1. Introduction

Earth Observation (EO) data platforms collect, process, aggregate, distribute, and commercialise information obtained from satellites, aerial sensors, synthetic-aperture radar (SAR), multispectral imagery, hyperspectral sensors, weather instruments, and other remote-sensing systems. Modern EO platforms increasingly combine raw satellite imagery with AI analytics, cloud computing, geospatial databases, APIs, digital twins, and proprietary datasets.

Competition concerns arise when a platform becomes an important gateway between satellite operators, data providers, governments, businesses, researchers, and downstream users. Market power may arise not merely from ownership of satellites, but from control over data access, historical archives, processing infrastructure, APIs, analytical models, interoperability standards, and distribution channels.

The central competition-law question is therefore:

When does an Earth Observation data platform cease to be merely an efficient data intermediary and become an infrastructure bottleneck capable of excluding competitors or exploiting dependent users?

2. Relevant Markets

Several overlapping markets may need to be considered.

A. Raw EO data

This includes:

• optical satellite imagery;
• SAR imagery;
• multispectral and hyperspectral imagery;
• thermal imagery;
• weather and atmospheric observations;
• high-resolution commercial imagery.

A platform controlling unusually valuable imagery may possess significant bargaining power.

B. Processed EO data

The platform may transform raw observations into:

• orthorectified imagery;
• mosaics;
• vegetation indices;
• elevation models;
• change-detection datasets;
• climate datasets;
• maritime intelligence;
• agricultural intelligence.

Processed data can become more difficult to replicate than raw observations.

C. EO analytics

Platforms may provide AI-based services such as:

• object detection;
• crop classification;
• infrastructure monitoring;
• deforestation detection;
• disaster mapping;
• emissions monitoring;
• maritime surveillance.

Here, competition may occur at the analytics layer, even where underlying satellite data are available elsewhere.

D. EO data infrastructure

A platform can also compete in:

• cloud storage;
• data catalogues;
• API access;
• computing;
• geospatial databases;
• model hosting;
• data-processing pipelines.

This creates the possibility of vertical leverage from infrastructure into downstream EO markets.

3. Major Competition Issues

3.1 Control Over Unique or Non-Replicable EO Data

The first concern is control over data that competitors cannot reasonably reproduce.

A platform may possess:

• decades of historical imagery;
• proprietary high-resolution observations;
• frequently refreshed datasets;
• unique geographic coverage;
• calibrated sensor data;
• labelled datasets for AI training.

If competitors require access to such data to compete effectively, refusal or discriminatory access can become a competition concern.

The issue resembles the essential-facility doctrine, although courts generally apply that doctrine cautiously.

The relevant questions include:

1. Is the data genuinely indispensable?
2. Can competitors obtain equivalent data elsewhere?
3. Is duplication technically possible?
4. Is duplication economically realistic?
5. Does access require unreasonable investment?
6. Is the platform itself competing downstream?

4. Exclusive Data Agreements

An EO platform may negotiate exclusive arrangements with:

• satellite operators;
• governments;
• launch providers;
• sensor manufacturers;
• data aggregators.

Exclusive agreements may create efficiencies by financing satellite deployment or guaranteeing revenue.

However, extensive exclusivity can foreclose competing platforms.

For example, if the dominant EO distributor obtains exclusive rights to distribute the most commercially valuable imagery for an entire region, rival platforms may be unable to achieve sufficient scale.

Competition authorities would therefore examine:

• duration;
• geographic scope;
• market coverage;
• switching possibilities;
• alternative suppliers;
• exclusivity rebates;
• contractual termination rights.

5. Data Aggregation and Network Effects

EO platforms become more valuable when they aggregate multiple datasets.

For example:

Satellite imagery + weather + terrain + traffic + agriculture + historical imagery + AI models

may produce a substantially more valuable product than any individual dataset.

This can generate data-driven network effects.

