The Digital Markets Act: Reshaping Tech Competition and Machine Learning Observability in the EU

Executive Summary

The European Commission designated six major tech companies—Alphabet, Amazon, Apple, ByteDance, Meta, and Microsoft—as “gatekeepers” under the Digital Markets Act in September 2023, fundamentally transforming how digital platforms operate in the European Union. As enforcement began in March 2024, the DMA has emerged as one of the most consequential regulatory frameworks in the digital economy, with profound implications for competition, innovation, and emerging technologies like machine learning observability systems.

This analysis examines the multidimensional impacts of the DMA, explores its ripple effects across economic, technological, and geopolitical dimensions, and proposes strategic adaptations for machine learning observability platforms navigating this new regulatory landscape.

I. Understanding the Digital Markets Act: Framework and Objectives

The Regulatory Architecture

The DMA targets companies that hold considerable market power and provide at least one core platform service, with gatekeepers required to comply with new EU requirements by March 2024. The legislation establishes three criteria for gatekeeper designation: significant impact on the internal market, provision of services acting as important gateways between business users and end users, and an entrenched, durable market position.

To qualify as a gatekeeper, undertakings must generate EU revenues of at least €7.5 billion in each of the last three financial years or have an average market capitalization of at least €75 billion, provide core platform services in at least three Member States, and maintain at least 45 million EU monthly active end users and 10,000 yearly active EU business users.

The twenty-two designated core platform services span eight sectors: online search engines, app stores, social networking services, video-sharing platforms, messenger services, online advertising, operating systems, and web browsers. In April 2024, the Commission designated Apple’s iPadOS as a gatekeeper, and in May 2024, Booking was designated as a gatekeeper for its online intermediation service.

Key Obligations and Prohibitions

The DMA imposes both positive obligations and negative prohibitions on gatekeepers:

Obligations include:

• Allowing third-party interoperability with gatekeeper services
• Providing business users access to data generated on the platform
• Enabling app developers to freely steer consumers to alternative offers
• Offering tools for advertisers to conduct independent verification
• Permitting users to uninstall pre-installed software
• Allowing choice of default services (browsers, search engines)

Prohibitions include:

• Self-preferencing own services in rankings over competitors
• Combining personal data across services without explicit consent
• Preventing users from linking to businesses outside the platform
• Tying access to one service to subscription of another

If gatekeepers fail to comply with the DMA, they can face fines of up to 10% of their global revenues—a penalty that can rise to 20% for repeat offenders. For systematic infringements, the Commission reserves authority to impose structural remedies, including forced divestiture.

II. Economic Ramifications: Competition, Innovation, and Market Dynamics

Immediate Market Impacts

The DMA’s enforcement has produced tangible market shifts. In April 2024, Reuters reported that in the first month after regulations were implemented, independent browsers saw significant user increases—Cyprus-based Aloha Browser reported users in the EU jumped 250% in March, while Norway-based Vivaldi, Germany-based Ecosia, and US-based Brave also experienced rising user numbers.

As of March 2024, merchants can choose an external payment service provider to take payments through apps on Apple devices, with Apple’s commission on in-app transactions for digital goods and services reduced to 17% under the new scheme. This represents a substantial reduction from the previous 30% commission rate, potentially saving developers billions in fees annually.

Compliance Costs and Business Model Disruption

The financial burden of compliance extends beyond potential fines. Major gatekeepers have invested hundreds of millions of euros in technical infrastructure changes, legal consultations, and compliance monitoring systems. Experts expect three major changes: the ability to sideload apps outside designated app stores, increased interoperability requirements, and fundamental alterations to data sharing practices.

However, compliance costs disproportionately impact different stakeholders:

For Gatekeepers: Technical restructuring, legal teams, compliance reporting systems, and potential revenue losses from reduced commissions and eliminated self-preferencing practices.

For Developers: New opportunities to reach customers directly, but complexity in navigating multiple compliance frameworks across jurisdictions.

For Consumers: Potential benefits from increased choice and lower prices, but also concerns about security risks from alternative app stores and confusion from proliferated options.

