Innovation travels business class, but not for everyone

A patented invention in India will be reused three times faster by an American researcher than by an Indian researcher. This single figure, drawn from the 2026 WIPO report on global intellectual property, summarizes a reality that enthusiasm for the “democratization” of knowledge tends to mask: technological diffusion is anything but equitable.

The 2026 WIPO report brings new measures of asymmetric diffusion through breakthrough inventions, drawing on fifty years of patent citation data. What it reveals is not a lucky accident. It is an architecture.

The essentials

  • The United States absorbs Indian inventions three times faster than India itself, according to the 2026 WIPO report
  • Fundamental technologies — biotech, AI, quantum — take an average of ten years to transform into commercial patents
  • The speed of diffusion depends structurally on human capital, intellectual property regimes, and trade agreements in place
  • This gap creates a cumulative advantage for already-dominant economies, which capture science produced elsewhere before the countries of origin can benefit from it

An Indian invention belongs to whoever can exploit it

The WIPO figure deserves careful reading. It’s not that the United States steals Indian patents. It’s that they possess the instruments — universities, laboratories, venture capital, engineers trained to translate an idea into a product — that allow them to absorb a foreign discovery before the country of origin even understands its commercial scope.

Technological diffusion functions like a reading race: everyone has access to the same text, but those who read fast and comprehend quickly benefit first. And in the knowledge economy, arriving first is not a detail — it is often the entirety of the competitive advantage.

WIPO measures this speed by tracing patent citations across borders. When an American patent cites an Indian patent, it is documented proof that knowledge has traveled. The frequency and speed of these citations reveal who captures what, and how much time passes after original publication. The finding is clear: rich countries cite inventions from emerging countries much earlier than the latter cite themselves or cite each other.

This mechanism has been documented in the economics literature for several decades, but the 2026 report brings it new and systematic measures, notably through trajectories of breakthrough inventions over fifty years. And precise measures are unsettling.

Ten years between discovery and patent: a window only some can open

The report’s second finding is perhaps even more structural. In sectors of fundamental research — biotechnology, artificial intelligence, quantum computing — an invention takes an average of ten years to transform into a commercializable patent. Ten years during which knowledge circulates freely in scientific publications, conferences, doctoral theses.

This period is often described as a common good of knowledge. It is, in theory. In practice, it is a race. And like any race, it favors those who have the right shoes.

Transforming a scientific publication into an exploitable patent requires lawyers specialized in intellectual property, laboratories capable of moving from prototype to product, financiers who understand early-stage deeptech risk. These resources are not distributed randomly across the planet. They concentrate in Boston, San Francisco, Munich, Singapore.

A researcher publishing a biotech breakthrough from Nairobi or Dhaka contributes to the common good of world science. He or she is also, objectively, less well positioned to capture the economic value of their own discovery in the decade that follows. This is not a fatality. It is a finding that calls for specific policies.

This phenomenon is not unrelated to broader debates about who actually benefits from the productivity growth induced by AI. The race for AI concentrates its benefits where infrastructure already exists, reinforcing the advantages of economies that already dominate the technological value chain.

The rules of the game are written by those who win

Technological diffusion does not happen in a legal vacuum. It operates within a framework: intellectual property agreements, trade treaties, technology transfer clauses. And this framework, essentially, was negotiated by and for the economies that export technology.

The WTO TRIPS agreement, which harmonizes patent regimes globally since 1995, imposes on developing countries intellectual protection standards modeled on those of the United States and Europe. The economic logic is understandable: protect innovation to encourage it. But the concrete effect is to restrict the capacity of less advanced countries to use, adapt, and repackage imported technologies — what economists call learning by imitation, which was nonetheless the engine of industrial catch-up for South Korea, Taiwan, even Japan.

These countries took off during a period when intellectual property rules were much less strict. They were able to copy, adapt, improve, then innovate. This path is much narrower today for a Bangladesh or Tanzania. The WIPO report does not state it in these terms — it is, after all, the UN body tasked with defending intellectual property — but the data it produces invites this reading.

Human capital: the least visible bottleneck

Legal regimes matter. But the WIPO report insists on a factor perhaps even more determinant: human capital. The capacity to absorb foreign technology depends first on the density of engineers, researchers, technicians, and managers capable of understanding what they read in a patent or publication.

This absorption capacity — a concept formalized by economists Wesley Cohen and Daniel Levinthal as early as 1990 — is the silent bottleneck of technological diffusion. A country can sign all the technology transfer agreements it wants; if it does not have the human resources to understand and deploy what is transferred to it, the agreement remains a dead letter.

Examples abound. Dozens of African countries have signed agreements with Chinese solar manufacturers to install renewable capacity. Many have obtained the panels. Fewer have obtained training to maintain them, repair them, integrate them into smart grids. The technology arrived; the capacity to sustainably benefit from it, less so.

WIPO measures this gap by correlating the speed of technological diffusion with indicators of tertiary education, R&D spending, and researcher density per capita. The correlation is strong. This is not a surprise — but documenting it rigorously makes it possible to respond to discourses that attribute technological inequalities to cultural factors or a lack of “entrepreneurial mentality.” The reality is more prosaic and more actionable: it is a matter of public investment in education and research.

