Competition Law and Machine-Directed Mergers and Acquisitions

1. Introduction

Machine-directed mergers and acquisitions (M&A) refers to situations in which artificial intelligence (AI), machine-learning systems, automated valuation tools, algorithmic decision systems, or autonomous corporate software substantially assists—or potentially directs—the identification, evaluation, negotiation, execution, or post-merger integration of acquisitions.

Competition law does not generally create a separate legal category called “machine-directed M&A.” The existing rules on merger control, market power, substantial lessening of competition, abuse of dominance, coordination, and protection of innovation continue to apply. However, machine-directed transactions create new questions because machines may identify targets, predict competitive threats, evaluate data, recommend acquisitions, or even initiate transactions according to pre-programmed objectives.

The central competition-law question is therefore:

Does the use of a machine in making or implementing an acquisition change the competitive effects of the transaction, or merely change the method by which the transaction is carried out?

Usually, the technology itself is not unlawful. The important issues are market effects, control, competitive significance of the target, data and technology concentration, potential competition, innovation, and the degree of human responsibility for the transaction.

2. Meaning of Machine-Directed M&A

Machine-directed M&A may involve several levels of automation.

A. Machine-assisted M&A

Humans remain the decision-makers, but AI assists with:

identifying acquisition targets;

valuation;

due diligence;

market analysis;

competitor mapping;

financial forecasting;

patent analysis;

customer analysis;

risk assessment.

B. Machine-recommended M&A

The system goes beyond analysis and recommends:

which company should be acquired;

when the acquisition should occur;

the proposed price;

which competitors represent future threats;

whether a potential competitor should be eliminated through acquisition.

C. Machine-executed M&A

More advanced systems could automate portions of:

target selection;

bidding;

negotiation;

contractual drafting;

transaction execution;

post-acquisition integration.

D. Autonomous or machine-directed M&A

The most advanced scenario involves an autonomous system pursuing corporate objectives and making acquisition decisions with limited human intervention.

This raises particularly important questions about:

accountability;

merger notification;

human oversight;

competition effects;

algorithmic decision-making;

strategic acquisitions;

acquisition of potential competitors;

data concentration.

3. Why Machine-Directed M&A Creates Competition-Law Issues

Traditional merger analysis assumes that identifiable human decision-makers determine the transaction.

Machine-directed M&A potentially changes that assumption.

An AI system might determine that acquiring a small company is advantageous because the target possesses:

unique data;

a valuable algorithm;

patents;

AI talent;

cloud infrastructure;

customer relationships;

a developing technology;

a potential future competitor.

A conventional financial analysis might view the target as a small company.

An AI-based competition analysis may identify it as a strategically important future competitor.

Therefore, traditional indicators such as current turnover or market share may not fully capture competitive significance.

4. Main Competition-Law Issues

4.1 Relevant Market

Competition authorities must first determine the relevant market.

Depending on the transaction, this may involve:

AI models;

cloud computing;

semiconductor technology;

data services;

online platforms;

autonomous systems;

financial technology;

logistics technology;

software;

advertising technology;

robotics.

Machine-directed M&A can make market definition difficult because technological markets evolve rapidly.

A target may operate in a small current market but possess technology capable of entering a much larger future market.

5. Market Power and Algorithmic Decision-Making

The use of AI does not itself establish market power.

Competition authorities may examine:

market share;

barriers to entry;

network effects;

economies of scale;

access to data;

intellectual property;

computing resources;

switching costs;

interoperability;

ecosystem control;

financial resources;

technological advantages.

A machine may help a company accumulate market power, but the legal question remains whether the resulting transaction or conduct harms competition.

6. Acquisition of Potential Competitors

One of the most important issues is the acquisition of a potential competitor.

A large technology company might use an AI system to identify small firms that could eventually become competitors.

For example:

A dominant AI platform uses automated market intelligence to identify a small start-up developing a competing model and acquires it before the start-up becomes commercially significant.