Innovation Concerns and Product Delays

The DMA has led to unintended consequences, including gatekeeper reluctance to roll out new services in the EU—for example, Meta’s Threads was introduced in the EU five months after its global release due to data-processing concerns, and certain product enhancements like iPhone Mirroring were not released to EU consumers over compliance concerns.

This phenomenon, dubbed “regulatory chill,” raises questions about whether the DMA’s ex-ante approach appropriately balances competition protection with innovation incentives. Some economists argue that pre-emptive regulations risk prohibiting potentially beneficial practices before market impacts can be assessed.

III. Enforcement Actions: Testing the DMA’s Teeth

Early Investigations and Penalties

On March 25, 2024, the Commission opened non-compliance investigations against Alphabet, Apple, and Meta concerning rules on steering in Google Play and self-preferencing in Google Search, Apple’s rules on steering in the App Store and browser choice screen design, and Meta’s ‘pay or consent’ model.

On April 23, 2025, Apple and Meta were both found non-compliant with the law and fined €500 million and €200 million respectively, with 60 days given to comply with the Commission’s decisions or risk periodic penalty payments.

These enforcement actions demonstrate the Commission’s willingness to impose substantial penalties. The speed of investigation—moving from preliminary findings to final decisions within 12-13 months—signals a more aggressive enforcement posture than traditional competition law proceedings, which often span years.

Ongoing Investigations and Preliminary Findings

On July 1, 2024, Meta received preliminary findings that its ‘pay or consent’ advertising model is in breach of the DMA, with the Commission viewing the model as forcing users to give consent on the use of their personal data without an alternative of a less personalized free version.

On March 19, 2025, Alphabet was informed of the Commission’s preliminary view that it is in breach of DMA rules regarding Google Play app store preventing developers from freely steering consumers to better offers, while Google search treats Alphabet’s services more favorably than others.

These investigations reveal the Commission’s interpretation of DMA obligations, particularly around genuine user choice and fair treatment of competitors. The outcomes will establish precedents for how similar provisions are enforced against other gatekeepers.

IV. Geopolitical and Global Regulatory Dimensions

The Brussels Effect: Global Regulatory Contagion

Governments around the globe are implementing their own versions of the DMA—Japan enacted its Smartphone Software Competition Promotion Act in June 2024, while discussions and proposals have emerged in Australia, Brazil, Canada, India, Indonesia, Kenya, Malaysia, Mexico, Morocco, New Zealand, South Africa, South Korea, Thailand, and Turkey.

This phenomenon, termed the “Brussels Effect,” reflects the EU’s ability to project regulatory influence globally. Kazakhstan, Uzbekistan, and Nigeria have adopted rules emulating aspects of the DMA, while content regulation regimes have been introduced or expanded in Australia, Canada, India, Brazil, South Korea, Singapore, and the United Arab Emirates.

The global proliferation of DMA-inspired legislation creates both opportunities and challenges:

Opportunities:

• Harmonized compliance frameworks reduce complexity for global platforms
• Shared regulatory standards facilitate international cooperation
• Competitive pressure may drive gatekeepers toward more open practices globally

Challenges:

• Regulatory fragmentation if countries adopt incompatible variations
• Compliance burden multiplies for companies operating in multiple jurisdictions
• Risk of regulatory arbitrage as companies optimize for most permissive regimes

Transatlantic Tensions and Trade Implications

The US government identified both the DMA and DSA as unfair trade barriers in the US Trade Representative’s 2025 National Trade Estimates Report, with the State Department in early August directing diplomats to condemn “undue” restrictions imposed by the DSA.

These tensions reflect deeper disagreements about digital regulation philosophy. The US traditionally favors ex-post antitrust enforcement targeting specific harms, while the EU’s ex-ante approach establishes preventive rules. Critics in the US technology sector argue the DMA disproportionately targets American companies and creates barriers to market entry.

According to reports, Brussels regulators were reassessing investigations against companies like Apple, Meta, and Google as US companies urged President Trump to intervene against what they characterized as overzealous EU enforcement, though Commission spokespeople denied any review was taking place.

