Competition Law and Strategic Competition Policy for Intelligence-Centric Societies

1. Introduction

An intelligence-centric society is an economy in which competitive advantage increasingly depends upon the ability to collect, process, analyse and deploy information through artificial intelligence, data analytics, algorithms, cloud computing, digital platforms and automated decision-making.

In such an economy, traditional productive assets—factories, machinery and physical distribution networks—are increasingly supplemented by:

data;

AI models;

computing power;

algorithms;

digital platforms;

cloud infrastructure;

proprietary datasets;

digital identity systems;

automated decision systems;

intellectual property;

digital ecosystems.

This transformation creates a strategic competition-policy problem:

How should competition law preserve competitive markets when market power increasingly depends upon control over intelligence-generating resources rather than merely physical assets?

Competition law must therefore address not only price and output, but also:

data access;

innovation;

AI infrastructure;

interoperability;

algorithmic dependence;

ecosystem control;

digital gatekeeping;

computational resources;

potential competition.

2. Meaning of an Intelligence-Centric Society

An intelligence-centric economy can be represented as:

Data → Computing → Algorithms → Intelligence → Decisions → Economic Power

For example:

Consumer data → AI model → personalised recommendations → consumer allocation → platform power

or:

Industrial data → predictive AI → better production → lower costs → competitive advantage

The competitive significance of each layer may therefore be substantial.

3. Why Competition Policy Must Adapt

Traditional competition analysis often focuses upon:

market share;

prices;

output;

costs;

barriers to entry.

In intelligence-centric markets, competitive advantage may instead arise from:

Data economies of scale

More users generate more data.

Algorithmic learning

More data can improve an AI system.

Network effects

More users make a platform more valuable.

Computing economies

Large firms may obtain better access to GPUs and cloud infrastructure.

Ecosystem effects

A firm controlling several interconnected services can reinforce its position across markets.

This can produce:

Data → better intelligence → more users → more data → stronger intelligence.

That feedback loop can create durable market power.

4. Strategic Competition Policy

Strategic competition policy should protect contestability of intelligence markets.

Its objectives include:

preventing exclusionary control of critical data;

maintaining access to computing infrastructure;

preserving interoperability;

preventing abusive AI-platform tying;

controlling anti-competitive acquisitions;

preventing algorithmic collusion;

protecting innovation competition;

preserving access to digital ecosystems;

promoting competitive neutrality;

preventing permanent technological lock-in.

5. Competition Law Framework

A. Article 101 TFEU

Article 101 can address:

algorithmic collusion;

information exchange;

data-sharing agreements;

technology licensing restrictions;

market allocation;

standard-setting arrangements.

B. Article 102 TFEU

Article 102 is particularly relevant where a dominant undertaking controls:

critical datasets;

AI infrastructure;

cloud computing;

search;

app distribution;

digital advertising;

operating systems.

Potential abuses include:

refusal of access;

discriminatory access;

tying;

self-preferencing;

exclusionary rebates;

margin squeeze;

leveraging.

C. Indian Competition Act, 2002

In India, intelligence-centric competition issues can principally arise under:

Section 3

Anti-competitive agreements, including:

algorithmic coordination;

data-sharing arrangements;

technology restrictions;

market allocation.

Section 4

Abuse of dominance involving:

discriminatory access;

denial of market access;

tying;

leveraging;

exclusionary conduct.

Sections 5 and 6

Merger control involving:

AI companies;

data businesses;

cloud providers;

digital platforms;

emerging technologies.

6. Data as a Strategic Competition Asset

Data may function as an important competitive input.

Consider:

Platform A → 100 million users → extensive behavioural data → superior AI model → better recommendations → more users.

A smaller competitor may be unable to replicate the same data advantage.

Competition law must therefore ask:

Is the data genuinely unique?

Is it necessary for competition?

Are substitutes available?

Can competitors obtain equivalent data?

Is data being used to exclude competitors?

Does combining datasets strengthen dominance?

7. Data Advantage Does Not Automatically Mean Dominance

Possession of a large dataset does not automatically establish:

dominance;

abuse;

market foreclosure.

Data can often be:

replicable;

purchasable;

obtainable from public sources;

generated independently.

Therefore, competition authorities should determine whether the data creates a material and durable competitive advantage.

8. AI and Market Power

AI can reinforce existing market power through several mechanisms.

Data advantage

Large datasets improve model development.

Computing advantage

Large firms may obtain superior access to computing resources.