The target's current market share may be tiny.

Nevertheless, the transaction could eliminate an important source of future competition.

This is sometimes discussed through the concept of nascent or potential competition.

7. Killer Acquisitions

A killer acquisition generally refers to an acquisition where an established firm purchases an emerging competitor or promising innovation partly to prevent the development of competitive pressure.

Machine-directed systems could potentially make such acquisitions easier because algorithms can continuously monitor:

start-up funding;

patents;

research publications;

employee movements;

product launches;

user growth;

developer activity;

technology performance.

The system could identify promising competitors much earlier than traditional corporate monitoring.

However, not every acquisition of a start-up is a killer acquisition.

Authorities must establish the relevant competitive theory of harm.

8. Data Concentration

Machine-directed M&A creates significant data-related concerns.

A transaction may combine:

customer databases;

behavioral data;

transaction information;

location information;

search data;

training data;

proprietary datasets;

industrial data.

The competitive importance of the data depends on factors such as:

uniqueness;

scale;

quality;

exclusivity;

replicability;

relevance to the product;

availability of alternatives.

Data concentration can increase barriers to entry, but data ownership alone does not automatically constitute unlawful market power.

9. AI Model and Compute Concentration

Modern AI competition can depend on access to:

GPUs;

specialized processors;

cloud computing;

training infrastructure;

foundation models;

model-development tools;

specialized datasets.

An acquisition that combines a powerful AI model with scarce computing resources may have effects extending beyond the target's existing product market.

Competition authorities may therefore examine vertical and ecosystem effects.

10. Vertical M&A

Machine-directed acquisitions may involve companies operating at different levels of the supply chain.

For example:

Semiconductor company → cloud provider → AI model developer → AI application → distribution platform

Acquiring a company at one level may allow the acquiring firm to restrict competitors at another level.

Potential concerns include:

foreclosure;

tying;

bundling;

discriminatory access;

refusal to supply;

self-preferencing;

interoperability restrictions.

Vertical integration, however, can also generate legitimate efficiencies.

11. Algorithmic Target Selection

A machine may rank potential acquisition targets according to criteria such as:

probability of becoming a competitor;

technological capability;

patent portfolio;

customer growth;

employee expertise;

valuation;

market-entry potential.

Competition authorities may ask:

Was the target selected because it was an efficient acquisition opportunity, or because its elimination would reduce future competition?

The algorithm's objective function therefore becomes relevant evidence.

12. The Problem of “Future Competition”

Traditional merger analysis often examines current competitive conditions.

Machine-directed M&A makes future competition particularly important.

A small company may have:

low revenue today;

few customers;

no significant market share;

but possess technology that could substantially disrupt an established market.

Therefore, authorities may examine:

innovation pipelines;

R&D capabilities;

patents;

product roadmaps;

technical talent;

venture financing;

customer adoption;

internal documents;

strategic plans.

13. AI and Merger Due Diligence

AI can improve due diligence by examining enormous quantities of:

contracts;

emails;

financial records;

patents;

customer information;

employment records;

litigation;

regulatory documents.

This can help identify competition risks.

However, automated due diligence may also create problems if:

training data is incomplete;

important documents are incorrectly classified;

the model produces false positives;

the model misses strategically important evidence;

the system has embedded assumptions.

Therefore, human verification remains important.

14. Algorithmic Pricing After the Merger

Competition concerns can continue after acquisition.

If an acquiring company integrates the target's AI system with its own pricing algorithm, the combined system might influence:

prices;

discounts;

supply;

allocation;

advertising;

bidding.

This could create risks of:

coordinated pricing;

exclusionary pricing;

discriminatory pricing;

algorithmic facilitation of collusion.

The acquisition itself and subsequent conduct should therefore be analysed separately.

15. Machine-Directed M&A and Merger Notification

Merger notification obligations generally depend on the applicable jurisdiction's legal thresholds.

The fact that an AI system made the decision does not ordinarily eliminate the legal responsibility of the acquiring undertaking.