V. Implications for Machine Learning Observability Systems

Understanding ML Observability in the DMA Context

Machine learning observability platforms represent a critical infrastructure layer for AI systems, providing monitoring, debugging, and performance optimization capabilities. Nearly three-quarters (72%) of organizations have adopted AIOps solutions to tackle the complexity of their multicloud environments, though 97% of technology leaders say probabilistic machine learning approaches limit the value that AIOps tools deliver given the manual effort required to gain reliable insights.

The adoption of AI technologies was the top strategy driving the need for observability (41%), with about two in five organizations (42%) deploying AI monitoring, 29% machine learning model monitoring, and 24% AIOps capabilities.

The intersection of the DMA and ML observability creates unique regulatory challenges and opportunities:

Data Access and Portability Requirements

The DMA’s data portability obligations fundamentally affect ML observability platforms that process data across gatekeeper services. Article 6(9) requires gatekeepers to provide effective portability of data generated through business user or end-user activity.

Implications for ML Observability:

1. Training Data Transparency: Observability platforms monitoring ML models on gatekeeper infrastructure must ensure compliance with data access requirements. If a business trains models using data from a gatekeeper platform, they must be able to export that training data and associated metadata.
2. Model Performance Metrics: Observability systems tracking model performance across gatekeeper platforms need standardized interfaces for data extraction. This includes inference latency, prediction accuracy, and resource utilization metrics.
3. Cross-Platform Monitoring: As businesses exercise DMA-granted rights to use alternative services, observability platforms must support monitoring across heterogeneous infrastructure, not just within gatekeeper ecosystems.

Interoperability Requirements for AI Infrastructure

The Commission has started to assess whether Microsoft and Amazon should be designated as gatekeepers for their cloud computing services, which would bring cloud AI infrastructure under DMA scrutiny. This development has profound implications for ML observability.

Strategic Considerations:

1. Open Telemetry Standards: ML observability vendors should prioritize OpenTelemetry and other open standards to facilitate interoperability across cloud providers. Built on top of OpenTelemetry, observability systems achieve agnosticism of vendor, framework, and language, granting flexibility in an evolving generative landscape.
2. Multi-Cloud Architecture: The average multicloud environment spans 12 different platforms and services, with organizations using 10 different observability or monitoring tools on average. DMA compliance may accelerate multi-cloud adoption as businesses reduce dependency on single gatekeepers.
3. API Standardization: Observability platforms should develop standardized APIs allowing seamless data exchange between competing cloud providers, enabling businesses to switch providers without losing historical monitoring data.

Competitive Dynamics in AI Monitoring Markets

The DMA’s anti-self-preferencing provisions create opportunities for independent ML observability vendors competing against gatekeeper-affiliated offerings.

Opportunity Vectors:

1. Level Playing Field: If cloud providers are designated as gatekeepers, they cannot preferentially promote their own observability tools over third-party alternatives. This creates market opportunities for specialized vendors.
2. Access to Performance Data: Gatekeeper obligations to provide access to data generated on their platforms may enable observability vendors to build more comprehensive monitoring solutions, accessing previously proprietary metrics.
3. Neutral Marketplaces: App store provisions requiring fair treatment could benefit observability tool providers seeking distribution through gatekeeper-controlled channels.

Risk Factors:

1. Compliance Complexity: Observability vendors must navigate gatekeeper compliance requirements, potentially requiring different implementations for EU versus non-EU deployments.
2. Feature Parity Challenges: While gatekeepers cannot self-preference, they may still enjoy natural advantages in instrumenting their own infrastructure more deeply than third parties can access.
3. Interoperability Overhead: Supporting data portability across competing platforms increases technical complexity and maintenance burden for observability vendors.

VI. Environmental and Sustainability Considerations

Energy Efficiency and Carbon Footprint

The DMA’s impact on environmental sustainability operates through multiple channels:

Infrastructure Proliferation: Organizations struggle to maintain visibility into cloud-native architectures as Kubernetes continues to become the platform of choice for modern applications, with 76% of technology leaders saying it’s more difficult to maintain visibility into this architecture compared with traditional technology stacks. Increased infrastructure complexity from multi-cloud strategies may increase aggregate energy consumption.

Optimization Opportunities: Conversely, improved observability driven by DMA compliance could enable better resource utilization. ML observability platforms detecting inefficient model deployments or over-provisioned infrastructure can reduce unnecessary computation and associated carbon emissions.

