A few cloud and platform providers now exert de facto governance over digital services by embedding asymmetric interdependencies in data-driven supply chains; effective policy should target these chain-level dynamics rather than only individual apps or firms.

Main Finding

Data-driven supply chains — the modular, API-mediated networks of cloud services, infrastructure providers, and component platforms that underpin most digital applications — function as distributed sites of governance in informational capitalism. Through data flows and modular integrations, these chains simultaneously produce application functionality and generate commodified data. Asymmetric interdependencies among actors (especially hyperscale cloud and platform providers) create concentrated technical, economic, and regulatory power that shapes outcomes, accountability, and value extraction. Law and regulation are not external to these dynamics but actively shape interdependencies and therefore provide levers to rebalance power.

Key Points

• Platformisation and modularisation: Modern apps are built by stringing together cloud and infrastructure services (compute, databases, AI/ML APIs, identity, payments, CDN, analytics, adtech). A single visible application is typically many systems controlled by multiple actors.
• Dual role of supply chains: They (a) enable application functionality through API-mediated processing, and (b) produce, capture and route large volumes of data used for profiling, targeting and AI development.
• Capture and datafication: Human activities are reconceptualised as data (capture), which is commodified and monetised by platforms. This is the principal logic of value extraction under informational capitalism.
• Concentration of infrastructure: A few hyperscalers (AWS, Azure, Google Cloud, plus major platform firms) control a large share of cloud services, creating structural dependencies across many downstream applications and sectors.
• Interdependence as governance: Control is distributed across many actors; interdependence is dynamic and instance-specific (data processing changes subsequent flows), producing multipolar governance where power flows through networked relations rather than single owners.
• Asymmetric interdependencies: Differences in providers’ control over compute, data, identity, or routing translate into bargaining power and the ability to shape technical and legal terms, entrench market position, and extract value.
• Invisibility and accountability gaps: Users and many downstream actors often lack visibility into the supply-chain composition and data flows, complicating oversight, liability, and regulatory enforcement.
• Law matters: Contractual terms, liability regimes, data protection rules, procurement law, and competition policy influence how interdependencies form and which actors accrue power; thus law/regulation can be used to reshape supply-chain governance.

Data & Methods

• Methodological approach: Conceptual and normative analysis grounded in political economy, legal scholarship, and prior empirical literature on platforms, cloud computing, adtech, and supply-chain security.
• Evidence and illustrations: Uses a canonical case study (a food-ordering mobile app) to exemplify how client apps integrate cloud, payment, identity, CDN, analytics and adtech providers; discusses adtech and tracking chains as empirical examples of extensive data-extractive supply chains.
• Literature synthesis: Draws on scholarship on platformisation, capture/datafication (Agre; Cohen), supply-chain security guidance (NCSC, CISA), adtech investigations, and contemporary analyses of market concentration in cloud services.
• Analytical lens: Interdependence — examining how legal, technical and economic allocations of control generate asymmetric dependencies and distributed governance effects.
• Not an empirical econometric study: the chapter is a conceptual/legal analysis and policy sketch rather than quantitative causal estimation.

Implications for AI Economics

Practical and research implications for the economics of AI, markets, and policy:

1. Market structure, concentration and barriers to entry

   • Concentration of cloud and platform services creates high switching costs and network effects that can entrench dominant suppliers of compute, data, and ML tooling.
   • Economists should quantify these concentration effects (market shares, HHI), vertical integration, and impacts on competition in AI model provisioning and downstream applications.

2. Data access, rents, and distribution of value

   • Data generated in supply chains is a source of competitive advantage. Unequal access to training/behavioral data produces rents for upstream providers and affects innovation incentives.
   • Research directions: measure the relationship between data control and AI model performance, pricing, and firm profitability; estimate distributional consequences across firms and workers.

3. Externalities, systemic risk and resilience

   • Interconnected supply chains produce systemic dependencies (single-point failures, contagion, security vulnerabilities). These are externalities not internalised by market actors.
   • Policy and modelling: assess systemic risk metrics for cloud-dependent industries; value of redundant public or open infrastructure; cost–benefit of mandated diversification.

4. Governance, regulation and antitrust

   • Legal and regulatory tools can reshape interdependencies: mandates on interoperability, data portability, provenance/auditability, liability allocation, incident reporting, and procurement rules can reduce lock-in and increase accountability.
   • Antitrust remedies might consider data and infrastructure access (not just consumer prices), vertical conduct, and bundling of AI services with upstream infrastructure.
   • Economists should model effects of these interventions on investment, innovation, and welfare.

5. Measurement and empirical strategies

   • Important empirical tasks: mapping supply chains (service dependencies, API calls), estimating data flows, measuring visibility/transparency gaps, and tracing provenance of datasets used for AI.
   • Methods: network analysis of service dependencies, event/incident logs, platform usage telemetry, procurement and billing data, and case studies of outages or data breaches.

6. Policy prescriptions with economic rationale

   • Promote transparency and auditability: require provenance logs and disclosure of critical dependencies to reduce information asymmetries and facilitate contestability.
   • Encourage interoperability and data portability to lower switching costs and spur competition; evaluate via counterfactual competition models.
   • Support public or open compute/data infrastructure to lower entry barriers and internalise social value of data-intensive public goods (health, safety-relevant models).
   • Strengthen supply-chain security standards and mandatory incident reporting to internalise negative externalities and protect welfare.
   • Revisit taxation and revenue rules to address value capture from data and platform coordination.

7. Research agenda examples

   • Estimate how hyperscaler market power translates into price markups or reduced product variety for downstream AI services.
   • Quantify welfare effects of data-sharing mandates vs. investment incentives for data collection.
   • Model bargaining over API/compute access in a vertical chain and explore efficient regulatory interventions.
   • Empirically map the overlap between adtech/tracking data and datasets used for commercial AI models to assess concentration of training inputs.

Concluding note: Viewing cloud- and API-mediated systems as data-driven supply chains reframes many AI-economics questions: market power is exercised not only through product markets but via control of the infrastructure and data flows that enable AI. Effective economic analysis and policy must therefore measure and address interdependencies, visibility gaps, and the institutional rules that shape them.