Competition Law and Governance of Resource Intelligence Systems

1. Introduction

Resource Intelligence Systems (RIS) may be understood as digital or AI-enabled systems that collect, analyse, predict and allocate information about scarce or strategically important resources. These resources may include data, computing capacity, energy, transport capacity, logistics networks, cloud infrastructure, financial liquidity, spectrum, raw materials, labour, inventory, charging infrastructure and platform attention.

Although “Resource Intelligence Systems” is not a settled statutory category in competition law, the concept is highly relevant to modern antitrust because control over the information layer that determines how resources are discovered, priced, prioritised and allocated can create market power.

An RIS can perform functions such as:

• demand forecasting;
• dynamic pricing;
• resource allocation;
• algorithmic ranking;
• capacity optimisation;
• predictive maintenance;
• supply-chain optimisation;
• credit/resource scoring;
• matching buyers and sellers;
• allocation of cloud or computing resources;
• energy-grid optimisation;
• logistics routing;
• inventory allocation; and
• automated recommendations.

The competition-law question is therefore not merely who owns the physical resource, but also who controls the intelligence system that determines access to, prioritisation of, and conditions for using that resource.

The importance of this issue is reinforced by the growing recognition that data and computational resources can function as critical inputs for AI and digital markets, potentially increasing entry barriers and enabling vertical integration.

2. Meaning of Resource Intelligence Systems

A useful analytical model is:

Resources → Data → Intelligence → Prediction → Allocation → Market Outcomes

For example, an energy platform may collect:

• consumption data;
• grid-capacity data;
• weather information;
• generation forecasts;
• battery availability; and
• wholesale prices.

An algorithm may then determine:

• which consumer receives electricity;
• at what price;
• from which supplier;
• at what time;
• with what priority.

The RIS therefore becomes an intermediary between resource availability and competitive access.

Core components

3. Why Resource Intelligence Creates Competition Concerns

A. Control over essential information

A dominant undertaking may possess information that competitors cannot reasonably replicate.

Examples include:

• real-time demand information;
• consumer behaviour;
• supplier performance;
• inventory levels;
• traffic flows;
• energy consumption;
• pricing histories;
• logistics data.

If competitors depend upon such information, withholding or discriminatory provision of it may become a competition concern.

B. Data-driven entry barriers

RIS systems frequently improve as they receive more data.

This can produce a feedback loop:

More users → more data → better predictions → better service → more users → still more data.

Such feedback effects can make entry difficult even where the underlying technology is technically replicable.

The competition issue becomes particularly important where a dominant undertaking combines:

1. a large user base;
2. extensive data;
3. computing infrastructure;
4. algorithms; and
5. distribution channels.

4. Relevant Competition-Law Framework

Resource Intelligence Systems can potentially implicate several traditional competition-law doctrines.

4.1 Abuse of dominance

A dominant RIS operator may engage in:

• discriminatory access;
• refusal to supply;
• self-preferencing;
• tying;
• leveraging;
• exclusionary rebates;
• predatory conduct;
• discriminatory ranking;
• exploitative data practices.

In the EU, Article 102 TFEU provides the principal framework.

In India, Section 4 of the Competition Act 2002 addresses abuse of dominant position.

In China, the Anti-Monopoly Law provides the principal framework, supplemented by rules addressing platform-economy conduct.

4.2 Restrictive agreements

RIS operators may facilitate:

• algorithmic price coordination;
• information exchange;
• allocation of customers;
• exclusionary agreements;
• data-sharing arrangements;
• vertical restrictions.

Algorithms do not automatically immunise conduct from competition law merely because the final decision is automated.

4.3 Merger control

Acquisitions involving RIS systems may create competition concerns where a large undertaking acquires:

• unique datasets;
• AI models;
• cloud capacity;
• optimisation technology;
• logistics intelligence;
• predictive analytics;
• resource-management platforms.

Traditional turnover thresholds may sometimes fail to capture the competitive significance of strategically important data or technology assets, making transaction-value and below-threshold theories increasingly relevant.

4.4 Essential-facility considerations

A resource intelligence platform can potentially become strategically indispensable where competitors cannot realistically reproduce the relevant:

• data;
• infrastructure;
• interface;
• network;
• computational capacity; or
• allocation mechanism.

However, mere usefulness or commercial importance does not automatically establish an essential facility. The established legal requirements must still be examined.

5. Six Major Case Laws

Case 1 — Google Search (Shopping)

Google Search (Shopping), Case AT.39740; Google LLC v European Commission, T-612/17

This is one of the most important authorities for understanding how an intelligence system can affect competitive access.

Google's general-search infrastructure functioned as an important information gateway. The European Commission found that Google favoured its own comparison-shopping service in search results while competing comparison-shopping services were subjected to different ranking treatment.