Data Center Competition: As the DMA promotes competition, alternative cloud providers may differentiate through superior sustainability practices, potentially accelerating green data center adoption.

Sustainable AI Development

ML observability platforms play a crucial role in sustainable AI development by:

1. Model Efficiency Monitoring: Tracking computational efficiency metrics (FLOPs per inference, memory utilization) enables identification of wasteful models.
2. Carbon Attribution: Advanced observability systems can attribute carbon emissions to specific models or training runs, enabling accountability and optimization.
3. Resource Right-Sizing: Continuous monitoring allows dynamic resource allocation, ensuring models use only necessary computational resources.

The DMA’s emphasis on data access could facilitate development of sustainability benchmarking tools, allowing comparison of model efficiency across platforms and driving competitive pressure toward greener AI practices.

VII. Sociopolitical Dimensions: Power, Governance, and Digital Rights

Platform Power and Democratic Discourse

The DMA represents a fundamental assertion of public authority over private platform governance. European lawmakers hope the DMA will increase fairness and boost competition, addressing concerns that concentrated platform power undermines democratic values by controlling information flows and economic opportunities.

Democratic Accountability: By imposing transparency requirements and limiting data combination practices, the DMA seeks to make platform power more accountable to public oversight. This is particularly relevant for ML systems making consequential decisions about content visibility, economic opportunities, and service access.

Algorithmic Transparency: ML observability systems could play a role in DMA compliance by documenting algorithmic decision-making processes. If gatekeepers must explain ranking or recommendation systems, observability platforms providing audit trails become critical infrastructure for regulatory compliance.

Privacy and User Autonomy

Article 5(2) of the DMA requires gatekeepers to obtain consent from users when they intend to combine or cross-use their personal data, and if users refuse consent, they should have access to a less personalized but equivalent alternative.

This provision has profound implications for ML systems relying on cross-service data:

Training Data Constraints: ML models trained on combined datasets from multiple gatekeeper services must ensure proper consent frameworks. Observability platforms should monitor consent status for training data to ensure compliance.

Model Personalization: The prohibition on forced data combination affects personalized ML models. Observability systems must track whether models inappropriately leverage cross-service data without consent.

Performance Trade-offs: As gatekeepers offer less personalized alternatives, ML observability becomes critical for measuring performance degradation and ensuring “equivalent” functionality as required by the DMA.

Labor Market and Skills Implications

The DMA’s enforcement creates demand for new professional competencies:

Compliance Specialists: Understanding DMA requirements and implementing technical compliance measures requires specialized legal-technical expertise, creating employment opportunities.

Interoperability Engineers: Building systems that work across heterogeneous gatekeeper platforms demands sophisticated engineering capabilities.

ML Ethics and Governance: As observability platforms document algorithmic decision-making for regulatory purposes, roles bridging technical ML work and regulatory compliance will expand.

Organizations deploying AI monitoring, ML model monitoring, and AIOps capabilities estimated higher annual total value received from observability than those who hadn’t deployed them, suggesting these investments generate tangible business returns alongside regulatory compliance.

VIII. Strategic Adaptations for Machine Learning Observability Platforms

Based on the comprehensive analysis above, ML observability platforms should consider the following strategic adaptations:

1. Embrace Open Standards and Interoperability

Action Plan:

• Prioritize OpenTelemetry implementation for data collection and export
• Develop standardized APIs for cross-platform monitoring
• Participate in industry standards bodies shaping observability protocols
• Design architecture assuming multi-cloud, multi-platform deployments

Rationale: The DMA’s interoperability requirements will accelerate demand for vendor-neutral observability solutions. Platforms locked into proprietary formats face displacement by interoperable alternatives.

2. Build DMA-Compliant Data Governance Features

Action Plan:

• Implement granular data access controls aligned with DMA portability requirements
• Develop automated consent verification for training data sources
• Create audit trails documenting data provenance and usage
• Build data export functionality enabling seamless platform switching

Rationale: As gatekeepers face DMA data access obligations, observability platforms must support compliant data handling. Failure to do so creates liability and limits market opportunities.