More users produce:

• more demand;
• more commercial revenue;
• more labelled data;
• more feedback;
• better algorithms;
• more integrations.

Those improvements attract still more users.

The result can be a self-reinforcing competitive advantage.

6. Data Portability and Switching Costs

Customers may become dependent on a platform because their workflows are built around its:

• APIs;
• proprietary formats;
• metadata structures;
• cloud architecture;
• analytical models;
• dashboards;
• historical archives.

Switching may therefore require:

• data migration;
• software redevelopment;
• retraining;
• revalidation of analytical models;
• API replacement;
• employee retraining.

High switching costs can create customer lock-in even when nominally competing services exist.

Competition authorities may consequently consider:

• interoperability obligations;
• standardized APIs;
• export rights;
• machine-readable data formats;
• reasonable migration periods.

7. Self-Preferencing

A vertically integrated EO platform might supply both:

1. EO data/infrastructure; and
2. downstream analytical products.

It could potentially prefer its own downstream services.

Examples include:

• ranking its own analytics above rivals;
• giving its own applications faster API access;
• providing its own subsidiary with better-resolution data;
• giving affiliated services preferential cloud processing;
• withholding metadata from competitors.

This resembles competition concerns arising in digital-platform markets generally.

8. Discriminatory Access

A dominant platform may provide different:

• prices;
• API limits;
• processing speeds;
• data resolutions;
• licensing terms;
• refresh frequencies

to different customers.

Discrimination becomes particularly problematic where the platform supplies a critical input to downstream competitors.

For example:

Platform A supplies high-resolution satellite data to several analytics firms but gives its affiliated analytics division substantially better refresh rates and API quotas.

That could potentially disadvantage independent downstream providers.

9. Predatory Pricing and Cross-Subsidisation

A vertically integrated platform might offer EO data below cost while recovering losses through another business.

Potential strategies include:

• free satellite-data APIs;
• heavily discounted processing;
• bundled cloud storage;
• zero-price access for selected users;
• below-cost analytics.

Low prices are not inherently unlawful.

The competition concern arises where pricing is strategically designed to:

1. eliminate competitors;
2. prevent entry;
3. create dependency; and
4. permit later exploitation.

10. Bundling and Tying

EO platforms may bundle:

Satellite imagery + cloud processing + AI analytics + storage + API access

into one contract.

Bundling may generate substantial efficiencies.

However, a dominant platform might use control over an indispensable EO input to force customers to purchase unrelated services.

For example:

A dominant satellite-data provider requires customers purchasing high-resolution imagery to use its proprietary cloud-processing service.

If competitors are excluded from the downstream processing market, tying concerns may arise.

11. API and Interoperability Restrictions

APIs increasingly constitute competitive infrastructure.

A platform may technically permit access while imposing:

• restrictive rate limits;
• excessive authentication requirements;
• discriminatory API latency;
• limited historical queries;
• incompatible formats;
• restrictive terms of use.

Thus, formal access does not necessarily equal effective access.

Competition analysis should consider whether API restrictions materially increase rivals' costs.

12. Algorithmic Advantages

EO platforms increasingly use AI to transform imagery into commercially valuable information.

A dominant platform may benefit from:

more data → better models → better predictions → more customers → more data.

This feedback loop may make entry progressively more difficult.

Potential competition concerns include:

• exclusive AI-training datasets;
• preferential access to labelled imagery;
• discriminatory model access;
• refusal to provide model outputs;
• proprietary classification systems;
• interoperability barriers.

13. Government Data and Public Resources

EO markets frequently involve governments.

Public authorities may own or finance:

• satellite systems;
• weather observations;
• mapping databases;
• environmental datasets;
• defence-related imagery;
• public-sector geospatial information.

Competition problems can arise if a private platform obtains preferential access to public data and then uses it to compete against other commercial providers.

This raises questions of:

• non-discrimination;
• transparent licensing;
• public-data access;
• state aid;
• procurement;
• public-sector neutrality.

