Competition Law and Future Competition Governance for Intelligent Infrastructures

1. Introduction

Intelligent infrastructures are physical and digital infrastructures whose operation increasingly depends upon software, artificial intelligence, sensors, data, cloud computing, automated decision-making and network connectivity. Examples include:

• smart electricity grids and energy-management systems;
• intelligent transport and autonomous mobility infrastructure;
• 5G/6G telecommunications networks;
• cloud and edge-computing infrastructure;
• smart ports, airports and logistics systems;
• digital payment and financial infrastructure;
• AI-enabled healthcare infrastructure;
• connected buildings and smart cities;
• satellite and space communications infrastructure; and
• industrial IoT and automated manufacturing infrastructure.

Competition law traditionally concentrated on physical bottlenecks—railways, electricity networks, telecommunications networks, ports and pipelines. Intelligent infrastructure creates new bottlenecks in software, data, interfaces, algorithms, cloud capacity and interoperability.

The central future competition-law question is therefore not simply who owns the physical infrastructure, but also:

Who controls the intelligent layer through which competitors, users and complementary services obtain access to that infrastructure?

The European Union's current regulatory direction illustrates this shift. The EU has increasingly combined traditional Article 102 TFEU principles with ex-ante digital regulation, interoperability requirements and infrastructure-access regulation. In July 2026, the European Commission adopted binding specification measures concerning Google's Android interoperability for competing AI services and access to Google Search data.

2. Meaning of Intelligent Infrastructure

Intelligent infrastructure can be understood as a system having five interconnected layers:

A. Physical layer

Examples:

• fibre networks;
• electricity grids;
• railway tracks;
• ports;
• charging stations;
• data centres;
• satellites.

B. Connectivity layer

This includes:

• 5G/6G;
• APIs;
• network interconnection;
• cloud connectivity;
• edge computing;
• IoT communications.

C. Data layer

Infrastructure increasingly depends on:

• real-time operational data;
• customer data;
• location data;
• sensor data;
• machine-generated data;
• traffic information;
• energy-consumption data.

D. Intelligence layer

This contains:

• AI models;
• algorithms;
• predictive systems;
• automated allocation systems;
• machine-learning models;
• optimisation software.

E. Interface/ecosystem layer

This includes:

• APIs;
• operating systems;
• application stores;
• digital dashboards;
• authentication systems;
• interoperability protocols.

Competition can be restricted at any of these layers.

3. Why Intelligent Infrastructure Creates New Competition Problems

3.1 Traditional monopoly becomes algorithmic monopoly

A traditional infrastructure monopoly may arise because duplicating a railway, pipeline or electricity network is economically inefficient.

An intelligent infrastructure operator can acquire an additional advantage through:

• proprietary algorithms;
• accumulated data;
• machine-learning feedback loops;
• network effects;
• proprietary APIs;
• technical standards;
• switching costs.

Thus, an infrastructure operator may control both the physical bottleneck and the informational bottleneck.

3.2 Data can become an infrastructure input

Data may be indispensable for:

• AI training;
• predictive maintenance;
• autonomous vehicles;
• energy balancing;
• traffic optimisation;
• fraud detection;
• demand forecasting.

Future competition cases may therefore concern refusal to provide access to commercially important infrastructure data, rather than refusal to provide access to physical infrastructure.

The European Commission's 2026 DMA measures concerning Google Search data illustrate this direction: third-party search engines are being given access to search data that Google can collect at scale.

4. Essential Facilities Doctrine and Intelligent Infrastructure

The traditional essential-facilities doctrine asks whether a dominant undertaking's refusal to provide access to an infrastructure or input can amount to abuse.

Important authorities include:

1. Commercial Solvents v Commission
2. Magill
3. Bronner
4. IMS Health
5. Microsoft
6. Slovak Telekom
7. Lithuanian Railways
8. Alphabet/Android Auto

The doctrine becomes particularly important where intelligent infrastructure is:

• difficult to replicate;
• technically complex;
• network-dependent;
• supported by public investment;
• necessary for interoperability.

5. Case Law

Case 1: Bronner v Mediaprint

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

Principle

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

The infrastructure generally must be indispensable, meaning there must be no actual or potential substitute, and refusal must be capable of eliminating competition while lacking objective justification.

Importance for intelligent infrastructure

Bronner remains an important baseline for privately developed infrastructure.

Future disputes could concern:

• proprietary cloud infrastructure;
• private AI computing infrastructure;
• autonomous transport platforms;
• private charging networks;
• industrial IoT infrastructure.

However, applying Bronner mechanically to every digital platform may be inappropriate where the platform was deliberately designed to accommodate third-party users.

