){kind=logo}
The World Intellectual Property Organization (WIPO) recently published its 2026 Technology SPARK Report on patent trends in generative artificial intelligence (GenAI), expanding on WIPO's 2024 Patent Landscape Report. The report draws on global patent publication data to track how the GenAI innovation landscape is evolving across geographies, industries and technology categories.
The key takeaway will not surprise those following the rapid development of AI technology and its uses: growth of the GenAI patent landscape is accelerating dramatically as it broadens in both technological scope and geography. As noted in the report:
[T]he GenAI patent landscape is evolving faster than ever, and the strategic decisions being made now—about where to file, what to protect and how to navigate an increasingly crowded field—will shape competitive positions for years to come.
A Two-Year Surge That Dwarfs the Prior Decade
Over 56,000 new GenAI patent families were published in just 2024 and 2025. This number exceeds the entire output for the preceding 10 years combined. Keep in mind, the lag between filing and publication of up to 18 months in the U.S. and many other jurisdictions, so the more recent numbers are likely even higher. Within the context of all AI-related patents, GenAI's share of AI patenting has jumped to nearly 9%, up from roughly 6% in 2023.
Some Surprising New Filing Leaders
Chinese tech companies (Tencent Holdings, Ping An Insurance Group and Baidu) remain among the largest patent owners, and Alphabet (parent of Google) leads the way among U.S. companies in GenAI publications.
But perhaps the most striking finding is that SoftBank is now the world's largest GenAI patent holder by volume. This Japanese company published nearly 3,000 GenAI patent families, almost all of them in 2025 alone. The report notes that this may signal the emergence of a new type of patent owner—diversified conglomerates building vertically integrated AI strategies—alongside traditional tech companies.
Looking ahead, the report predicts that "the integration of GenAI into sectors such as health care, manufacturing, finance and energy is likely to expand the diversity of actors filing patents in this field.”
Geographic Concentration
The report notes that China-based inventors lead GenAI patent publications by a wide margin, and the top four countries (China, U.S., Japan and Korea) account for the vast majority of filings. But the growth elsewhere is still noteworthy. Germany, Canada and Switzerland all posted compound annual growth rates above 85%, and seven of the top 10 inventor locations have published more GenAI patent families in the last two years than in the entire prior decade.
Under the Hood: LLMs Take Over
For those more technically curious, the underlying technology mix has shifted decisively over the past couple of years. Unsurprisingly, large language models (LLMs) now dominate the patent landscape, outnumbering generative adversarial networks, or GANs (the previous leader), by roughly three to one. Diffusion models, which power some image and video generation tools, rank third but are growing quickly as a share of the whole. Text-based applications are catching up to image/video, and newer categories like software/code generation (one of the fastest-growing in relative terms), 3D modeling and agentic AI systems are starting to appear in filings.
* * *
The full WIPO report is available at Patent Analytics Reports on AI Inventions and their press release is here: GenAI Innovation Soaring, With Patent Activity Nearly Tripling in Two Years.
Our data shows that Gen-AI patents are no longer just being filed by tech companies but also by large enterprises in areas like finance, telecommunications, infrastructure and other digital services. While the initial impact of Gen-AI has been on content, the next frontier will be in its impact on the way we innovate.
https://www.wipo.int/...
-
Partner