Competition Law and Predictive Infrastructure Market Dominance

1. Introduction

Predictive infrastructure market dominance refers to situations in which an enterprise uses predictive analytics, artificial intelligence (AI), machine learning, data infrastructure, forecasting systems, or other computational tools to obtain or reinforce a dominant position in markets that provide essential or strategically important infrastructure.

The concept becomes particularly important where infrastructure is not merely physical. Modern infrastructure may include:

cloud-computing infrastructure;

telecommunications networks;

payment infrastructure;

digital identity systems;

data centres;

logistics networks;

electricity and energy-management systems;

transport and mobility infrastructure;

AI-compute infrastructure;

digital mapping and location infrastructure;

app-store and API infrastructure;

industrial data platforms; and

automated forecasting and allocation systems.

A dominant infrastructure provider may use predictive information to anticipate capacity requirements, customer switching, demand patterns, congestion, investment needs, competitor expansion, pricing responses, and future market entry. Competition concerns arise where that predictive advantage is converted into exclusionary conduct.

The central competition-law question is therefore not whether predictive technology is itself unlawful. Rather, it is:

Whether predictive capabilities, combined with control over infrastructure or strategically important inputs, are being used to exclude competitors, raise entry barriers, discriminate between downstream users, or extend market power into adjacent markets.

2. Meaning of Predictive Infrastructure Market Dominance

Predictive infrastructure market dominance can be understood as a combination of three elements:

A. Infrastructure control

An undertaking controls an infrastructure, network, platform, input, interface, or technological layer that competitors or customers significantly depend upon.

B. Predictive advantage

The undertaking possesses superior information concerning:

demand;

capacity;

network utilisation;

customer behaviour;

switching;

congestion;

investment;

competitor activity;

pricing;

future demand;

technological developments.

C. Strategic use of prediction

The undertaking uses this information to strengthen or protect market power, for example by:

reserving scarce capacity;

discriminatory allocation;

exclusionary pricing;

preferential access;

capacity withholding;

tying;

self-preferencing;

discriminatory interoperability;

strategic investment;

raising rivals' costs; or

preventing potential competitors from achieving sufficient scale.

3. Why Predictive Infrastructure Creates Competition Concerns

Traditional infrastructure dominance was often based upon physical characteristics such as:

ownership of pipelines;

railway tracks;

electricity grids;

ports;

telecommunications networks; or

distribution systems.

Predictive infrastructure introduces an additional source of competitive advantage: informational foresight.

A dominant infrastructure operator may know, for example, that a particular competitor will require additional network capacity six months later. It could potentially use that information to make strategic capacity decisions before the competitor expands.

This creates an important distinction:

Ordinary infrastructure advantage

“I own the infrastructure.”

versus

Predictive infrastructure advantage

“I own the infrastructure and can predict how competitors will use it before they act.”

The second situation can potentially strengthen structural entry barriers.

4. Market Definition

Competition authorities would normally begin by identifying the relevant market.

Several markets may need examination simultaneously.

4.1 Infrastructure market

Examples include:

cloud infrastructure;

telecommunications infrastructure;

payment-processing infrastructure;

data-centre services;

railway infrastructure;

port infrastructure.

4.2 Predictive analytics market

The relevant product may alternatively be:

forecasting software;

AI infrastructure;

demand prediction;

network optimisation;

predictive maintenance.

4.3 Data market

Where the predictive advantage depends upon exclusive access to data, the relevant competitive question may concern:

collection;

access;

aggregation;

portability;

interoperability; and

data sharing.

4.4 Downstream markets

The infrastructure provider may also compete downstream.

For example:

Cloud infrastructure → AI services → enterprise applications.

If the infrastructure operator supplies competitors at the first level but competes with them at the third level, predictive information may provide an opportunity for discriminatory treatment.

5. Dominance and Market Power

Dominance is not established merely because an undertaking possesses sophisticated AI or predictive technology.

Authorities may examine:

market share;

barriers to entry;

economies of scale;

network effects;

switching costs;

interoperability;

access to essential data;

infrastructure investment requirements;

customer dependency;

technological advantages;

vertical integration;

regulatory barriers; and

countervailing buyer power.

The predictive component becomes particularly significant where it is difficult for rivals to reproduce the dominant firm's informational advantage.

6. Predictive Information as a Strategic Competitive Asset

Predictive infrastructure may generate several forms of competitive advantage.

6.1 Demand prediction

A dominant provider can forecast which customers will increase infrastructure consumption.