Authorities may examine:

turnover;

assets;

transaction value;

market shares;

control;

voting rights;

economic influence.

In technology markets, transaction-value thresholds can become important where a start-up has limited revenue but substantial competitive potential.

16. Transaction Value and Start-Ups

Suppose:

Target revenue = very low;

Target market share = very low;

Purchase price = extremely high;

Target owns significant AI technology.

A traditional revenue threshold might fail to capture the transaction.

A transaction-value test can potentially capture economically significant acquisitions of emerging technology companies.

This is particularly relevant to acquisitions of:

AI start-ups;

biotech firms;

semiconductor designers;

cybersecurity companies;

data companies.

17. Machine-Directed M&A and Innovation Competition

Competition law increasingly considers innovation competition.

A transaction may reduce competition even where current price effects are limited.

Authorities can examine whether the acquisition reduces:

R&D;

product development;

technological experimentation;

independent innovation;

alternative business models.

This is especially important where companies compete through innovation rather than price.

18. Post-Merger Integration

The competition analysis does not necessarily end when the transaction closes.

The merged company may integrate:

datasets;

algorithms;

cloud infrastructure;

employees;

customer bases;

patents;

distribution networks.

Integration can create efficiencies but may also produce foreclosure.

Competition authorities may therefore examine whether the combined company can:

deny competitors access to inputs;

disadvantage rival applications;

restrict interoperability;

use confidential competitor information;

combine datasets to create an entry barrier.

19. Important Case Laws

The following cases are not all directly about autonomous or AI-directed M&A. Because machine-directed M&A is a developing concept, several are foundational or analogous authorities.

Case 1: United States v. Philadelphia National Bank

Citation: 374 U.S. 321 (1963)

Principle

The U.S. Supreme Court recognized that mergers between significant competitors can substantially lessen competition and that market concentration is an important consideration.

Relevance

Machine-directed M&A systems may identify acquisitions based on market concentration and competitive positioning. Authorities can therefore examine whether an AI-selected transaction materially increases concentration.

20. Case 2: FTC v. Heinz

Citation: 246 F.3d 708 (D.C. Cir. 2001)

Principle

The court scrutinized a merger that would substantially increase concentration in the relevant market.

Relevance

The case illustrates the importance of examining market structure rather than merely asking whether the acquiring company and target are individually large.

For machine-directed M&A, an algorithm might identify a small target, but the transaction could nevertheless materially alter competitive structure.

21. Case 3: FTC v. Staples, Inc.

Citation: 970 F. Supp. 1066 (D.D.C. 1997)

Principle

The court carefully examined the relevant product market and competitive effects of the proposed Staples–Office Depot merger.

Relevance

The case demonstrates the importance of precise market definition in merger analysis.

For machine-directed M&A, automated systems may identify a broader or narrower competitive landscape than conventional business analysis. Competition authorities must independently determine the legally relevant market.

22. Case 4: United States v. Microsoft Corp.

Citation: 253 F.3d 34 (D.C. Cir. 2001)

Principle

Microsoft's conduct concerning the browser market was examined in the context of monopoly power, exclusionary conduct, and protection of Microsoft's operating-system position.

Relevance

Although this was not an M&A case, it is highly relevant by analogy to machine-directed acquisitions because it demonstrates how an incumbent can use an existing technological ecosystem to protect its position against emerging competitive threats.

For AI-driven M&A, authorities may examine whether an acquisition strengthens an existing ecosystem in a way that forecloses future competition.

23. Case 5: Google Shopping

Case: Google and Alphabet v European Commission

Citation: Case T-612/17, General Court, 2021

Principle

The European Union courts considered Google's treatment of competing comparison-shopping services and the competitive effects of preferential placement within Google's ecosystem.

Relevance

The case is important for understanding ecosystem power and self-preferencing.

A machine-directed acquisition may combine an important technology with a powerful platform, creating opportunities for preferential treatment or foreclosure after the acquisition.

