Platforms convert unpaid attention into monetisable data, concentrating market power and rents while embedding 'immaterial' digital production in costly physical infrastructure; this reorganisation of accumulation intensifies extraction, precarious platform labour, and distributional tensions.

Main Finding

Digital capitalism is best understood not as a break from capitalism nor simply a continuation of older markets, but as a reorganisation of capitalist accumulation around platforms, data, attention and digital labour. Platforms convert unpaid and paid user activity into measurable attention and data, which are then monetised across a set of business models. This reproduces capitalism’s core logic (value extraction and accumulation) while producing distinctive contradictions: commodified unpaid activity, the material basis of supposedly immaterial production, concentrated market power, and a separation between user participation and algorithmic control.

Key Points

• Third-position argument: digital capitalism reorganises capitalist accumulation rather than replacing or merely extending prior forms.
• Shared logic across media business models: advertising, subscriptions, creator monetisation, in‑game spending and platform commissions all capture user activity as attention → measure it as data → convert it into revenue.
• Digital labour: both unpaid (user-generated activity and attention) and paid (platform labour, gig work, microtasks) are central to value creation.
• Infrastructure duality: “weightless” digital products rest on dense physical infrastructure (data centres, networks, cloud hardware), making production materially grounded.
• Market concentration: advertising markets, cloud services, and major platforms show high concentration, amplifying rent extraction and control.
• Four core contradictions:
• Commodification of unpaid user activity — revenue extraction from nonwaged activity.
• Material basis of immaterial production — digital outputs require substantial physical inputs and costs.
• Concentration of market power — winner-take-most dynamics in platforms, ads, cloud computing.
• Gap between participation and algorithmic control — users participate but lack control over how their activity is processed and monetised.
• Conceptual contribution: reframes new media as economic infrastructures built for value extraction rather than merely channels for content or interaction.

Data & Methods

• Method: critical narrative review anchored in Marxian value theory and literature on platforms, datafication and surveillance capitalism.
• Evidence: curated set of publicly reported indicators (institutional and market sources) used as contextual description rather than causal testing. Examples of indicators:
  • Platform reach (user penetration across world population).
  • Market concentration metrics for advertising and cloud markets.
  • Scale and geographic spread of platform labour and gig work.
  • Physical infrastructure measures (data centre capacity, energy use, network backhaul).
• Approach: conceptual synthesis + contextual empirical indicators (no formal econometric causal identification).
• Limitations: indicators are descriptive contextual evidence; the study’s aim is theoretical and conceptual clarification rather than statistical causation.

Implications for AI Economics

• Data as foundational input for AI: platforms’ capture of attention and user data becomes the raw material for training AI systems, concentrating an important input and increasing barriers to entry.
• Amplified concentration and rents: control over large-scale datasets, model infrastructure (cloud, specialized chips), and distribution channels gives incumbents stronger market power and economic rents.
• Value capture and distribution: AI-enabled monetisation intensifies questions about who captures surplus from user-generated data and AI outputs (platforms, advertisers, creators, users).
• Labour dynamics:
  • Increased unpaid extraction (user-generated data) and precarious paid digital labour (gig and microtasking) supply training data and model evaluation.
  • Potential displacement of some paid tasks by AI while creating new forms of platform-mediated work and downstream creator monetisation pressures.
• Importance of physical inputs: model training and inference are resource-intensive (compute, energy, datacenters), so policy and economics must account for capital, operating costs and environmental externalities—not just “weightless” digital value.
• Measurement needs for AI economics:
  • Better metrics for attention, data flows, and value per unit of user activity.
  • Revenue attribution across multi-sided platform chains (advertisers → platforms → creators → users).
  • Accounting for nonmarket transfers (unpaid activity as implicit subsidies).
  • Monitoring concentration across datasets, models, compute, and distribution.
• Policy levers and regulatory implications:
  • Antitrust and market-structure interventions to curb winner-takes-most dynamics.
  • Data governance: portability, access rules, and limits on unilateral data monopolisation to lower entry barriers for AI development.
  • Labor protections and new forms of compensation for data- and attention-providing activities (e.g., creator rights, remuneration frameworks).
  • Taxation of platform rents and public investment in alternative infrastructures (public data commons, open compute resources).
• Research agenda suggestions:
  • Model platforms as multi-sided markets where attention/data are scarce, monetisable inputs with nonrival characteristics but excludable control.
  • Empirically estimate the value capture per user-hour of attention and how AI changes that mapping.
  • Measure the distributional impacts of AI-enabled platforms on wages, rents and regional inequality.
  • Analyze the role of physical infrastructure constraints (compute availability, energy) in shaping AI market structure and incentives.
  • Study governance designs that re-balance bargaining power (data trusts, collective bargaining for platform workers, public provision).

Concluding note: treating new media and AI ecosystems as economic infrastructures for value extraction highlights the continuity with capitalist accumulation while clarifying the novel mechanisms—datafication, attention capture, platform-mediated labour and concentrated compute—that reshape distributional and policy challenges in the AI era.