6.2 Competitor prediction

It may identify which competitors are likely to expand.

6.3 Capacity prediction

It can forecast future shortages and allocate infrastructure capacity accordingly.

6.4 Switching prediction

Predictive systems can identify customers who are likely to switch suppliers.

6.5 Entry prediction

The provider may anticipate the timing and scale of new market entry.

6.6 Investment prediction

The provider can identify where rivals are likely to invest before those investments become publicly visible.

This information may be competitively legitimate when used for efficient planning. It becomes problematic where it facilitates exclusionary strategies.

7. Predictive Capacity Withholding

One important concern is strategic capacity allocation.

Suppose an infrastructure operator predicts that a rival will require additional capacity.

The operator could theoretically:

reserve capacity for itself;

sign long-term contracts with customers;

delay infrastructure expansion;

allocate scarce capacity to affiliated firms;

increase the rival's costs; or

make expansion commercially unattractive.

The competition-law issue would depend on evidence showing that the conduct is exclusionary rather than ordinary commercial capacity management.

8. Predictive Self-Preferencing

Vertical integration creates another concern.

Suppose:

Infrastructure provider → predictive platform → downstream service.

If the infrastructure operator competes downstream, it may use predictive information generated from infrastructure customers to favour its own downstream business.

Potential practices include:

earlier access to capacity;

superior APIs;

preferential latency;

better forecasting information;

preferential network routing;

priority access to infrastructure;

discriminatory technical standards.

This resembles broader self-preferencing concerns in digital markets, although the legal assessment depends on the applicable jurisdiction and conduct.

9. Raising Rivals' Costs Through Predictive Systems

A dominant infrastructure provider can potentially increase competitors' costs by strategically altering:

access fees;

capacity;

technical specifications;

interoperability;

service quality;

authentication requirements;

API access;

network priority.

Predictive analytics can make such strategies more targeted because the dominant undertaking can identify which rivals are most vulnerable.

The relevant theory is often described as raising rivals' costs.

10. Refusal to Supply and Infrastructure Access

Infrastructure markets frequently raise the essential-facilities/refusal-to-deal question.

A refusal to provide access is not automatically abusive merely because competitors desire access.

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

indispensability;

lack of realistic alternatives;

elimination or substantial weakening of competition;

objective justification;

feasibility of access; and

effects on consumers and downstream competition.

Predictive systems may aggravate the problem where the infrastructure owner selectively denies access based upon its forecast of a competitor's future growth.

11. Interoperability and Predictive Infrastructure

Interoperability is especially important in technologically complex infrastructure.

A dominant provider may control:

APIs;

data formats;

authentication;

technical protocols;

cloud interfaces;

payment standards;

identity systems.

If predictive information is combined with control over interoperability, competitors may face a double barrier:

Infrastructure dependency + informational dependency.

This can make switching and entry more difficult.

12. Exclusive Contracts and Predictive Forecasting

Predictive systems can also be used to identify customers whose switching would be particularly important to rivals.

A dominant infrastructure provider could then offer:

long-term exclusivity;

loyalty discounts;

capacity reservations;

minimum-purchase commitments;

rebates;

bundled services.

The competition analysis would depend upon the structure and effects of the arrangement.

The important issue is whether predictive analytics merely improves ordinary commercial targeting or instead facilitates exclusion of equally efficient competitors.

13. Predatory or Strategic Investment

Prediction may also influence infrastructure investment.

A dominant undertaking may know that a competitor needs a particular infrastructure segment to enter.

It could respond by:

rapidly expanding capacity;

constructing redundant infrastructure;

acquiring scarce inputs;

entering the competitor's target region;

signing exclusive contracts;

increasing network density.

Investment itself is normally pro-competitive or efficiency-enhancing. Competition concerns arise only where the evidence establishes an exclusionary strategy and the relevant legal test is satisfied.

14. Network Effects and Predictive Feedback Loops

Predictive infrastructure markets can generate self-reinforcing feedback loops:

More users

↓

More data

↓

Better predictions

↓

Better infrastructure optimisation

↓

Better service

↓

More users

This can produce substantial economies of scale.

Competitors may consequently face a problem known as data-network feedback.

A new entrant may possess a technically comparable algorithm but lack sufficient historical data to generate predictions of equivalent quality.

15. Data as an Entry Barrier

Predictive infrastructure dominance can therefore depend upon data accumulation.

Important questions include:

Is the data exclusive?

Can competitors obtain comparable data?

Is the data generated by customers?

Is data portability available?

Are interoperability standards open?