6. Case 2: IMS Health v NDC Health

Case C-418/01, IMS Health GmbH & Co OHG v NDC Health GmbH

The Court dealt with refusal to license an intellectual-property-protected system that was important to competing businesses.

The judgment reinforced the exceptional nature of compulsory access but recognised circumstances in which refusal involving an indispensable input can constitute abuse.

Intelligent-infrastructure significance

Modern infrastructure frequently combines:

• software;
• intellectual property;
• data;
• standards;
• APIs.

Consequently, future cases may involve demands for interoperability information or access to proprietary technical systems.

7. Case 3: Microsoft v Commission

Case T-201/04, Microsoft Corp. v Commission

Microsoft's refusal to provide interoperability information to competing work-group server products was found problematic under Article 102.

The case is particularly significant because interoperability information was treated as an important competitive input.

Relevance today

The same logic has growing importance for:

• cloud interoperability;
• AI-agent interoperability;
• smart-grid systems;
• connected vehicles;
• industrial IoT;
• smart-city platforms.

The European Commission's later competition work expressly identifies Microsoft and other cases when considering access restrictions involving infrastructure, networks, intellectual property and data.

8. Case 4: Slovak Telekom

Case C-165/19 P, Slovak Telekom a.s. v Commission

The case concerned access to telecommunications infrastructure and wholesale broadband services.

The Court clarified the relationship between Article 102 TFEU and sector-specific regulation.

Principle

Where regulation already requires access, competition law need not necessarily apply the exceptionally strict Bronner conditions in the same way.

Intelligent-infrastructure significance

This is crucial for:

• fibre networks;
• 5G infrastructure;
• cloud connectivity;
• electricity-grid access;
• smart transport infrastructure.

Where legislators have already established an access regime, competition law and sectoral regulation can operate together.

9. Case 5: Lithuanian Railways

Case C-42/21 P, Lietuvos geležinkeliai AB v Commission

The case involved removal of railway infrastructure that could have been used by competitors.

The Court accepted that the circumstances surrounding infrastructure developed with public funds and the deliberate destruction of infrastructure can affect the application of the Bronner framework.

Intelligent-infrastructure significance

The case is particularly relevant to infrastructure that has:

• substantial public financing;
• strategic significance;
• limited replicability;
• long asset lives.

Future disputes could involve the deliberate disabling of:

• interoperability interfaces;
• network connections;
• APIs;
• data-access mechanisms;
• charging infrastructure;
• cloud interfaces.

10. Case 6: Alphabet/Android Auto

Case C-233/23, Alphabet and Others v AGCM, judgment of 25 February 2025

This is one of the most directly relevant cases for intelligent infrastructure.

Google's Android Auto platform allowed third-party applications to interact with vehicle infotainment systems, but a third-party app sought interoperability that Google had not provided.

The Court held that refusal to ensure interoperability could potentially constitute abuse even though Android Auto was not indispensable in the strict Bronner sense.

Major development

The Court distinguished a digital platform that was created only for the dominant firm's internal needs from one that was designed to accommodate third-party applications.

This means that where a platform is open to third-party complementors, refusal to provide interoperability can be assessed differently from refusal to provide access to a purely internal infrastructure.

Importance

This principle is highly relevant to:

• autonomous vehicles;
• smart homes;
• AI assistants;
• smart-city systems;
• connected appliances;
• digital healthcare;
• industrial platforms.

The Android Auto decision therefore represents an important evolution of essential-facilities reasoning into the intelligent-platform environment.

11. Case 7: Google Shopping

Case C-48/22 P / Google Shopping

The Google Shopping litigation concerned preferential treatment by Google's search infrastructure.

The case demonstrates that competition problems can arise not only from refusal of access but also from preferential treatment within an infrastructure controlled by a dominant undertaking.

This is relevant to intelligent infrastructure because algorithms can determine:

• ranking;
• visibility;
• access;
• resource allocation;
• routing;
• recommendation;
• priority.

A future infrastructure operator could therefore distort competition without explicitly excluding a competitor.

12. Case 8: Intel/McAfee

The Intel/McAfee merger review is particularly relevant to interoperability.

The Commission examined whether the integration of security software with Intel's dominant hardware position could affect competing security solutions and adopted interoperability commitments.

The Commission's digital-merger materials identify Intel/McAfee as an example of concern about interoperability degradation.

Future significance

This demonstrates why intelligent infrastructure merger review cannot stop at conventional market shares.

Authorities may need to assess:

• software-hardware integration;
• access to technical interfaces;
• data advantages;
• ecosystem foreclosure;
• compatibility degradation.

13. From Essential Facilities to Essential Intelligence

A major future development is the movement from:

essential physical facility

to

essential digital facility

and ultimately:

essential intelligence facility.