14. Merger and Acquisition Concerns

EO consolidation may involve acquisitions of:

• satellite operators;
• imagery providers;
• geospatial analytics companies;
• cloud providers;
• AI firms;
• data brokers.

A transaction may remove an important emerging competitor even if the target has relatively little current revenue.

Authorities should therefore consider:

• data concentration;
• innovation competition;
• future competition;
• access to unique datasets;
• vertical foreclosure;
• interoperability;
• AI-training advantages.

15. Case Laws

Case 1 — United Brands Company v Commission (1978)

The European Court of Justice examined dominance and the concept of an undertaking's ability to behave independently of competitors and customers.

Relevance to EO platforms

The principle is useful where an EO platform controls a commercially important data ecosystem.

A platform's dominance cannot be assessed solely by looking at its market share. Authorities can examine:

• customer dependence;
• alternative suppliers;
• barriers to entry;
• infrastructure advantages;
• bargaining power.

EO application: A platform controlling a uniquely valuable EO dataset may possess significant market power even where several smaller imagery providers remain active.

Case 2 — Magill — RTE and ITP v Commission (1995)

The Court considered refusal to license information and established important principles concerning exceptional circumstances in which refusal to supply intellectual-property-protected material may constitute abuse.

Relevance

The case is particularly important for EO data platforms because satellite imagery, databases, metadata, and analytical datasets can involve intellectual-property rights.

The case illustrates that:

Intellectual-property protection does not automatically immunise exclusionary conduct from competition law.

However, intervention requires demanding conditions.

EO application: If a dominant platform controls indispensable data that competitors require to develop competing products, refusal to license may attract scrutiny where the exceptional Magill conditions are satisfied.

Case 3 — Bronner v Mediaprint (1998)

The Court imposed a demanding test for treating infrastructure as indispensable under Article 102.

Relevance

The case is highly relevant to EO infrastructure.

A competitor cannot simply argue:

"It is expensive or inconvenient to build another satellite-data platform."

The stronger question is whether duplication is impossible or economically unrealistic and whether there is no viable alternative.

EO application: A rival would have a stronger case if a dominant platform controlled unique historical imagery that could not realistically be recreated because the underlying satellite observations occurred years earlier.

Case 4 — IMS Health v NDC Health (2004)

The Court further developed the exceptional circumstances surrounding compulsory access to protected information and interoperability-type structures.

Relevance to EO

EO data platforms can possess highly structured datasets that become industry standards.

For example:

• proprietary geographic grids;
• standardized metadata;
• historical databases;
• analytical classifications.

If an entire downstream industry has become dependent upon one proprietary data architecture, refusal of access may generate competition concerns.

The case demonstrates that commercial importance alone is insufficient; the legal threshold remains demanding.

Case 5 — Microsoft v Commission (2007)

The General Court upheld findings involving Microsoft's refusal to provide interoperability information and its tying conduct.

Relevance to EO platforms

This is one of the most useful analogies for technologically integrated EO platforms.

An EO platform may control:

• data;
• APIs;
• processing infrastructure;
• proprietary software;
• interoperability protocols.

If rivals cannot effectively interoperate with the platform, competition may be weakened.

EO application: A dominant EO infrastructure provider could face scrutiny if it intentionally withholds technical information necessary for competing analytics systems to interoperate with its platform.

Case 6 — Google Shopping — Google and Alphabet v Commission (2024)

The EU courts considered Google's preferential treatment of its own comparison-shopping service through its general search results.

Relevance to EO

The case is highly relevant to self-preferencing.

Suppose an EO platform operates:

1. a data marketplace; and
2. its own downstream environmental analytics service.

It could potentially manipulate:

• search rankings;
• API responses;
• marketplace placement;
• recommendation systems;
• default analytical tools

to favour its own downstream products.

The Google Shopping litigation demonstrates how discriminatory treatment within a platform ecosystem can be assessed under abuse-of-dominance principles.