Can rivals replicate the predictive model?

Are switching costs high?

Does the infrastructure provider combine data from multiple markets?

The competition concern is stronger where data accumulation is difficult to replicate and materially affects competitive performance.

16. Algorithmic Discrimination

A predictive infrastructure system may allocate different conditions to different users.

For example:

Not every differentiation is unlawful. Objective efficiency, risk, congestion, security, and cost differences may justify different treatment.

17. Merger Control and Predictive Infrastructure

Merger control becomes important where a large infrastructure provider acquires:

predictive AI firms;

data analytics companies;

infrastructure startups;

cloud-management platforms;

network optimisation firms;

potential competitors.

The traditional market-share approach may not fully capture the competitive significance of the transaction.

Authorities may examine:

Horizontal effects

Does the transaction eliminate an existing competitor?

Vertical effects

Could the infrastructure provider foreclose downstream competitors?

Data effects

Does the transaction combine uniquely valuable datasets?

Innovation effects

Could the transaction eliminate an emerging predictive technology?

Potential competition

Could the acquired firm have become a significant future competitor?

18. Killer Acquisitions and Predictive Infrastructure

A particularly important scenario arises when an infrastructure incumbent acquires a small company developing a predictive technology.

The startup may have:

low current revenue;

low market share;

substantial technological potential.

Traditional concentration measures could therefore underestimate the transaction's significance.

Authorities may instead examine:

pipeline products;

innovation capabilities;

internal documents;

customer adoption;

technological substitutability;

likely future expansion.

19. Algorithmic Collusion

Predictive infrastructure systems can also create horizontal competition concerns.

Suppose several infrastructure operators use highly sophisticated forecasting systems that continuously observe:

prices;

capacity;

utilisation;

demand;

congestion.

If the systems independently optimise prices, outcomes could become more coordinated.

However:

Parallel algorithmic pricing does not automatically establish an unlawful cartel.

Authorities generally need evidence satisfying the applicable legal test for concerted action, agreement, communication, or other prohibited coordination.

20. Hub-and-Spoke Risks

A common predictive algorithm or infrastructure intermediary can create additional risks.

For example:

Provider A
↓
Common predictive platform
↓
Provider B

If competing firms exchange competitively sensitive information through a common intermediary, questions may arise concerning:

information exchange;

concerted practices;

hub-and-spoke coordination;

price coordination;

capacity coordination.

The fact that software performs the communication rather than human employees does not necessarily eliminate competition-law scrutiny.

21. Relevant Case Laws

1. United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001)

Microsoft concerned Microsoft's dominant position in operating systems and its conduct affecting competition from web browsers.

The case is important for predictive infrastructure because it demonstrates how control over an important technological layer can be used to protect or extend market power into adjacent technological markets.

Relevance:

technological platform dominance;

exclusionary conduct;

leveraging;

interoperability;

strategic control of an infrastructure layer.

It provides an important conceptual foundation for analysing digital infrastructure dominance.

2. United Brands v Commission, Case 27/76 (1978)

The European Court of Justice examined the abuse of a dominant position and discriminatory commercial conditions.

The case is particularly relevant to infrastructure markets because dominance involves more than simply identifying a high market share. The conduct of the dominant undertaking and its effects on trading partners must also be examined.

Relevance to predictive infrastructure:

A dominant infrastructure provider cannot necessarily rely upon its technological sophistication as a justification for discriminatory treatment of dependent customers or competitors.

3. Bronner v Mediaprint, Case C-7/97 (1998)

Bronner concerned access to a newspaper home-delivery system.

The Court established stringent conditions for treating refusal of access to infrastructure as abusive.

The case is especially important for predictive infrastructure because it illustrates the limits of compulsory-access theories.

Relevance:

infrastructure access;

indispensability;

alternative facilities;

refusal to supply;

elimination of competition.

Where predictive infrastructure becomes strategically important, Bronner provides an important framework for determining whether access can legitimately be demanded.

4. IMS Health v NDC Health, Case C-418/01 (2004)

IMS Health concerned access to a commercially important information structure and the relationship between intellectual property and competition law.

The Court identified demanding circumstances in which refusal to license intellectual property could constitute abuse.

Relevance:

Predictive infrastructure may depend upon:

proprietary datasets;

software;

data architectures;

technical standards;

prediction models.

IMS Health therefore illustrates the careful balance between legitimate property rights and competition-law intervention.