3. Develop Cross-Platform Monitoring Capabilities

Action Plan:

• Invest in connectors for multiple cloud providers and ML platforms
• Build unified dashboards aggregating metrics across heterogeneous infrastructure
• Develop cross-platform cost optimization features
• Create compatibility layers abstracting platform-specific differences

Rationale: The average multicloud environment spans 12 different platforms and services. DMA compliance will likely increase multi-cloud adoption, creating demand for unified observability across platforms.

4. Position as Neutral, Independent Alternative

Action Plan:

• Emphasize independence from gatekeeper platforms in marketing messaging
• Develop partnerships with smaller cloud providers seeking DMA compliance
• Create “DMA compliance certification” demonstrating adherence to fair competition principles
• Avoid exclusive distribution agreements that could be construed as self-preferencing

Rationale: The DMA creates opportunities for independent vendors by limiting gatekeeper self-preferencing. Market positioning emphasizing neutrality and fairness aligns with DMA values and differentiates from gatekeeper-affiliated offerings.

5. Invest in Algorithmic Transparency and Explainability

Action Plan:

• Build features documenting ML model decision-making processes
• Develop bias detection and fairness monitoring capabilities
• Create compliance reporting modules for regulatory submissions
• Implement model lineage tracking from training through deployment

Rationale: As regulators scrutinize algorithmic decision-making, observability platforms providing transparency become essential infrastructure. This positions platforms as enablers of responsible AI development.

6. Develop Sustainability Monitoring Features

Action Plan:

• Implement carbon footprint tracking for ML workloads
• Build energy efficiency benchmarking across models and platforms
• Create recommendations for resource optimization
• Develop sustainability reporting aligned with emerging ESG disclosure requirements

Rationale: Environmental sustainability increasingly influences technology purchasing decisions. Observability platforms enabling green AI practices differentiate in the market while supporting corporate sustainability goals.

7. Prepare for Global Regulatory Expansion

Action Plan:

• Monitor DMA-inspired legislation in other jurisdictions
• Design flexible compliance frameworks adaptable to varying requirements
• Establish regional partnerships for localized compliance expertise
• Build configurable policy engines supporting jurisdiction-specific rules

Rationale: Governments around the globe are implementing their own versions of the DMA. Platforms anticipating global regulatory expansion will capture first-mover advantages in emerging markets.

8. Foster Community and Standards Participation

Action Plan:

• Contribute to open-source observability projects
• Participate in industry working groups defining ML monitoring standards
• Collaborate with academic researchers studying AI governance
• Engage constructively with regulators to shape practical compliance requirements

Rationale: Active participation in standards development and regulatory dialogue positions platforms as thought leaders and ensures emerging requirements align with technical realities.

IX. Risk Mitigation Strategies

ML observability platforms face several risk categories requiring mitigation:

Regulatory Compliance Risks

Risk: Non-compliance with DMA data portability or interoperability requirements could result in customer liability.

Mitigation:

• Conduct regular compliance audits of data handling practices
• Engage legal counsel specializing in EU digital regulation
• Implement automated compliance checking in product workflows
• Maintain detailed documentation of compliance measures

Technology Fragmentation Risks

Risk: Supporting multiple incompatible regulatory regimes across jurisdictions increases complexity and costs.

Mitigation:

• Design modular architecture with pluggable compliance components
• Implement feature flags enabling region-specific functionality
• Develop robust testing frameworks validating compliance configurations
• Establish clear product roadmap balancing global consistency with regional requirements

Competitive Displacement Risks

Risk: Gatekeeper-affiliated observability offerings may retain advantages despite DMA anti-self-preferencing rules through deeper platform integration.

Mitigation:

• Develop unique value propositions beyond basic monitoring (e.g., advanced AI-powered analytics, superior user experience)
• Build strong customer relationships through exceptional service and support
• Create ecosystem partnerships with complementary tool providers
• Innovate rapidly to maintain technical leadership

Market Access Risks

Risk: Gatekeepers may use technical complexity or non-compliance with platform-specific requirements to limit third-party observability tool effectiveness.