This is precisely what Tyler Cowen observed in his analysis of American productivity: innovation gains do not diffuse spontaneously, even within a developed economy. Absorption remains a deliberate act, one that demands specific investments.

What some countries are attempting, and why it deserves attention

The picture would be incomplete if it stopped at diagnosis. Several strategies show that the diffusion gap is not impermeable.

India, precisely, offers a partially interesting counter-example. Its generic pharmaceutical sector was built largely on a deliberately strict interpretation of patentability in its 2005 Patent Law. Section 3(d) of that law requires that a molecule be substantially more effective than a prior form to be patentable — which allowed manufacturers like Cipla to produce antiretrovirals at prices 90% lower than patented versions, directly contributing to expanding access to HIV treatments in low-income countries. India used the margins of maneuver in international law to build large-scale capacity for absorption and adaptation.

Brazil followed similar logic in the 2000s for antiretrovirals, issuing compulsory licenses. These strategies faced intense pressure from the United States and European Union through commercial sanction mechanisms — which says something about the balance of power.

In the AI field, several countries are seeking to accelerate their absorption capacity by other means. Kenya, Ethiopia, and Rwanda have launched national AI programs that bet less on fundamental research — too costly, timelines too long — than on adapting existing models to local contexts: agriculture, community health, public services. The approach is pragmatic: rather than seeking to produce innovation first, maximize adoption speed of available tools and build local competencies in the process.

This is a strategy consistent with what WIPO data reveals: the speed of diffusion depends less on who invents than on who is ready to absorb. Building this absorption capacity is the most rational investment for an emerging country wishing to capture a share of global technological value.

India, moreover, is not resigned to its role as an involuntary donor. The Modi government has massively invested in STEM programs since 2014. According to WIPO (WIPI 2024), Indian patent and industrial design applications more than doubled between 2018 and 2023; for patents alone, the increase is approximately 40% over this period, from roughly 46,000 to 64,480 applications. The question is not whether India can reverse the asymmetry measured in the report — but at what horizon, and at what cost.

The next frontier: who writes the rules for generative AI and quantum computing

The 2026 WIPO report arrives at a pivotal moment. The technologies that today take ten years to transform into patents — generative AI, quantum computing, synthetic biotechnology — are precisely those that will reshape the next decade of the global economy.

The intellectual property regimes applied to these technologies remain largely undefined. Who owns an AI model trained on public data? How do you patent a quantum algorithm? What rights does a country retain whose cultural, linguistic, or medical data was used to train a commercial model?

These questions are under negotiation, principally in forums where developing countries are poorly represented. The European Union and the United States are advancing their own frameworks — the European AI Act, American Executive Orders on AI — without coordination with the Global South. WIPO seeks to play the role of neutral forum, but its mandate remains limited to patents and copyright, not data or model governance.

What the report’s data suggests is simple: if the rules of the game for tomorrow’s fundamental technologies are written before emerging countries have strengthened their absorption capacity, the asymmetry measured today will reproduce itself, amplified. If, conversely, negotiation spaces allow mechanisms of technology transfer, training, and flexibility in patent regimes, the cycle can be partially broken.

Latin America offers an example of what deliberate upskilling strategy can produce when public policies are seriously deployed to this end. It is neither automatic nor rapid. But it is documented.

The real question posed by the 2026 WIPO report is not technical. It is political: who sits at the table when the rules of AI and quantum diffusion are written? And who funds the training of engineers that will allow emerging countries to absorb these technologies before they are entirely captured by those who produced them?

Sources

  1. WIPO — World Intellectual Property Report 2026 (full report): https://www.wipo.int/web-publications/world-intellectual-property-report-2026/en/index.html
  2. WIPO – WIPR 2026 Executive Summary: https://www.wipo.int/web-publications/world-intellectual-property-report-2026/en/executive-summary.html
  3. WIPO – Video WIPR 2026 Main Findings: https://www.wipo.int/en/web/videos/w/wipr-2026-main-findings
  4. WIPO – WIPR 2026 Chapter 2 (knowledge diffusion): https://www.wipo.int/web-publications/world-intellectual-property-report-2026/en/2-global-trends-of-technological-knowledge-diffusion.html
  5. Cohen, W. M. & Levinthal, D. A. (1990). Absorptive Capacity: A New Perspective on Learning and Innovation. Administrative Science Quarterly, 35(1), 128-152: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1504447
  6. Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), WTO, 1995: https://www.wto.org/english/tratop_e/trips_e/intel2_e.htm
  7. Indian Patents Act, Section 3(d), Amendment 2005: https://www.drugpatentwatch.com/blog/indian-pharmaceutical-patent-prosecution-the-changing-role-of-section-3d/
  8. IP Helpdesk EU – WIPI 2024 (India patent statistics): https://intellectual-property-helpdesk.ec.europa.eu/news-events/news/world-intellectual-property-indicators-2024-india-moved-sixth-place-patent-applications-2024-11-11_en
  9. WIPO data on patent filings by country, 2018-2023 — WIPO IP Statistics Data Center