GenAI Patent Boom Global Inventions Nearly Triple in Just Two Years

Global innovation around generative artificial intelligence has accelerated sharply, with published GenAI patent families rising from about 14,000 in 2023 to more than 37,800 in 2025, according to the World Intellectual Property Organization (WIPO). More than 56,000 GenAI patent families were published across 2024 and 2025 combined, exceeding the total from the previous decade. Large language models have overtaken GANs as the leading model category, while image and video generation, text, software/code and 3D technologies are expanding rapidly.

Generative AI is no longer simply a software trend measured by the number of chatbots, image generators or AI applications entering the market.

The technology is increasingly becoming an intellectual-property race.

New data from the World Intellectual Property Organization (WIPO) shows that the number of published patent families related to generative AI increased from approximately 14,000 in 2023 to more than 37,800 in 2025. That represents nearly a threefold increase in only two years.

Even more striking is the pace of recent activity: more than 56,000 new GenAI patent families were published during 2024 and 2025 combined, exceeding the cumulative number published during the entire 2014–2023 period.

The numbers suggest that the next phase of the AI race is increasingly about protecting the underlying technologies that make generative systems useful across industries.

GenAI Patent Activity Has Entered a New Phase

WIPO’s latest Patent Trends Update in GenAI, published in 2026, extends its earlier analysis through 2025.

The organization uses published patent families as a measure of new inventions. According to the updated data, GenAI inventions roughly doubled from 18,862 in 2024 to 37,808 in 2025.

GenAI now represents 8.7% of all AI-related patent family publications, compared with 6.1% in 2023 and 4.2% in 2017.

That change reflects how quickly generative systems have moved from experimental research into commercial development.

The shift accelerated following the emergence of large-scale generative AI systems after 2022, when increasingly capable models demonstrated that AI could generate text, images, software, audio and other forms of content at scale.

Large Language Models Have Taken the Lead

One of the most significant changes identified by WIPO is the rise of large language models (LLMs).

LLMs have now overtaken generative adversarial networks, or GANs, as the largest GenAI model category by patent volume. WIPO reports approximately 14,100 LLM-related patent family publications in 2025, compared with around 5,200 for GANs.

The change reflects the evolution of generative AI itself.

GANs played an important role in earlier generations of image-generation technology. Today’s GenAI ecosystem, however, increasingly revolves around transformer-based architectures and large language models capable of processing and generating complex sequences of information.

LLMs are also becoming foundations for systems that work beyond text, including multimodal AI capable of processing combinations of text, images, audio, video and other data.

Image and Video AI Remain Major Innovation Areas

Although LLMs are rapidly expanding, generative AI patent activity is not limited to chatbots.

WIPO’s data shows that image and video generation remains the largest GenAI output category. Between 2014 and 2025, approximately 40,000 published patent families were associated with image and video generation, including more than 13,800 published in 2025 alone.

Text-related patent activity is also accelerating.

WIPO says text-related patent filings more than tripled over two years, narrowing the gap with image and video technologies. Software and code generation, 3D modelling and other emerging application areas are also expanding.

This means the GenAI patent landscape is becoming increasingly multimodal rather than being concentrated around a single type of generated content.

AI Is Moving Into Science, Healthcare and Industry

The expansion of GenAI patents also reflects the technology’s movement into specialized industries.

Generative systems are increasingly being developed for applications involving scientific research, healthcare, engineering, finance, telecommunications, infrastructure and industrial processes.

WIPO notes that major patent activity is no longer coming exclusively from companies traditionally associated with software or internet services. Large enterprises in areas such as finance, telecommunications, infrastructure and digital services are increasingly seeking intellectual-property protection for GenAI innovations.

This broadening matters because it suggests that GenAI is increasingly being treated as an enabling technology rather than a standalone consumer product.

The underlying AI can become part of a much larger industrial system.

China Remains the Largest Source of GenAI Patents

The geographical distribution of GenAI innovation is another major feature of the WIPO data.

China remains the largest source of GenAI patent publications by a substantial margin. WIPO reports that more than 43,000 GenAI patent families were published by China-based inventors during 2024 and 2025. China’s GenAI patent-family publication growth between 2023 and 2025 corresponds to a compound annual growth rate of 64%.

The United States is also seeing rapid growth, with a 92% CAGR over the same period.

Japan recorded a particularly sharp relative increase, with a 210% CAGR between 2023 and 2025, moving from fourth to third among the leading inventor locations. Germany also recorded strong growth, while Canada and Switzerland posted triple-digit growth rates.

WIPO’s 2025 ranking places China, the United States, Japan, South Korea and India among the top five GenAI-patenting economies.

India Appears Among the Top Five GenAI-Patenting Economies

India’s inclusion in WIPO’s top five GenAI-patenting economies is particularly relevant to the country’s growing AI research ecosystem.

The WIPO data places India behind China, the United States, Japan and the Republic of Korea in its 2025 ranking of GenAI-patenting economies.

The patent data does not by itself measure the commercial value or scientific quality of each invention. But it does provide an indication of where organizations and researchers are seeking intellectual-property protection for emerging GenAI technologies.

For India, that places the country’s research institutions, technology companies and startups within a rapidly expanding global patent landscape.

The Companies Behind the Patent Race Are Changing

The GenAI patent race is also becoming broader at the corporate level.

WIPO identifies Japan’s SoftBank as the leading GenAI patent applicant, with nearly 3,000 patent families published during 2024 and 2025. It is followed by Tencent Holdings, Ping An Insurance Group and Baidu. Research organizations and infrastructure companies also appear prominently.

WIPO’s 2025 data also places the Chinese Academy of Sciences, State Grid Corporation of China, Alphabet/Google, Zhejiang University, Microsoft and IBM among the top patent owners.

The composition of this group illustrates how GenAI is spreading beyond companies whose primary business is building AI models.

Infrastructure operators, universities, financial institutions and industrial organizations are increasingly developing their own AI-related intellectual property.

Why Patents Matter in the GenAI Race

Patent activity does not automatically translate into successful products.

A patent may protect a technical invention, but commercialization depends on factors including engineering performance, data, computing infrastructure, market demand, regulatory requirements and the ability to scale.

Nevertheless, patents provide a window into where organizations are investing research and development resources.

For companies, intellectual property can help protect technical approaches and create potential licensing opportunities. For researchers and policymakers, patent analytics can reveal emerging technological directions and areas of increasing competition.

WIPO describes patent monitoring as a way to understand where GenAI innovation is accelerating and how competitive dynamics are changing.

The AI Patent Race Is Moving Beyond Chatbots

The biggest message from the latest WIPO data may be that the GenAI revolution is becoming much larger than conversational AI.

The technology is expanding across language, images, video, software, 3D modelling and specialized scientific and industrial applications. At the same time, the companies seeking patents increasingly come from sectors outside traditional technology.

That changes the nature of the competition.

The first phase of the generative AI boom was largely defined by the public arrival of powerful models and consumer applications. The next phase is increasingly about the underlying technologies, specialized models, multimodal systems and industrial applications that can turn AI capabilities into proprietary products and processes.

With published GenAI patent families rising from roughly 14,000 in 2023 to more than 37,800 in 2025, WIPO’s data shows just how quickly that transition is taking place.

The numbers do not establish which companies or countries will ultimately dominate the technology. They do show that the race to develop and protect the next generation of generative AI has already expanded into one of the world’s most active areas of technological innovation.