Over the past ten years, artificial intelligence has taken on a central role in China\u2019s economic and strategic development. Its advance is the result of long-term planning involving the central government, local administrations, major technology companies, universities and research centres. It finds its framework in the 15th Five-Year Plan, which sets the goal of technological and scientific self-reliance by 2030. In this vision, intellectual property is a pillar of national competitiveness.<\/p>\n\n\n\n
The WIPO report \u201cPatent Trends Update in GenAI\u201d, published in 2026, documents the outcomes of this trajectory in a sector that has moved from experimentation to ever wider commercial and industrial deployment. In the 2024-2025 biennium, more than 56,000 new patent families in generative AI were published, more than the entire cumulative output of the 2014-2023 decade.<\/p>\n\n\n\n
China retains the absolute lead by volume, with more than 43,000 patent families attributable to inventors located in the country over the two-year period. Tencent, Ping An and Baidu are among the largest holders. Growth extends to other systems as well: between 2023 and 2025, compound annual growth rates reach 210% in Japan, 124% in Switzerland, 109% in Canada and 92% in the United States.<\/p>\n\n\n\n
IP rights as competitiveness tools<\/strong><\/p>\n\n\n\n China\u2019s lead rests in large part on intellectual property. The policy of supporting patenting makes it possible to build, alongside the development of the technologies, portfolios designed for their economic exploitation. A patent gives the owner the right to prevent third parties from exploiting the invention, within the territory and the limits of the claims, and therefore also to permit such exploitation on negotiated terms through licences and technology agreements.<\/p>\n\n\n\n The choice of protection instrument remains a strategic decision. Comparing the portfolios of the large Chinese groups with that of OpenAI, which according to WIPO held 35 filings worldwide at the end of 2025, highlights two different orientations: extensive investment in patenting on the one hand, and greater reliance on trade secrets and speed of execution on the other.<\/p>\n\n\n\n A patent confers an exclusive right enforceable against third parties within the territorial scope of protection, as a rule for twenty years from the filing date of the application. In exchange, it requires publication of the application and a description sufficient to enable a person skilled in the art to carry out the invention. A trade secret, by contrast, requires no registration and lasts for as long as its requirements continue to be met, but it does not confer an exclusive right. It protects confidential information that has commercial value because it is secret and has been subject to reasonable steps to keep it so, against unlawful acquisition, use and disclosure, without precluding independent discovery. Its management therefore demands organisational, technological and contractual safeguards. The two instruments can be complementary, provided that what is disclosed and what remains confidential are clearly delimited, since whatever is described in a patent application ceases to be confidential once the application is published.<\/p>\n\n\n\n Patentability and guidelines<\/strong><\/p>\n\n\n\n Published patent families highlight the massive volume of inventive activity, but they do not guarantee that a corresponding number of patents will be granted. The strength of a portfolio depends on the inventions protected, the breadth of the claims and the robustness of the rights. These are the elements that determine how investment in patenting can translate into rights that can be used in the market.<\/p>\n\n\n\n In this perspective, the guidelines published by patent offices on applications relating to AI-related inventions take on particular importance.<\/p>\n\n\n\n The comparison between China and Europe is especially significant. The EPO and CNIPA converge in denying protection to abstract algorithms considered as such, but differ in their legal bases and evaluative approach. For the purposes of inventive step, the EPO takes into account only the features that contribute to technical character, whereas CNIPA assesses the solution as a whole, in terms of technical problem, technical means and technical effects. Protection must therefore be built around the technical solution, making the contribution of the algorithmic features recognisable. Description and claims must be drafted with the respective practices in mind: a result obtained in China cannot be assumed to carry over automatically to Europe, or vice versa.<\/p>\n\n\n\n Territoriality, litigation and the international landscape<\/strong><\/p>\n\n\n\n WIPO notes that families published in at least two jurisdictions, the International Patent Families, account for roughly 9% of 2025 publications. The figure reflects the weight of Chinese domestic filings and, in part, the young age of the portfolios, which may acquire a greater international dimension over time.<\/p>\n\n\n\n Since a patent is territorial, its effects arise only in the States or regions for which the right has been granted: a Chinese patent does not protect in Europe, and a European patent does not protect in China. There is no worldwide patent, and each extension requires a separate national or regional procedure under the rules of the relevant legal system. The choice of countries in which to file, tied to target markets, places of manufacture and the presence of competitors, therefore also conditions the ability to enforce the exclusive right in foreign markets. In countries not covered, the technology remains freely usable once the application is published.<\/p>\n\n\n\n The same applies to judicial enforcement: disputes must be coordinated across the competent fora, taking into account differences in the level of protection. For businesses, filing planning and cross-border litigation strategy are therefore interlinked. The compatibility between legal systems affects the ability to leverage abroad the technologies developed in the domestic market.<\/p>\n\n\n\n China v. EU competitiveness: models comparison<\/strong><\/p>\n\n\n\n The comparison between China and Europe is not limited to patent filing numbers, because it reflects two different models of competitiveness. China has built its lead on long-term industrial policy, a strong capacity for investment and a scale of data, infrastructure and filings that Europe is not at present able to match. Europe, for its part, leverages the strength of its regulatory framework, from intellectual property protection to AI regulation, and aims to make the technology trustworthy and legally compliant as well as technologically competitive.<\/p>\n\n\n\n The two models are not directly comparable, and neither is measured by volume alone. The quantity of patents indicates the direction of a strategy, but the value of rights depends on their quality, their territorial scope and the ability to enforce them. Likewise, regulation is a factor of competitiveness only if it does not become a brake on development. It is in this relationship between development, rights and rules that intellectual property becomes a lever of technological autonomy.<\/p>\n\n\n\n <\/p>\n","protected":false},"excerpt":{"rendered":" Over the past ten years, artificial intelligence has taken on a central role in China\u2019s economic and strategic development. Its advance is the result of long-term planning involving the central government, local administrations, major technology companies, universities and research centres. It finds its framework in the 15th Five-Year Plan, which sets the goal of technological […]<\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"inline_featured_image":false,"footnotes":""},"categories":[170],"tags":[213,210,218],"class_list":["post-1830","post","type-post","status-publish","format-standard","hentry","category-patents-and-utility-models","tag-ai","tag-china","tag-patents"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/posts\/1830","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/comments?post=1830"}],"version-history":[{"count":1,"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/posts\/1830\/revisions"}],"predecessor-version":[{"id":1831,"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/posts\/1830\/revisions\/1831"}],"wp:attachment":[{"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/media?parent=1830"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/categories?post=1830"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.osservatorio-ip.it\/en\/wp-json\/wp\/v2\/tags?post=1830"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}