24. Case 6: Intel v Commission

Citation: Case C-413/14 P, Court of Justice of the European Union, 2017

Principle

The Court emphasized the importance of examining the actual or potential exclusionary effects of certain practices involving dominant firms rather than relying solely on formal classification.

Relevance

For machine-directed M&A, competition authorities should examine actual competitive effects rather than assuming that the presence of AI or technological integration is itself harmful.

The case supports a careful effects-based approach.

25. Case 7: United Brands v Commission

Citation: Case 27/76, [1978] ECR 207

Principle

The Court established important principles concerning market definition and dominance.

Relevance

Machine-directed M&A can involve sophisticated technologies where the relevant market is not immediately obvious.

United Brands illustrates the continuing importance of:

substitutability;

market definition;

dominance;

competitive constraints.

26. Case 8: Hoffmann-La Roche v Commission

Citation: Case 85/76, [1979] ECR 461

Principle

The Court explained the concept of dominance and the special responsibility of dominant undertakings not to undermine genuine competition.

Relevance

If a dominant technology company uses machine systems to systematically identify and acquire emerging competitors, the competitive consequences may require particularly careful scrutiny.

The important point is that dominance increases the competition-law sensitivity of conduct, but an acquisition is not automatically unlawful merely because the purchaser is dominant.

27. Case 9: FTC v. Meta Platforms

The broader U.S. litigation concerning Meta's acquisitions of Instagram and WhatsApp illustrates a contemporary competition-law concern: whether acquisitions of emerging platforms can preserve or eliminate potential competition.

Relevance

The case demonstrates the importance of analysing:

nascent competition;

potential competition;

network effects;

platform ecosystems;

innovation;

acquisition history.

For machine-directed M&A, these considerations become particularly significant because an automated system could systematically identify emerging competitors before they reach substantial scale.

Important: The precise legal status and findings of contemporary litigation should be distinguished from established historical precedents; the broader lesson is the competition-law relevance of acquisitions of emerging digital competitors.

28. Case 10: Illumina/GRAIL

European Commission merger-control proceedings

The Illumina/GRAIL dispute is particularly important for technology and innovation-oriented merger control.

Principle

The transaction raised questions about the acquisition of an innovative company with limited conventional revenue but potentially important competitive significance.

Relevance

It demonstrates why merger control may need to consider:

innovation;

future competition;

nascent markets;

technological development;

competitive potential.

This is closely connected to machine-directed acquisition strategies.

29. Case 11: Microsoft/Activision Blizzard

European Commission merger-control decision and related international proceedings

The transaction demonstrates the importance of examining digital ecosystems and vertical relationships in technology-sector mergers.

Relevant concerns included:

gaming;

cloud services;

content;

distribution;

licensing;

foreclosure.

Relevance

A machine-directed acquisition system may view several apparently separate markets as parts of a single technological ecosystem.

Competition authorities therefore need to examine both:

horizontal competition + vertical/ecosystem effects.

30. Machine-Directed M&A and Killer Acquisitions

A machine may be particularly effective at identifying acquisitions that traditional management would overlook.

For example:

AI system identifies a start-up with only 20 employees because its patent portfolio, research publications and developer activity indicate that it may become a significant competitor within five years.

The competition authority should not automatically treat the transaction as unlawful.

Instead, it should investigate:

What technology does the target possess?

Could it become a competitor?

Does the acquirer already possess substantial market power?

Are there alternative purchasers?

Would the target continue independent development?

What efficiencies does the acquisition generate?

Would competition or innovation be substantially reduced?

31. Autonomous Acquisition and Corporate Responsibility

An important legal principle is:

Automation does not normally eliminate the legal responsibility of the undertaking using the system.

If a company delegates acquisition analysis to an AI system, the company generally remains responsible for complying with applicable merger and competition laws.

The machine cannot ordinarily be treated as an independent legal person merely because it made the recommendation.