The General Court substantially upheld the infringement, although it annulled part of the Commission's reasoning concerning certain markets and effects.

Relevance to RIS

The case demonstrates that competition concerns can arise where:

Information infrastructure + algorithmic ranking + market power = competitive foreclosure risk.

The important issue is not merely ownership of a search engine but control over the intelligence architecture determining visibility.

Principle

A dominant information intermediary cannot necessarily use its control over an important algorithmic gateway to systematically favour its own downstream service.

The UK's competition authorities have likewise identified Google Shopping as an important example of algorithmic self-preferencing.

6. Case 2 — Amazon Marketplace

CMA Investigation into Amazon's Marketplace

The UK's Competition and Markets Authority investigated Amazon concerning the use of:

• third-party seller data;
• Buy Box selection;
• Prime eligibility and related criteria.

The CMA ultimately accepted commitments from Amazon and closed the investigation. The investigation therefore should not be described as a final infringement finding.

RIS significance

This case illustrates a particularly important RIS problem:

The platform may possess intelligence concerning competitors while simultaneously competing against those competitors.

Amazon's marketplace can obtain extensive information regarding:

• seller performance;
• product demand;
• pricing;
• inventory;
• consumer behaviour.

If a platform uses competitively sensitive third-party information to optimise its own downstream business, competitors may face a structural disadvantage.

Principle

Competition analysis must consider not only access to the market, but also control over market intelligence generated inside the platform.

7. Case 3 — SAMR v Alibaba

Alibaba Group Holding Limited — SAMR, 2021

China's State Administration for Market Regulation found Alibaba responsible for implementing an exclusivity arrangement commonly described as “choose one from two”, under which merchants were pressured to choose Alibaba rather than competing platforms.

SAMR imposed a RMB 18.228 billion penalty.

RIS significance

Alibaba illustrates the relationship between:

platform intelligence + merchant dependence + network effects + exclusion.

A platform possessing information about merchants and consumers can potentially use that ecosystem position to reinforce exclusivity.

The case is especially relevant to Resource Intelligence Systems because competition may be affected when the operator controls both:

1. the infrastructure through which resources are allocated; and
2. the intelligence generated from transactions occurring through that infrastructure.

Principle

Digital platform power can extend beyond traditional price control to control over commercial relationships, data and access conditions.

8. Case 4 — Shenzhen Weiyuanma v Tencent

Shenzhen Weiyuanma Software Development Co Ltd v Tencent Technology

This Chinese competition case concerned alleged abuse of dominance by Tencent.

The Chinese courts emphasised the importance of properly defining the relevant market for comprehensive internet platforms and distinguishing basic services from value-added services.

RIS significance

This is particularly important for RIS because an intelligence ecosystem may contain numerous interconnected functions.

For example, one system might simultaneously provide:

• data storage;
• analytics;
• recommendation;
• payment;
• logistics;
• cloud computing;
• advertising.

It would therefore be inappropriate automatically to treat the entire ecosystem as one market.

Principle

Competition authorities must examine the particular service affected by the allegedly abusive conduct, while considering the technological and network characteristics of the broader ecosystem.

9. Case 5 — United States v RealPage

United States v RealPage

The RealPage litigation illustrates the competition risks associated with algorithmic pricing systems.

The central issue concerns the use of algorithmic systems incorporating competitively sensitive information to generate rental-price recommendations.

The case is important because the competitive problem may arise not from competitors directly communicating with one another, but from the use of a common algorithmic infrastructure that can facilitate coordinated outcomes.

Recent antitrust discussion has increasingly focused on the possibility that algorithmic pricing systems may facilitate coordination where firms use shared or sensitive competitor information.

RIS significance

This creates a distinctive RIS problem:

Can independent resource-allocation decisions remain genuinely independent when competitors delegate important decisions to the same intelligence infrastructure?

The answer depends upon the facts, including:

• the information supplied to the algorithm;
• the degree of commonality;
• contractual arrangements;
• human involvement;
• communications between competitors;
• algorithm design; and
• actual competitive effects.

Principle

Automation does not by itself remove conduct from antitrust scrutiny.

10. Case 6 — Bronner v Mediaprint

Oscar Bronner GmbH & Co KG v Mediaprint, C-7/97

The European Court of Justice considered whether a dominant undertaking's newspaper-delivery system had to be made available to a competitor under the essential-facilities doctrine.

The Court established demanding conditions for compulsory access.

RIS significance

This case is highly relevant to modern intelligence systems.

Suppose a dominant company controls:

• a unique resource database;
• an indispensable allocation platform;
• an irreplaceable logistics intelligence network; or
• critical computational infrastructure.