5. Microsoft Corp. v Commission, Case T-201/04 (General Court, 2007)

The European Union's Microsoft case concerned Microsoft's refusal to provide interoperability information and the integration of products within its dominant operating-system ecosystem.

The General Court upheld the Commission's intervention concerning interoperability.

Relevance to predictive infrastructure:

It demonstrates the importance of:

interoperability;

technical information;

ecosystem control;

leveraging;

downstream competition.

Where predictive infrastructure is controlled by a dominant undertaking, discriminatory interoperability can become a significant competition issue.

6. Google Shopping, Google and Alphabet v Commission, Case T-612/17 (General Court, 2021)

The Google Shopping litigation concerned the treatment of Google's comparison-shopping service within its search results.

The case is relevant to predictive infrastructure because a platform controlling a key technological access point may potentially use that control to advantage its own downstream service.

Relevance:

self-preferencing;

platform leverage;

control over access;

downstream foreclosure;

data and visibility advantages.

It illustrates why competition authorities may examine the relationship between an infrastructure layer and affiliated downstream services.

7. Slovak Telekom and Deutsche Telekom, Joined Cases C-152/19 P and C-165/19 P (2021)

The Court of Justice addressed exclusionary conduct involving access to telecommunications infrastructure.

The case is particularly relevant because telecommunications networks represent classic infrastructure markets in which access conditions can affect downstream competition.

Relevance:

telecommunications infrastructure;

margin squeeze;

access conditions;

dominant infrastructure provider;

downstream foreclosure.

The case provides an important bridge between traditional network regulation and modern data-driven infrastructure markets.

8. T-Mobile Netherlands, Case C-8/08 (2009)

T-Mobile Netherlands concerned the competition-law treatment of communications between competitors.

The Court emphasised the significance of exchanges capable of reducing strategic uncertainty between competitors.

Relevance to predictive infrastructure:

Modern predictive systems can process large amounts of competitively sensitive information. Where competitors exchange information through algorithms or infrastructure platforms, competition authorities may need to examine whether the system facilitates prohibited coordination.

9. Eturas, Case C-74/14 (2016)

Eturas concerned an electronic platform through which restrictions concerning discounts were communicated to participating travel agencies.

The case is highly relevant to digital infrastructure because it demonstrates that an electronic intermediary can become relevant to competition-law analysis where its system facilitates coordinated conduct.

Relevance:

digital platforms;

algorithmic communication;

information exchange;

coordinated pricing;

intermediary infrastructure.

10. AC-Treuhand, Case C-194/14 P (2015)

AC-Treuhand concerned the liability of an intermediary that facilitated cartel activity.

The case demonstrates that competition-law responsibility can extend beyond firms directly selling the cartelised products where an intermediary intentionally contributes to the implementation of anticompetitive conduct.

Relevance:

Predictive infrastructure providers should therefore consider whether their technology merely provides neutral infrastructure or actively facilitates prohibited coordination.

22. Case-Law Comparison

23. Application Under Indian Competition Law

The Competition Act, 2002 provides several potentially relevant provisions.

Section 3 — Anti-Competitive Agreements

Predictive infrastructure can create concerns involving:

price fixing;

market allocation;

output restrictions;

bid rigging;

information exchange;

coordinated capacity decisions.

Section 3(3) is particularly important for agreements or concerted practices among competitors involving price fixing, market allocation, output restrictions and collusive bidding.

Section 4 — Abuse of Dominant Position

Section 4 becomes particularly relevant where a dominant infrastructure provider:

imposes unfair or discriminatory conditions;

limits or restricts markets;

denies market access;

uses dominance in one market to enter or protect another;

imposes discriminatory access conditions;

engages in exclusionary conduct.

Predictive infrastructure dominance would therefore primarily be analysed through the interaction between market power, infrastructure dependency, predictive advantage, and exclusionary effects.

Sections 5 and 6 — Combinations

Acquisitions involving:

AI infrastructure;

cloud computing;

predictive analytics;

data platforms;

telecommunications infrastructure;

network optimisation;

may raise merger-control concerns where the transaction crosses applicable thresholds and creates or strengthens market power.

Particular attention may be warranted where a dominant infrastructure firm acquires a potential future competitor possessing strategically important predictive technology.

24. Consumer-Welfare Issues

Predictive infrastructure dominance can affect consumers through:

Higher prices

Competitors may face increased infrastructure costs.

Reduced innovation

Potential entrants may be unable to obtain sufficient infrastructure access.

Reduced choice

Exclusion of downstream firms can reduce available services.

Lower quality

A dominant infrastructure provider may have weaker incentives to improve service where rivals are effectively excluded.