For example:

The competitive bottleneck may therefore be the decision-making system controlling the infrastructure.

14. Interoperability as a Future Competition Remedy

Interoperability may become one of the most important competition remedies.

Possible requirements include:

Technical interoperability

Systems must technically communicate.

Functional interoperability

Third-party products must be able to perform comparable functions.

Data interoperability

Users and competitors can access relevant data.

API interoperability

Dominant infrastructure operators must provide effective API access.

AI interoperability

Third-party AI systems may need access to operating-system capabilities.

The EU's 2026 DMA proceedings concerning Android are a current illustration: the Commission required Google to provide effective interoperability to competing AI services with relevant Android features.

15. Data Access and Competition Governance

Future competition authorities may have to determine:

1. Who owns infrastructure-generated data?
2. Who controls access?
3. Can competitors use the data?
4. Can customers transfer it?
5. Is data portability sufficient?
6. Can the dominant operator use competitors' data?
7. Can the operator combine infrastructure data with platform data?

The competition problem becomes especially serious where:

more users → more data → better AI → better service → more users.

This can produce a self-reinforcing feedback loop.

16. Algorithmic Discrimination

Intelligent infrastructure can automatically discriminate between competitors.

For example, an infrastructure algorithm might:

• give affiliated services faster access;
• reduce competitors' network priority;
• increase competitors' latency;
• provide better API functionality to affiliated businesses;
• rank affiliated services higher;
• allocate scarce electricity capacity preferentially;
• provide better charging availability to affiliated vehicles.

This creates a new category of potentially abusive conduct:

algorithmic infrastructure discrimination.

Competition authorities may therefore need access to:

• source-code documentation;
• model documentation;
• audit logs;
• API specifications;
• decision rules;
• training-data descriptions;
• performance measurements.

17. Self-Preferencing in Intelligent Infrastructure

Self-preferencing may occur where an infrastructure owner operates downstream services.

Examples:

Smart-grid operator → electricity retailer

Cloud provider → cloud-based AI service

Vehicle platform → affiliated mobility service

Smart-city platform → affiliated service provider

Airport platform → affiliated logistics company

The competition question is whether the infrastructure controller is using control over an upstream bottleneck to disadvantage downstream competitors.

18. Vertical Integration

Intelligent infrastructures encourage vertical integration because one undertaking may control:

hardware → connectivity → data → AI → application → customer interface.

This can produce efficiencies but may also create foreclosure risks.

Merger authorities may therefore examine:

• access foreclosure;
• interoperability degradation;
• data foreclosure;
• ecosystem effects;
• switching costs;
• network effects;
• cross-market leverage.

19. Network Effects

Intelligent infrastructure often has strong network effects.

For example:

More users
↓
More operational data
↓
Better algorithm
↓
Better service
↓
More users

This can create rapid concentration.

Competition governance must therefore examine dynamic competition, not only current market shares.

20. Public Infrastructure and Competition

Many intelligent infrastructures will be partly publicly funded.

Examples include:

• national broadband;
• smart electricity grids;
• public transport systems;
• government cloud;
• public digital identity;
• smart-city infrastructure.

Public investment creates a special competition issue:

Should publicly financed infrastructure be accessible to competing commercial operators on transparent and non-discriminatory terms?

EU infrastructure policy increasingly emphasises access, interconnection and interoperability, including access to physical network elements, technical interfaces and operational-support systems.

21. Future Governance Model

A future intelligent-infrastructure competition framework could contain eight pillars.

Pillar 1 — Market definition

Authorities should identify:

• infrastructure markets;
• data markets;
• interoperability markets;
• AI markets;
• downstream markets.

Pillar 2 — Infrastructure-access rules

Dominant infrastructure operators could face:

• FRAND access;
• non-discrimination;
• transparent pricing;
• access timelines;
• technical-access obligations.

Pillar 3 — Interoperability

Regulation should address:

• APIs;
• protocols;
• data portability;
• technical standards;
• AI-agent interoperability.

Pillar 4 — Data governance

Rules should regulate:

• access;
• portability;
• sharing;
• aggregation;
• use restrictions.

Pillar 5 — Algorithmic neutrality

Authorities should investigate whether algorithms systematically favour affiliated services.

Pillar 6 — Merger control

Authorities should consider:

• data accumulation;
• ecosystem expansion;
• interoperability;
• future competition;
• vertical foreclosure.

Pillar 7 — Continuous monitoring

Traditional competition enforcement is often retrospective.

Intelligent infrastructure may require:

• continuous monitoring;
• technical audits;
• interoperability testing;
• algorithmic compliance reports.

