into another?

6. Future competition

Could a small AI company become an important competitive constraint?

This is particularly relevant because conventional turnover-based merger thresholds may fail to capture strategically important acquisitions of low-revenue technology firms.

XIII. Essential Facilities in Machine Economies

The essential-facilities concept may become relevant where access to a technological input is indispensable for competition.

Potential examples include:

• cloud infrastructure;
• specialised computing;
• payment infrastructure;
• app stores;
• interoperability interfaces;
• critical datasets;
• digital identity infrastructure;
• AI model interfaces.

However, technological importance alone does not automatically establish an essential facility. Traditional legal requirements concerning indispensability, dominance and competitive foreclosure remain important.

XIV. Interoperability as a Competition Remedy

Interoperability may become one of the most important remedies in machine economies.

For example:

Platform A

↕

Common technical interface

↕

Platform B

This can reduce:

• switching costs;
• network-effect barriers;
• ecosystem lock-in;
• data silos.

The objective is not necessarily to make every system identical but to prevent technical architecture from becoming an artificial barrier to competition.

XV. Algorithmic Auditing

Competition authorities may increasingly need technical expertise.

An algorithmic investigation may require examination of:

• source code;
• model architecture;
• training datasets;
• model outputs;
• API logs;
• version histories;
• parameter changes;
• pricing records;
• communications;
• system documentation;
• employee instructions.

The evidentiary process therefore becomes partly technical and computational, rather than purely documentary.

XVI. Evidence Problems

Machine economies create unusual evidence.

A traditional cartel may leave:

“Let's increase the price by 10%.”

An algorithmic cartel may instead leave:

configuration files + API calls + model logs + pricing outputs + training data.

Therefore, competition authorities need sophisticated methods for reconstructing causation.

Important evidence can include:

Input

What information did the machine receive?

Processing

What rules or model transformed the information?

Output

What recommendation or action resulted?

Human intervention

Did employees approve or modify the result?

Market effect

Did competitors subsequently change their behaviour?

XVII. Strategic Governance Model

A useful governance framework can be represented as:

DATA

↓

COMPUTATIONAL INFRASTRUCTURE

↓

ALGORITHM / AI MODEL

↓

AUTOMATED DECISION

↓

MARKET BEHAVIOUR

↓

COMPETITIVE EFFECT

↓

LEGAL INTERVENTION

The regulator should therefore ask six questions:

1. Who controls the data?
2. Who controls the computational infrastructure?
3. Who designs the algorithm?
4. Who controls its objectives?
5. Who benefits from the resulting market behaviour?
6. Can rivals realistically discipline the system through competition?

XVIII. Difference Between Traditional and Machine-Economy Competition Law

XIX. Indian Competition-Law Relevance

For India, these issues principally arise under the Competition Act, 2002, particularly:

• Section 3 — anti-competitive agreements;
• Section 4 — abuse of dominant position;
• Sections 5 and 6 — combinations;
• Section 19 — inquiry powers;
• Section 26 — investigation;
• Section 27 — orders and remedies.

The Competition Commission of India may therefore need to analyse algorithmic conduct through existing doctrines while developing appropriate economic and technical methodologies.

Potential Indian machine-economy sectors include:

• digital payments;
• UPI-related ecosystems;
• e-commerce;
• food-delivery platforms;
• ride-hailing;
• online advertising;
• cloud computing;
• AI services;
• digital lending;
• online travel;
• logistics;
• energy-management systems; and
• autonomous mobility.

XX. Future Regulatory Architecture

A mature competition regime for machine economies could contain seven components.

1. Algorithmic transparency

Not necessarily complete disclosure of source code, but sufficient information for regulators to understand competitive effects.

2. Independent algorithmic auditing

High-risk dominant platforms could be subject to technical audits.

3. Data-access rules

Critical data advantages could be scrutinised where they create foreclosure.

4. Interoperability

Dominant ecosystems could be required to permit technically reasonable interoperability.

5. Real-time competition monitoring

Authorities could develop systems capable of detecting suspicious algorithmic pricing patterns.

6. AI merger scrutiny

Competition agencies could examine acquisitions of AI assets even where conventional turnover measures appear small.

7. Accountability

Companies should maintain clear responsibility for algorithmic decisions.

XXI. Important Doctrinal Lessons From the Cases

The cases collectively demonstrate several propositions:

First

Technology does not immunise anticompetitive conduct.

Topkins, Eturas and the algorithmic pricing litigation illustrate this principle.

Second

An algorithm can become the mechanism of coordination.

Eturas demonstrates the relevance of a common computerised system to concerted-practice analysis.

Third

Platform algorithms can affect competition without fixing prices.

Google Shopping illustrates the importance of algorithmic ranking and self-preferencing.

Fourth

Data-sharing architecture matters.

The RealPage allegations illustrate how a common software system receiving sensitive information from competitors can become central to an antitrust theory.

Fifth

Machine economies require ecosystem analysis.

Amazon Marketplace demonstrates the competition concerns that can arise where a platform simultaneously controls marketplace infrastructure, seller information and algorithmic selection mechanisms.

XXII. Conclusion

Competition law and strategic governance of machine economies represent the transition from regulating markets operated primarily by humans to regulating markets increasingly structured by computational systems.

The central legal principle should remain technologically neutral:

An algorithm should not make otherwise unlawful coordination lawful merely because the coordination is automated.

At the same time, competition law must avoid treating every form of algorithmic parallelism as a cartel. Independent algorithms may legitimately respond to common market conditions. The difficult cases arise where technology is deliberately structured to facilitate coordination, exclusion, self-preferencing, information exchange or durable market power.

The emerging legal framework therefore needs to combine:

antitrust doctrine + economics + data governance + algorithmic auditing + digital-market regulation + merger control + technical expertise.

The six-plus cases above—particularly Eturas, Meyer v Kalanick, Topkins, Google Shopping, Cornish-Adebiyi and RealPage, together with Amazon Marketplace—show the evolution from traditional human coordination toward software-mediated and data-driven competition problems. The CMA's continuing digital-market work, including investigations and conduct requirements concerning major digital platforms, also illustrates the movement toward more proactive governance of technologically structured markets.

Exam proposition: The principal challenge of machine-economy competition law is not to regulate machines as independent economic persons, but to ensure that firms cannot use computational architecture, data, algorithms and autonomous decision systems to evade, reproduce or entrench conduct that competition law is designed to prevent.