The competitor might argue that access is necessary for effective competition.

Bronner demonstrates that commercial usefulness alone is insufficient.

The claimant generally needs to establish circumstances approaching genuine indispensability and satisfy the stringent conditions associated with compulsory access.

Principle

Competition law must balance:

access to indispensable infrastructure

against

the legitimate incentives of infrastructure owners to invest and innovate.

11. Additional Relevant Authority — Aspen Skiing

Aspen Skiing Co v Aspen Highlands Skiing Corp., 472 U.S. 585 (1985)

The U.S. Supreme Court considered a refusal-to-deal situation involving ski-resort ticketing arrangements.

The case remains an important U.S. authority concerning exceptional circumstances in which a dominant firm's termination of a profitable course of dealing can raise monopolisation concerns.

RIS relevance

The principle can become relevant where a dominant intelligence infrastructure previously provides competitors with access to:

• data;
• interfaces;
• allocation services;
• technical interoperability; or
• infrastructure.

A sudden termination of access may require examination of the undertaking's prior conduct, business justification and competitive consequences.

12. Resource Intelligence and Self-Preferencing

One of the most significant risks is self-preferencing.

Consider an RIS controlling the ranking of suppliers:

Supplier A → Supplier B → Supplier C

If the RIS operator owns Supplier A, it may have an incentive to manipulate:

• rankings;
• recommendations;
• access priority;
• resource allocation;
• search visibility;
• capacity reservations.

The competitive problem becomes more serious where competitors cannot independently verify the allocation mechanism.

This is particularly analogous to Google Shopping, where control over the search-ranking architecture was central to the competition analysis.

13. Algorithmic Discrimination

RIS systems can discriminate between market participants through:

Explicit discrimination

The operator directly programs preferential treatment.

Implicit discrimination

The algorithm learns patterns that systematically disadvantage certain competitors.

Data-based discrimination

Certain firms receive better treatment because the system possesses richer information about them.

Dynamic discrimination

The algorithm continuously changes allocation according to observed behaviour.

This raises an important legal issue:

Should competition law regulate the outcome, the algorithm, the underlying data, or all three?

Generally, competition law focuses on competitive conduct and effects rather than simply the existence of an algorithm.

14. Algorithmic Collusion

RIS systems can also facilitate coordination.

Suppose four suppliers use the same intelligence provider:

Supplier A → Algorithm X
Supplier B → Algorithm X
Supplier C → Algorithm X
Supplier D → Algorithm X

If Algorithm X receives extensive competitor-sensitive information and recommends prices or allocation strategies, the system could potentially facilitate coordinated outcomes.

Potential competition-law issues include:

• exchange of competitively sensitive information;
• hub-and-spoke coordination;
• price coordination;
• output restriction;
• market allocation;
• coordinated capacity reductions.

The critical distinction is between:

legitimate independent optimisation

and

coordination facilitated through a common intelligence architecture.

15. Data as a Competitive Resource

Data can itself become a strategically important resource.

A dominant RIS may possess:

• historical data;
• real-time data;
• behavioural data;
• transaction data;
• geospatial data;
• predictive data;
• metadata.

The competitive advantage can become cumulative:

Data → Better prediction → Better allocation → More users → More data.

This creates a potential data-network-effect cycle.

Research concerning AI infrastructure similarly identifies data and computational resources as important inputs capable of increasing entry barriers and facilitating vertical integration.

16. Access to Computational Resources

Modern RIS platforms may require enormous computing capacity.

Control over:

• GPUs;
• cloud computing;
• data centres;
• specialised chips;
• model-training infrastructure;

may therefore become a competition issue.

A dominant firm could potentially disadvantage rivals through:

• discriminatory cloud access;
• preferential computing allocation;
• exclusivity agreements;
• tying;
• capacity reservation;
• discriminatory pricing;
• interoperability restrictions.

Thus, competition analysis may move from market share analysis toward infrastructure-control analysis.

17. Vertical Integration

RIS markets often involve multiple levels:

Hardware

↓

Cloud/Computing

↓

Data

↓

AI/Algorithm

↓

Platform

↓

Distribution

↓

Consumer

Vertical integration may produce efficiencies, but it may also create foreclosure risks.

For example, an undertaking controlling both an AI resource-allocation system and a downstream service could have an incentive to:

• degrade rival access;
• favour its own products;
• obtain rivals' data;
• restrict interoperability;
• bundle services.

18. Interoperability and Data Portability

Interoperability becomes important when an RIS becomes deeply embedded in an ecosystem.

Competition authorities may need to examine:

• API access;
• data portability;
• interoperability;
• technical standards;
• switching costs;
• interface compatibility;
• data export.

