The updated patent data over the past two years confirm that GenAI has entered a new phase of technological and commercial development. What emerged as an acceleration in the 2024 WIPO Patent Landscape Report: Generative Artificial Intelligence
Several structural shifts in the GenAI patent landscape are also becoming clearer. LLMs have emerged as the dominant architectural model, overtaking earlier generative approaches such as GANs. Diffusion models and multimodal systems are expanding the innovative scope of GenAI, while emerging technologies such as reasoning models and agentic systems are beginning to appear in patent filings. At the same time, the range of actors participating in GenAI patenting is broadening, with telecommunications firms, infrastructure providers, industrial companies and utilities joining traditional software and internet companies as leading applicants.
The geographic distribution of innovation also continues to evolve. China remains the largest source of GenAI patent publications by a wide margin, but strong recent growth in the United States, Japan and several European countries suggests that the global innovation landscape is becoming more dynamic. Japan’s rapid rise in recent patent activity illustrates how the entry or strategic shift of a single major player can reshape national innovation statistics within a short period.
At the same time, several characteristics of the patent landscape highlight the relatively early stage of commercialization. International patent families (IPFs) still account for a small share of GenAI filings, indicating that many inventions remain concentrated in domestic markets or have not yet reached the stage where applicants pursue global protection strategies. As GenAI applications mature and business models stabilize, it is likely that a greater share of patents will be extended internationally.
Looking ahead, the pace of change in GenAI suggests that the patent landscape will continue to evolve rapidly. Areas such as multimodal systems, reasoning-oriented models and AI agents, as well as efficiency improvements in training and inference, are likely to become increasingly important directions for both research and patenting. In addition, the integration of GenAI into sectors such as healthcare, manufacturing, finance and energy is likely to expand the diversity of actors filing patents in this field.
For policymakers, businesses and IP professionals, the implications are clear. The GenAI patent landscape is becoming larger, more global and more strategically significant. Monitoring these developments will remain essential for understanding technological trajectories, competitive dynamics and the evolving role of IP in one of the fastest-moving fields of innovation ever observed.