Privacy and data concerns

Large-scale predictive systems may depend upon extensive collection and aggregation of user information.

Reduced technological diversity

Infrastructure dominance can encourage a single technical ecosystem.

25. Efficiency Justifications

Predictive infrastructure should not automatically be treated as anticompetitive.

Predictive technology can produce significant efficiencies through:

better capacity planning;

reduced congestion;

lower infrastructure costs;

predictive maintenance;

energy efficiency;

improved network reliability;

fraud prevention;

better resource allocation;

reduced downtime.

For example, a telecommunications operator using AI to predict congestion and dynamically allocate network capacity may substantially improve consumer welfare.

Competition law therefore needs to distinguish:

Predictive efficiency from predictive foreclosure.

26. Key Competition-Law Test

A useful analytical framework is:

Step 1 — Identify the infrastructure

What infrastructure does the undertaking control?

Step 2 — Establish market power

Is the undertaking dominant or otherwise capable of materially affecting competitive conditions?

Step 3 — Identify the predictive advantage

What information does its predictive system generate?

Step 4 — Identify the conduct

How is the predictive capability being used?

Step 5 — Examine foreclosure

Does the conduct make it more difficult for rivals to compete?

Step 6 — Examine effects

Does it affect:

prices;

output;

quality;

innovation;

entry;

consumer choice?

Step 7 — Consider efficiencies

Are there objective technological or economic justifications?

Step 8 — Examine proportionality and remedies

Could competition concerns be addressed through:

non-discriminatory access;

interoperability;

data portability;

transparency;

behavioural commitments;

structural remedies?

27. Possible Competition-Law Remedies

Where unlawful conduct is established, possible remedies could include:

Non-discriminatory access

Infrastructure access should be offered on objectively comparable terms.

Interoperability

Technical interfaces may need to remain accessible.

Data portability

Customers may be allowed to transfer relevant data to competing infrastructure providers.

Separation of information

Sensitive downstream information may be separated from affiliated businesses.

Non-discrimination obligations

The infrastructure provider may be required to apply comparable conditions to similarly situated users.

Monitoring

Independent monitoring can be used where algorithmic systems are difficult for regulators to observe.

Structural remedies

In exceptional circumstances, separation between infrastructure and downstream operations could be considered under applicable law.

28. Emerging Competition Issues

Predictive infrastructure dominance is likely to become increasingly important in:

AI-compute infrastructure

Cloud computing

Data centres

5G and 6G networks

Electricity grids

Electric-vehicle charging

Digital payments

Autonomous transportation

Smart ports

Satellite infrastructure

Digital identity

Industrial IoT

AI model infrastructure

Quantum computing infrastructure

In these markets, competitive power may depend not only on physical assets but also on data, prediction, computing capacity and control over interfaces.

29. Central Legal Tension

The central tension can be expressed as follows:

Predictive infrastructure

→ improves efficiency
→ reduces waste
→ improves capacity utilisation
→ strengthens reliability

but potentially also:

Predictive infrastructure

→ increases informational asymmetry
→ strengthens incumbent advantages
→ enables selective access
→ raises rivals' costs
→ reinforces entry barriers.

Competition law must therefore examine how predictive capability is deployed, rather than treating predictive technology itself as an antitrust violation.

30. Conclusion

Predictive infrastructure market dominance represents an emerging form of competition concern in which traditional infrastructure control is reinforced by advanced forecasting, AI, data accumulation and computational optimisation.

The most important competition-law risks arise where a dominant infrastructure provider combines:

infrastructure control + unique data + predictive capability + vertical integration + exclusionary conduct.

The principal legal issues include refusal to supply, discriminatory access, interoperability restrictions, self-preferencing, raising rivals' costs, exclusive dealing, strategic capacity allocation, algorithmic coordination, data foreclosure and acquisitions of potential predictive competitors.

The cases of Microsoft, Bronner, IMS Health, Microsoft T-201/04, Google Shopping, Slovak Telekom, T-Mobile Netherlands, Eturas and AC-Treuhand demonstrate the principal doctrinal foundations for analysing these issues. Together they show that competition law can address both traditional infrastructure foreclosure and newer forms of technological and informational leverage.

For India, Sections 3 and 4 of the Competition Act, 2002, together with merger-control provisions under Sections 5 and 6, provide the principal statutory framework. The critical task for enforcement will be distinguishing genuine predictive efficiencies from situations where predictive capabilities are strategically deployed to preserve or extend market power.