Pillar 8 — Regulatory coordination

Competition authorities should coordinate with:

• telecom regulators;
• energy regulators;
• transport regulators;
• data-protection authorities;
• cybersecurity authorities;
• AI regulators.

22. Ex-Ante and Ex-Post Regulation

Future governance will probably combine both.

Ex-post competition law

Used after problematic conduct occurs.

Examples:

• Article 102 TFEU;
• abuse of dominance;
• exclusionary conduct;
• discriminatory access.

Ex-ante regulation

Used before competitive harm becomes entrenched.

Examples:

• interoperability requirements;
• access obligations;
• data-sharing rules;
• platform obligations;
• technical standards.

The EU's Digital Markets Act represents an important example of this ex-ante approach.

23. Competition and Security

Intelligent infrastructure creates an additional tension between competition and cybersecurity.

A dominant operator may argue that interoperability creates:

• cybersecurity risks;
• privacy risks;
• operational instability;
• safety risks.

Such concerns can be legitimate.

Therefore, future competition law should distinguish between:

genuine security justification

and

security justification used as a pretext for exclusion.

This makes technical evidence and independent auditing particularly important.

24. Competition and Innovation

Infrastructure regulation must also preserve investment incentives.

If access is imposed too aggressively, infrastructure operators may argue that:

• investment incentives decline;
• innovation is reduced;
• infrastructure development becomes less attractive.

Conversely, insufficient access may allow a dominant infrastructure operator to entrench its position.

The appropriate framework therefore requires balancing:

access + innovation + investment + consumer welfare + security.

25. Future Role of Competition Authorities

Competition authorities may increasingly need technological capabilities in:

• AI;
• machine learning;
• cloud architecture;
• cybersecurity;
• telecommunications;
• distributed systems;
• data analytics.

The future competition investigator may need to understand not merely contracts and economics but also:

how an algorithm actually allocates access to infrastructure.

26. Remedies for Intelligent Infrastructure

Potential remedies include:

Structural remedies

• divestiture;
• separation of infrastructure and downstream businesses;
• functional separation.

Behavioural remedies

• non-discrimination;
• access obligations;
• transparent pricing;
• interoperability;
• data portability.

Technical remedies

• API access;
• open standards;
• interface publication;
• interoperability testing;
• algorithmic audits.

Governance remedies

• independent compliance monitors;
• periodic reporting;
• technical certification;
• regulatory sandboxes.

27. Major Future Competition Risks

The principal risks can be summarised as follows:

1. Infrastructure foreclosure
2. Algorithmic discrimination
3. Self-preferencing
4. Data foreclosure
5. Interoperability restrictions
6. Cloud dependency
7. AI ecosystem lock-in
8. Network-effect-driven concentration
9. Predatory or exclusionary algorithmic pricing
10. Vertical integration
11. Acquisitions of emerging competitors
12. Control over technical standards
13. Cybersecurity-based exclusion
14. Switching-cost exploitation
15. Control over autonomous decision-making

28. Conceptual Future Test

A useful future analytical framework could be:

Is the infrastructure important?
↓
Is access technically or economically difficult to replicate?
↓
Does the undertaking control a bottleneck?
↓
Does it also control data, algorithms or interfaces?
↓
Are competitors dependent upon the system?
↓
Is access discriminatory or selectively restricted?
↓
Does the conduct affect downstream competition?
↓
Is there an objective technical, security or efficiency justification?
↓
Can interoperability or access remedies preserve competition without undermining investment?

29. Relationship Between the Major Cases

The Alphabet/Android Auto judgment is especially significant for the future because the Court recognised that a digital platform opened to third-party undertakings should not necessarily be analysed under the strictest Bronner conditions.

30. Conclusion

Competition law for intelligent infrastructures is evolving from a relatively simple question of physical access toward a much broader governance problem involving:

physical infrastructure + networks + data + algorithms + AI + interoperability + ecosystems.

The classical essential-facilities doctrine remains important, particularly through Bronner, IMS Health, Microsoft, Slovak Telekom and Lithuanian Railways. But Alphabet/Android Auto demonstrates how competition law is adapting when infrastructure is a digital platform designed to interact with third-party services.

The future framework is therefore likely to combine:

Article 102-type abuse control

• sector-specific access regulation
• ex-ante digital regulation
• interoperability obligations
• data-access governance
• algorithmic oversight
• merger control
• technical auditing.

The central principle for future competition governance should be to ensure that control over an intelligent infrastructure does not automatically become control over every market that depends upon that infrastructure. At the same time, access obligations must account for legitimate investment, innovation, cybersecurity, privacy and safety considerations.

In this sense, the future of competition law for intelligent infrastructures is likely to move from merely policing market power toward governing technological control points—the physical, digital, data and algorithmic interfaces through which entire ecosystems compete.