Distribution advantage

A platform may integrate AI into an existing ecosystem.

User advantage

An established platform can rapidly distribute an AI product to millions of users.

Feedback loops

More users generate more data and feedback, improving the system.

This can create AI-specific barriers to entry.

9. Case Law 1 — Microsoft Corp. v Commission, Case T-201/04

Facts

Microsoft was found to have abused its dominant position in the PC operating-system market through conduct involving interoperability information and tying.

Principle

A dominant undertaking controlling an important technological platform cannot necessarily use that control to restrict interoperability or extend its dominance into related markets.

Relevance

The case is highly relevant to intelligence-centric economies.

Modern equivalents may involve:

AI operating layers;

cloud platforms;

AI APIs;

digital assistants;

foundation models.

If control of a dominant platform enables a firm to disadvantage competing AI services, competition-law concerns may arise.

10. Case Law 2 — United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001)

Facts

Microsoft's conduct involving Windows and competing browser technology was examined under U.S. antitrust law.

Principle

A dominant platform can violate antitrust law when it uses its platform position to suppress competitive threats or restrict distribution of rival technologies.

Relevance

The case demonstrates the importance of platform neutrality.

In an AI economy, similar questions can arise where a dominant operating system, cloud platform or digital ecosystem controls distribution of competing AI products.

11. Case Law 3 — Google Search (Shopping), Case T-612/17

Facts

Google was found to have systematically favoured its comparison-shopping service in search results over competing comparison services.

Principle

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

Relevance

The case is important for intelligence-centric societies because AI-powered search and recommendation systems increasingly determine:

what information consumers see;

which businesses receive traffic;

which products receive visibility.

Thus, control over information allocation can become a source of market power.

12. Case Law 4 — Google Android, Case T-604/18

Facts

The European Commission examined Google's arrangements concerning Android, search, browser distribution and app ecosystems.

Principle

A dominant ecosystem operator can unlawfully reinforce its position through contractual and technological arrangements that restrict competing services.

Relevance

AI services may increasingly be distributed through:

operating systems;

app stores;

cloud platforms;

digital assistants.

A dominant infrastructure provider could potentially use these gateways to favour its own AI services.

13. Case Law 5 — Intel v Commission, Case C-413/14 P

Facts

The case concerned Intel's rebate arrangements and their potential exclusionary effects.

Principle

The assessment of exclusionary conduct by a dominant undertaking requires attention to whether the conduct is capable of restricting effective competition.

Relevance

Intelligence-centric markets may involve:

AI-computing discounts;

cloud credits;

exclusive AI distribution;

preferential access;

loyalty incentives.

The same competition-law question remains:

Does the commercial arrangement merely reward efficiency, or does it foreclose effective competitors?

14. Case Law 6 — Deutsche Telekom v Commission, Case C-280/08 P

Facts

Deutsche Telekom's pricing arrangements concerning telecommunications access were examined under Article 102.

Principle

A vertically integrated dominant undertaking may abuse its position where its pricing structure makes effective downstream competition difficult.

Relevance

The principle is transferable to intelligence infrastructure.

For example:

Cloud infrastructure → AI computing → AI applications.

If the infrastructure provider controls an indispensable upstream input and simultaneously competes downstream, pricing may become an instrument of foreclosure.

15. Case Law 7 — Slovak Telekom v Commission, Case C-165/19 P

Facts

The case involved access to telecommunications infrastructure and exclusionary conduct.

Principle

Infrastructure control can affect downstream competition where access arrangements prevent competitors from competing effectively.

Relevance

In intelligence-centric markets, infrastructure may include:

cloud platforms;

AI compute;

APIs;

datasets;

digital identity;

communication networks.

The case therefore supports an infrastructure-based analysis of digital intelligence markets.

16. Case Law 8 — Bronner v Mediaprint, Case C-7/97

Facts

The dispute concerned access to a newspaper distribution system.

Principle

A dominant undertaking is not automatically required to share every asset with competitors.

The essential-facilities conditions must be satisfied, including genuine indispensability and the absence of realistic alternatives.

Relevance

This is particularly important for AI.

A company should not automatically be required to provide:

proprietary algorithms;

source code;

confidential training data;

commercially sensitive technology

merely because it is successful.

Competition law must distinguish legitimate intellectual-property protection from control over a genuinely indispensable competitive bottleneck.

17. Case Law 9 — United States v. Terminal Railroad Association of St. Louis, 224 U.S. 383 (1912)

Facts

A group of railroads controlled important terminal infrastructure.