Case 7 — Slovak Telekom v Commission (2021)

The Court considered exclusionary conduct involving access to telecommunications infrastructure.

Relevance to EO platforms

Telecommunications infrastructure and EO infrastructure are not identical, but the case provides a useful framework for analysing access to strategically important infrastructure.

For EO platforms, relevant infrastructure could include:

• data repositories;
• APIs;
• processing systems;
• distribution networks;
• satellite ground infrastructure.

The central question is whether access conditions allow effective competition downstream.

Case 8 — Servizio Elettrico Nazionale v Autorità Garante della Concorrenza e del Mercato (2022)

The Court addressed abuse of dominance and the use of advantages obtained from a former legal monopoly.

Relevance to EO

This is particularly useful where an EO platform has grown from a privileged public-sector position.

A platform may have obtained advantages through:

• government contracts;
• public satellite infrastructure;
• privileged datasets;
• regulatory access;
• public-sector information.

The case illustrates that advantages derived from a historically protected position may become competition concerns when leveraged into competitive markets.

16. Competition-Law Risk Matrix

17. How Competition Authorities Should Analyse EO Platforms

A modern investigation should examine at least five layers:

Layer 1 — Data

Who controls the underlying observations?

Layer 2 — Infrastructure

Who controls storage, processing and API infrastructure?

Layer 3 — Analytics

Who controls AI models and analytical outputs?

Layer 4 — Distribution

Who controls access to customers and marketplaces?

Layer 5 — Ecosystem

Does one undertaking control several layers simultaneously?

This last question is particularly important.

A company controlling only satellite imagery may face ordinary competition.

A company controlling:

satellites → historical data → cloud processing → AI models → API → marketplace → downstream applications

may possess substantially greater structural power.

18. Remedies

Potential remedies include:

Structural remedies

• divestiture;
• separation of businesses;
• limits on vertical integration.

Behavioural remedies

• non-discriminatory access;
• FRAND-style licensing;
• API interoperability;
• data portability;
• transparent ranking;
• prohibition of self-preferencing.

Data remedies

• access to historical datasets;
• standardized metadata;
• machine-readable exports;
• portability rights;
• reasonable licensing.

Merger remedies

• data-access commitments;
• interoperability commitments;
• restrictions on exclusive contracts;
• divestiture of overlapping analytics businesses.

19. Key Legal Principle

The central competition-law distinction is between legitimate technological integration and strategic exclusion.

An EO platform should generally be allowed to:

• innovate;
• integrate data;
• develop proprietary AI;
• negotiate commercial contracts;
• bundle complementary products;
• charge different prices where objectively justified.

The competition problem emerges where market power is used to foreclose rivals, raise their costs, prevent interoperability, deny indispensable inputs, or extend dominance from one EO layer into another.

20. Conclusion

Earth Observation data platforms represent a new form of data-intensive digital infrastructure. Their competitive significance extends beyond satellite imagery because control over EO data can provide advantages in AI training, analytics, cloud processing, environmental intelligence, agriculture, insurance, defence, logistics, and climate services.

The most important competition issues are therefore:

1. control of unique EO datasets;
2. refusal or discriminatory access;
3. exclusive data arrangements;
4. data-driven network effects;
5. platform lock-in and switching costs;
6. API and interoperability restrictions;
7. self-preferencing;
8. bundling and tying;
9. predatory or exclusionary pricing;
10. vertical integration and data leverage; and
11. concentration through mergers and acquisitions.

The cases of United Brands, Magill, Bronner, IMS Health, Microsoft, Google Shopping, Slovak Telekom, and Servizio Elettrico Nazionale collectively provide a useful legal framework for analysing these problems. The most important lesson is that data ownership alone does not establish dominance or abuse; competition analysis must establish market power, indispensability or competitive significance, exclusionary effects, and the absence or inadequacy of legitimate alternatives.