The objective is not necessarily to require every system to be interoperable with every competitor.

Rather, the question is whether interoperability restrictions are being used to exclude competitors or reinforce durable market power.

19. Merger Control and Resource Intelligence

RIS acquisitions require analysis beyond conventional market shares.

A transaction may be competitively significant because the target owns:

• a unique dataset;
• a proprietary optimisation model;
• a valuable algorithm;
• a large network;
• strategic computing capacity;
• specialised infrastructure.

A seemingly small technology company may therefore possess a strategically important competitive asset.

Relevant merger theories may include:

Horizontal effects

Two competing intelligence systems merge.

Vertical effects

A resource supplier acquires the intelligence platform controlling allocation.

Conglomerate effects

A large ecosystem acquires a complementary intelligence system.

Data foreclosure

The merged entity prevents rivals from obtaining strategically important information.

20. Governance Mechanisms

Effective governance of RIS should combine competition law with technical controls.

A. Transparency

Operators should maintain records concerning:

• major ranking changes;
• allocation criteria;
• data sources;
• material model changes;
• access decisions.

B. Non-discrimination

Where an RIS provides access to competing businesses, governance systems may establish:

• objective eligibility criteria;
• uniform access conditions;
• audit procedures;
• appeal mechanisms.

C. Data governance

Competition-sensitive data should be separated through:

• access controls;
• data firewalls;
• role-based permissions;
• logging;
• internal compliance systems.

D. Algorithmic auditing

Audits can examine whether algorithms systematically:

• favour affiliated companies;
• disadvantage competitors;
• coordinate prices;
• discriminate in access;
• manipulate rankings.

E. Human oversight

High-impact resource-allocation decisions should not necessarily be completely autonomous.

Governance can require:

Algorithm → Human review → Decision → Audit trail

particularly where the decision significantly affects competitors.

21. Competition Law and Resource Neutrality

An important emerging principle is resource neutrality.

A platform controlling a resource-allocation system should not use that control to transform itself from:

neutral infrastructure provider

into:

gatekeeper + competitor + allocator + data beneficiary.

This is especially important where the platform simultaneously:

1. collects market data;
2. controls access;
3. competes downstream; and
4. determines algorithmically which competitor receives resources.

22. Remedies

Competition authorities could potentially consider:

Structural remedies

• divestiture;
• separation of infrastructure and downstream operations.

Behavioural remedies

• non-discrimination;
• access obligations;
• interoperability;
• data-access requirements.

Algorithmic remedies

• independent auditing;
• algorithmic transparency;
• ranking-neutrality requirements.

Data remedies

• data portability;
• data-sharing safeguards;
• restrictions on use of rival data.

Governance remedies

• compliance monitoring;
• independent trustees;
• reporting obligations.

The appropriate remedy depends upon the actual competition problem and the applicable legal framework.

23. Key Case-Law Principles — Consolidated

24. Emerging Legal Issues

Resource Intelligence Systems are likely to generate several new competition-law questions.

1. Who owns intelligence derived from shared resources?

A platform may argue that its predictions are proprietary even though they are generated from data supplied by market participants.

2. Can competitors collectively use one optimisation system?

Such arrangements may create efficiencies but may also facilitate coordination.

3. Can an RIS operator use competitor-generated data?

The answer may depend upon the nature of the data, contractual arrangements, market power and competitive effects.

4. When does an algorithm become an essential facility?

Technical indispensability, replicability and competitive necessity become central questions.

5. Should algorithmic neutrality be mandatory?

This depends upon market power, legal framework and the nature of the service.

6. How should regulators investigate black-box systems?

Traditional evidence-gathering may need to be supplemented with:

• source-code examination;
• model testing;
• data-flow analysis;
• audit logs;
• simulation;
• counterfactual testing.

25. Conclusion

Resource Intelligence Systems represent a shift from competition over physical resources to competition over the intelligence that controls those resources.

The central competition-law concern is not simply that a company possesses an algorithm or large dataset. The crucial questions are:

1. Does the undertaking possess substantial market power?
2. What resource or information does the RIS control?
3. Is access to that resource important for effective competition?
4. Does the operator compete with the firms dependent upon the system?
5. Does the algorithm discriminate or self-preference?
6. Does the system facilitate coordination?
7. Does data accumulation reinforce market power?
8. Are interoperability or access restrictions exclusionary?
9. Can competitors reasonably replicate the intelligence infrastructure?
10. What remedy would preserve both competition and innovation?

The principal lesson from Google Shopping, Amazon Marketplace, Alibaba, Tencent, RealPage, Bronner and Aspen Skiing is that competition law increasingly has to examine the architecture through which markets are organised, rather than looking only at traditional prices and market shares.