Principle

Control over a strategically indispensable facility can create exclusionary power where competitors cannot effectively access the market.

Relevance

The modern equivalent could be:

AI compute infrastructure + cloud access + proprietary data + platform distribution.

The physical infrastructure may differ, but the competition principle remains relevant: control of a bottleneck can confer strategic market power.

18. Intelligence Infrastructure

The future competitive structure can be understood through four layers:

Control over multiple layers can create vertical ecosystem power.

19. AI Ecosystem Lock-In

Consider:

Cloud provider → foundation model → AI application → marketplace → consumers

If one company controls all five layers, competitors may face:

higher costs;

limited interoperability;

restricted distribution;

data disadvantages;

switching costs.

Competition policy should therefore examine ecosystem foreclosure, not simply individual product markets.

20. Algorithmic Collusion

Intelligence-centric markets create a new form of competition risk.

Independent firms may use AI systems to determine prices.

Independently:

AI A → ₹100
AI B → ₹105
AI C → ₹98

This is not automatically unlawful.

But competition concerns increase if businesses:

exchange sensitive information;

use a common pricing algorithm;

deliberately design algorithms to coordinate;

communicate through algorithmic systems;

use algorithms for customer or geographic allocation.

The underlying competition-law prohibition remains applicable even where coordination is technologically mediated.

21. AI-Powered Personalised Pricing

AI can classify consumers according to:

willingness to pay;

purchasing history;

location;

behavioural patterns.

Personalisation can generate legitimate efficiencies.

However, competition authorities may investigate whether personalised pricing is used by a dominant undertaking to:

exploit market power;

discriminate unfairly;

exclude competitors;

prevent switching.

The mere existence of personalised pricing does not establish an antitrust violation; market power and competitive effects remain crucial.

22. Self-Preferencing by AI Systems

An AI platform could recommend:

“Our own service”
above
“competing service.”

If the platform is dominant and controls the information-distribution gateway, self-preferencing can become a competition issue.

This is particularly significant because AI-generated answers may replace traditional search-result pages.

The competitive question becomes:

Who decides which supplier the AI recommends?

23. AI and Consumer Choice

Traditional search markets provide consumers with multiple results.

Generative AI may provide:

one answer.

This creates a potentially important competitive change.

If an AI system controls the interface through which consumers discover:

products;

travel;

financial services;

healthcare providers;

software;

then recommendation architecture itself may become a competitive bottleneck.

24. Competition and AI Training Data

AI companies may compete for access to:

books;

news;

images;

video;

scientific data;

industrial data;

consumer interaction data.

Competition law may become relevant where dominant firms:

foreclose access to essential datasets;

enter exclusionary data arrangements;

acquire competing data sources;

use contractual restrictions to prevent rival AI development.

But intellectual-property rights and privacy considerations must also be respected.

25. Competition and Compute Concentration

Advanced AI requires substantial computing resources.

If a small number of firms control:

GPUs;

cloud computing;

data centres;

specialised AI chips,

they may become critical suppliers to AI developers.

Potential competition concerns include:

discriminatory cloud access;

exclusive compute arrangements;

tying cloud services to AI models;

refusal of interoperability;

preferential allocation of scarce computing capacity.

26. Merger Control in Intelligence-Centric Markets

Traditional turnover thresholds may fail to capture some strategically important transactions.

An AI startup may have:

low revenue;

highly valuable technology;

unique datasets;

important researchers;

significant future competitive potential.

A large incumbent may acquire the startup before it becomes a significant competitor.

Competition authorities should therefore examine:

Innovation competition

Could the startup become a significant innovator?

Data competition

Does the acquisition consolidate unique datasets?

Technology competition

Does it eliminate an alternative technological architecture?

Ecosystem effects

Does it strengthen an existing platform?

27. Killer Acquisitions and AI

A hypothetical example:

Major cloud company → acquires promising AI startup → integrates its technology → competing AI developers lose an emerging source of innovation.

The competitive concern is not necessarily the startup's current market share.

It may be the future competitive constraint that disappears.

Therefore:

Potential competition can be more important than current revenue in rapidly developing intelligence markets.

28. Interoperability

Interoperability can reduce digital concentration.

Examples include:

AI model portability;

API interoperability;

cloud portability;

messaging interoperability;

data portability.

If switching between systems is difficult, users become locked into dominant ecosystems.