32. Human Oversight

A strong governance system should provide:

Human review of:

target identification;

market definition;

competitive-risk assessment;

merger filing;

remedies;

integration plans.

Auditability

The company should retain records showing:

what data the system used;

what assumptions it applied;

why the target was selected;

what alternatives were considered;

who approved the transaction.

This becomes particularly important if regulators later investigate the transaction.

33. Algorithmic Bias in Acquisition Decisions

AI systems can produce biased acquisition recommendations.

For example, an algorithm trained on historical acquisitions may learn:

“Acquire firms that resemble previous successful acquisitions.”

This could cause the system to repeatedly favour certain:

technologies;

geographic markets;

business models;

competitors.

From a competition perspective, this may result in systematic acquisition of particular classes of emerging competitors.

34. Algorithmic Screening of Competitors

A large platform could continuously monitor:

start-up financing;

patent applications;

employee movements;

product launches;

web traffic;

customer growth;

developer activity.

This creates a new form of continuous competitor surveillance.

The competition concern is not surveillance itself.

The concern arises where the information is used to systematically eliminate competitive threats through acquisitions or exclusionary conduct.

35. Efficiency Defences

Machine-directed M&A can produce substantial efficiencies.

Possible efficiencies include:

reduced R&D duplication;

improved AI models;

better cybersecurity;

lower infrastructure costs;

faster innovation;

improved interoperability;

better products;

reduced transaction costs.

Competition law should therefore distinguish:

harmful elimination of competition

from

legitimate integration producing verifiable efficiencies.

36. Remedies

If competition concerns arise, authorities may consider:

Structural remedies

prohibition of the transaction;

divestiture;

sale of particular assets;

separation of businesses.

Behavioural remedies

licensing;

interoperability;

access commitments;

non-discrimination;

data-access commitments;

restrictions on exclusive dealing.

Technology-oriented remedies

API access;

data portability;

interoperability;

separation of datasets;

restrictions on combining certain data;

independent monitoring.

37. Ex-Ante and Ex-Post Regulation

Machine-directed M&A requires both approaches.

Ex-ante

Before closing:

merger notification;

market analysis;

competitive-effects assessment;

potential-competition analysis;

innovation analysis.

Ex-post

After closing:

monitoring integration;

investigating foreclosure;

examining exclusionary conduct;

reviewing algorithmic pricing;

assessing interoperability;

monitoring acquisition-related commitments.

38. Machine-Directed M&A in Different Sectors

39. Key Legal Principles

The following principles are especially important:

AI involvement does not itself make an acquisition unlawful.

The acquiring undertaking remains responsible for compliance.

Current market share may not reveal future competitive significance.

Potential competition can matter.

Innovation competition can matter.

Data may constitute an important competitive asset.

Network effects can increase competitive significance.

Vertical integration requires effects-based analysis.

Technology concentration does not automatically equal unlawful dominance.

Efficiency claims should be supported by evidence.

Human oversight is important for high-impact decisions.

Post-merger conduct must be separately analysed.

40. Short Revision Table

41. Conclusion

Machine-directed M&A represents an emerging application of competition law rather than a separate branch of merger law.

The principal challenge is that an AI system can identify competitive threats much earlier and more systematically than traditional corporate decision-making. Consequently, competition authorities may need to examine not only present market share and turnover, but also:

potential competition;

innovation;

data;

algorithms;

AI models;

computing capacity;

network effects;

ecosystem power;

nascent competitors;

post-merger integration.

The foundational authorities such as Philadelphia National Bank, Heinz, Staples, Microsoft, United Brands, Hoffmann-La Roche, Intel, Google Shopping, Illumina/GRAIL, and Microsoft/Activision Blizzard provide principles that can be adapted to these emerging transactions.

The central legal principle is therefore:

The fact that a machine directed or assisted an acquisition does not determine its legality; competition law focuses on the transaction's effects on competition, innovation, market structure, and consumers, while ensuring that human corporate actors remain accountable for compliance.