Mitigation:

• Document any obstacles to fair market access and report to regulators
• Develop public advocacy emphasizing importance of observability competition
• Build coalition with other affected vendors to amplify voice
• Pursue legal remedies if anti-competitive practices are identified

X. Future Outlook: Emerging Trends and Considerations

AI System Designation Under DMA

MEPs expressed concern that no gatekeeper had been designated for cloud services, noting that cloud services are indispensable for the deployment and growth of AI technologies, and requested the Commission closely monitor both cloud services and AI services as they are tightly interdependent.

If AI systems or cloud AI platforms are designated as core platform services, ML observability becomes central to DMA compliance. Platforms enabling documentation and monitoring of AI system behavior could become regulatory necessities rather than optional tools.

Evolution of Enforcement Standards

The DMA has significantly reshaped the digital landscape in Europe and beyond, strengthening competition and bolstering the EU’s geopolitical influence, but has also introduced challenges including trade-offs and delays in launching new products and services, particularly in AI.

As enforcement proceeds, the Commission will refine its interpretation of DMA provisions. Observability platforms should monitor enforcement decisions to anticipate evolving compliance requirements and adjust product offerings accordingly.

Potential Legislative Amendments

The DMA includes provisions for periodic review and updating. Article 35 DMA requires the Commission to submit an annual report on the implementation of the DMA and the progress made towards achieving its objectives. Future amendments may explicitly address AI systems, algorithm transparency, or environmental sustainability, creating new opportunities and requirements for observability platforms.

Integration with Other Regulatory Frameworks

The DMA operates alongside the Digital Services Act (DSA), General Data Protection Regulation (GDPR), and emerging AI Act. ML observability platforms providing unified compliance across these frameworks will offer significant value. Developing “regulatory observability” features aggregating compliance evidence across multiple legal requirements represents a compelling product differentiation opportunity.

XI. Conclusion: Navigating the New Digital Regulatory Landscape

The Digital Markets Act represents a watershed moment in digital regulation, fundamentally reordering power relationships between platforms, businesses, and users. For machine learning observability platforms, the DMA creates both challenges and opportunities.

Key Takeaways:

1. Regulatory Inevitability: DMA-style regulation is proliferating globally. Platforms treating this as a temporary or regional phenomenon will be unprepared for the emerging regulatory reality.
2. Interoperability Imperative: Open standards and cross-platform compatibility transition from competitive differentiators to baseline requirements. Proprietary lock-in becomes increasingly untenable.
3. Transparency as Value: The DMA’s emphasis on transparency and fair treatment aligns with growing demand for responsible AI. Observability platforms enabling algorithmic accountability will thrive.
4. Strategic Agility Required: The regulatory landscape remains fluid. Platforms maintaining architectural flexibility and rapid adaptation capabilities will navigate uncertainty most effectively.
5. Collaboration Over Confrontation: Engaging constructively with regulators, contributing to standards development, and building coalitions advances industry interests more effectively than adversarial postures.

The DMA’s long-term impacts remain uncertain. The DMA ought to be rewritten to stipulate clear and cogent legal standards, return to a strengthened system of ex post control, and provide platforms with the opportunity to deliver efficiency defenses as part of corporate self-regulation, some analysts argue. Others contend the DMA represents necessary correction to market failures requiring ongoing vigilance and enforcement.

For machine learning observability platforms, success requires embracing the DMA not merely as a compliance burden but as a catalyst for innovation. Platforms that internalize DMA values—openness, fairness, user empowerment, transparency—into their product DNA will be best positioned to thrive in the evolving digital ecosystem.

The future of AI development increasingly depends on robust observability infrastructure. As regulatory scrutiny intensifies, ML observability platforms that facilitate responsible, compliant, and sustainable AI practices will become indispensable partners in the digital economy’s next chapter. The organizations that recognize this opportunity and invest accordingly will shape the future of both AI technology and digital governance.

About This Analysis: This comprehensive examination synthesizes information from 50+ primary and secondary sources including European Commission documents, legal analyses, industry reports, and academic research. It reflects the state of DMA implementation and enforcement as of December 2024, recognizing that this regulatory landscape continues to evolve rapidly.

About the Author

admin

Administrator

Visit Website View All Posts