Intellectual property has remade capitalism by turning knowledge into the prime means of production, concentrating power in patent-rich corporations; the rise of AI and cloud platforms is amplifying that concentration and deepening global inequalities.

Main Finding

Legal enclosure of knowledge — primarily through strengthened intellectual property rights (IPRs) and patents, institutionalized globally via TRIPS — has produced a distinct phase of "knowledge capitalism" in which intangible assets (knowledge, data, algorithms, AI models) are the primary means of production. This regime creates and reinforces monopolistic corporate power, skews the distribution of profits toward IP-rich firms (especially large tech firms), entrenches global inequalities between IP-rich and IP-poor countries, and is amplified by the rise of AI and proprietary cloud infrastructures.

Key Points

• Rise of intangibles: Corporate investment and firm valuation have shifted toward intangible assets (knowledge, software, data). Intangible investment intensity has grown markedly since the 1990s.
• Economic properties of intangibles: High fixed (up‑front) costs, near-zero marginal costs, scalability and network effects. These properties change competitive dynamics and market structure.
• Patents and IPRs as instruments of control: Patents are not only incentives for innovation but function as economic instruments — enabling exclusion, monopolistic rents, and strategic behavior (patent pools, acquisitions).
• Concentration of patenting and profits: Patenting activity and patent ownership have shifted from small inventors to large corporations; firms with robust IPRs (notably U.S. tech firms) capture disproportionate global profits.
• TRIPS and global institutionalization: The TRIPS regime globalized stringent IPR protections, making knowledge exclusion enforceable across borders and amplifying divergence between advanced and developing economies.
• Legal and regulatory lag: Rapid frontier innovation (especially digital/AI) outpaces regulation, creating "policy vacuums" that allow firms extended uncontested profit windows and to shape regulation via lobbying.
• Distributional consequences: Knowledge enclosure contributes to rising inequality (capital incomes vs labor), limits technology diffusion to poorer countries, and can produce "snowball dynamics" where IP-rich countries accumulate more advantage.
• AI and proprietary infrastructures: The growth of AI models and cloud/data-center infrastructure intensifies monopolistic tendencies — control over data, compute, and models becomes a new form of proprietary capital.

Data & Methods

• Approach: Conceptual and legal-historical analysis supplemented by empirical evidence from secondary sources. The paper synthesizes literature across law, economics and sociology and uses illustrative empirical indicators rather than primary new data collection.
• Empirical sources and indicators cited:
  • Trends in intangible investment and firm value: Financial Times reporting (Tei Parikh) and associated charts showing rising intangible investment (1995–2024) and growth of intangible-led sectors in corporate value (U.S., 1940–2020).
  • Patent statistics: Studies by Akcigit, Grigsby & Nicholas (patents filed with the USPTO), OECD/IP5 patent family data and World Bank figures.
  • Knowledge Economy Index: Diessner et al. (2025) combining technology (ICT, robots, patents/IP5 per employees) and skill indicators across OECD countries (1995–2019).
  • Corporate profit concentration: Analysis of Forbes Global 2000 firms and sectoral profit shares (Schwartz, 2019) and high-profile patent acquisitions (e.g., Google–Motorola, Microsoft–Nokia).
  • Legal-institutional evidence: TRIPS Agreement and scholarly critiques (Belloc & Pagano; Rikap).
• Methods used: literature review, legal analysis of IPR regimes, case examples and interpretation of patent/trade data to support theoretical claims.
• Limitations: The paper is primarily theoretical and interpretive; empirical claims rely on secondary aggregated data and illustrative examples rather than formal causal identification.

Implications for AI Economics

• Market structure and market power
  • AI economics models must incorporate stronger monopoly/oligopoly dynamics driven by IPRs, scale economies, and zero marginal costs of models and software.
  • Control over data, compute (data centers), and pretrained models becomes a key source of market power — affecting entry, pricing, and innovation rates.
• Innovation incentives vs exclusion
  • While IPRs can incentivize R&D, in AI they often enable exclusion (blocking competitors via patents, model copyrights, or data monopolies). Economic analysis should weigh dynamic innovation benefits against static and dynamic costs of restricted diffusion.
• Distributional effects and labor
  • Enclosure of AI knowledge concentrates returns to capital owners and can suppress wage growth in sectors where capital-intensive, proprietary AI substitutes labor — important for distributional and policy modeling.
• International divergence
  • TRIPS-style global IP enforcement interacts with AI to widen cross-country gaps: countries lacking domestic IP/AI assets face higher barriers to adoption and value capture. AI economics must model cross-border technology diffusion constrained by IP regimes.
• Policy and regulation
  • Economic research should evaluate policy tools: antitrust enforcement tailored to AI/platforms, IP reform (narrower patentability, compulsory licensing), data governance (data trusts, interoperability mandates), public investments in open models/infrastructure, and standards to reduce lock-in.
• Empirical research directions
  • Measure concentration of AI-related IPRs (patents, model copyrights) and their correlation with market shares, profitability, and R&D spillovers.
  • Causal studies exploiting policy changes (e.g., national IP reforms or TRIPS-like adoption) using difference-in-differences, synthetic controls or event studies to identify effects on innovation diffusion, wages, and firm dynamics.
  • Network / citation analyses of patent families and model reuse to quantify path-dependence and cumulative innovation dynamics.
  • Firm-level structural models incorporating zero marginal costs, network effects, and IPR rents to forecast market evolution under alternative policy regimes.
• Practical modeling suggestions
  • Replace perfect-competition assumptions for AI goods with models of monopolistic competition or oligopoly with asymmetric ownership of key intangibles.
  • Explicitly model fixed versus marginal cost structure for AI (training vs inference), and how IPRs and exclusive access to data/compute alter investment and pricing decisions.
  • Include cross-border IP enforcement as a constraint on technology transfer and as an input into growth/equity models for developing countries.

Overall, the paper suggests that the legal enclosure of knowledge transforms the economic logic around AI — turning data, models and compute into proprietary capital — and that AI economics must incorporate legal-institutional realities (IP regimes, cloud infrastructure control) to accurately analyze innovation, market structure, growth and inequality.