Competition policy can therefore support interoperability where justified by:

competition;

consumer welfare;

innovation;

technical feasibility.

29. Competitive Neutrality

Public authorities may increasingly operate AI and digital infrastructure.

Examples include:

government cloud;

public digital identity;

AI research infrastructure;

public data platforms.

Where public infrastructure competes with private businesses, competitive neutrality becomes important.

Public ownership should not automatically confer competitive advantages unrelated to legitimate public functions.

30. Intelligence-Centric Markets and SMEs

Small businesses can become dependent upon major AI platforms for:

advertising;

customer acquisition;

cloud computing;

software;

automated customer service;

content generation.

This can create a new form of AI-platform dependency.

Competition policy should preserve the ability of SMEs to:

switch providers;

export data;

use multiple AI systems;

integrate competing services;

reach consumers independently.

31. Strategic Competition Policy Model

A useful framework is:

1. Identify the intelligence asset

Is the strategic resource:

data?

compute?

algorithm?

model?

distribution platform?

2. Identify market power

Does the firm have durable power?

3. Examine bottlenecks

Can competitors realistically bypass the firm's infrastructure?

4. Examine interoperability

Can competing systems connect?

5. Examine data access

Can competitors obtain comparable inputs?

6. Examine vertical integration

Does the firm control multiple levels of the value chain?

7. Examine innovation

Could the conduct eliminate future competitors?

8. Examine efficiencies

Does the conduct produce legitimate technological benefits?

9. Apply proportionate remedies

Possible remedies include:

interoperability;

data portability;

nondiscrimination;

access obligations;

divestiture in exceptional merger cases;

restrictions on tying;

algorithmic monitoring.

32. Competition Risks Across the Intelligence Economy

33. Relationship Between Competition and Innovation

An intelligence-centric competition policy must preserve dynamic competition.

A market can have:

low prices;

high output;

while still suffering from reduced innovation.

For example:

Dominant AI platform → acquires emerging rival → integrates technology → fewer independent innovation paths.

Therefore, competition authorities should examine:

research pipelines;

patents;

developer ecosystems;

technical roadmaps;

potential entrants;

investment incentives.

34. Competition Law Should Not Become Technology Regulation

There is an important limit.

Competition law should not automatically regulate every:

AI error;

privacy problem;

cybersecurity failure;

ethical concern;

unfair algorithmic outcome.

Those may fall under other legal regimes.

Competition law should focus on competitive harm:

exclusion;

foreclosure;

collusion;

market power;

anti-competitive mergers;

discriminatory access;

barriers to entry.

35. Core Principles Emerging from the Case Law

The cases collectively support the following propositions:

Microsoft — platform control can be used to restrict technological competition.

Google Shopping — control over information distribution can affect downstream competition.

Google Android — ecosystem arrangements can reinforce dominance.

Intel — exclusionary effects matter when assessing dominant-firm conduct.

Deutsche Telekom — control of an upstream infrastructure layer can affect downstream competition.

Slovak Telekom — access to infrastructure can be essential to effective competition.

Bronner — essential-facilities obligations must remain limited and principled.

Terminal Railroad — control over an indispensable bottleneck can produce exclusionary power.

36. Exam-Ready Conclusion

Competition law in an intelligence-centric society must evolve from a primarily price-centred conception of competition toward a broader analysis of data, computing, algorithms, platforms, interoperability, innovation and information distribution.

The principal risk is not simply that a company becomes technologically successful. The deeper risk is that control over data + compute + algorithms + distribution creates a self-reinforcing ecosystem in which competitors cannot realistically enter or expand.

A strategic competition policy should therefore protect:

contestability + data access + interoperability + innovation + competitive neutrality + algorithmic independence + effective merger control.

At the same time, competition law should not treat every large dataset, AI model or proprietary technology as an essential facility. The principles illustrated by Microsoft, Google Shopping, Google Android, Intel, Deutsche Telekom, Slovak Telekom, Bronner and Terminal Railroad require careful examination of dominance, indispensability, foreclosure, competitive effects, legitimate business justification and efficiencies.

Core proposition

In an intelligence-centric society, competition policy must protect competition not only for consumers and markets, but also for the strategic resources from which economic intelligence is produced—data, computing capacity, algorithms, AI models and digital distribution. Market power becomes a competition-law concern where control over these resources enables a firm to foreclose rivals, restrict interoperability, discriminate in access, eliminate potential competitors or extend dominance across interconnected markets.