Introduction

When the WIPO Patent Landscape Report: Generative Artificial Intelligence ⁽¹⁾World Intellectual Property Organization (WIPO) (2024). Generative Artificial Intelligence. Patent Landscape Report. Geneva: WIPO. https://www.wipo.int/web-publications/patent-landscape-report-generative-artificial-intelligence-genai/en/index.html was published in 2024, based on an analysis of published patent applications up to the end of 2023, the field was already moving at extraordinary speed. The release of OpenAI’s ChatGPT in November 2022 had triggered what many described as an “iPhone moment” for GenAI: a sudden, visible shift in how the public, industry and policymakers perceived the technology.

The last two years have shown no sign of deceleration. On the contrary, the data now available prove that the GenAI patent landscape has continued to expand at an even faster pace. With tens of thousands of new patent families published in 2024 and 2025, the field has entered a new phase of technological maturity and commercial urgency.

Several technological developments have shaped the GenAI landscape over the last two years and have boosted patenting activity around the world. Following the broad public adoption of large language models (LLMs), 2024 saw a shift toward models capable of processing and generating content across multiple modalities – text, images, audio, video and code within a single system. The release of new ChatGPT versions, ⁽²⁾OpenAI (2024). Hello GPT-4o. Available at: https://openai.com/index/hello-gpt-4o/ (accessed May 29, 2026) Alphabet’s (Google's) Gemini family ⁽³⁾Pichai, S. and Hassabis, D. (2024). Our next-generation model: Gemini 1.5. Google. Available at: https://blog.google/innovation-and-ai/products/google-gemini-next-generation-model-february-2024/#sundar-note (accessed May 29, 2026) and Anthropic’s Claude series ⁽⁴⁾Anthropic (2024). Introducing the next generation of Claude. Available at: https://www.anthropic.com/news/claude-3-family (accessed May 29, 2026) established multimodal capability as the baseline expectation for frontier models – the leading-edge AI systems that represent the current state of the art.

A new architectural paradigm gained prominence in late 2024 with the release of OpenAI’s o1 series ⁽⁵⁾OpenAI (2024). Learning to reason with LLMs. Available at: https://openai.com/index/learning-to-reason-with-llms/ (accessed May 29, 2026) and comparable models from other developers. These systems extend computation at inference time, allowing models to “think” through complex problems step by step before producing a response, rather than relying solely on capabilities encoded during training. This approach has led to substantial performance gains on scientific, mathematical and coding-related benchmarks, and represents a new frontier for patenting activity around reasoning systems, inference-time scaling and model efficiency. ⁽⁶⁾Balachandran, V. et al. (2024). Inference-time scaling for complex tasks: Where we stand and what lies ahead. Available at: https://www.microsoft.com/en-us/research/publication/inference-time-scaling-for-complex-tasks-where-we-stand-and-what-lies-ahead/ (accessed May 29, 2026) 

The widespread release of capable open-weight models ⁽⁷⁾Open-weight models are AI models for which the trained parameters (weights) are publicly released, allowing developers to download, study, fine-tune and deploy the model on their own infrastructure. They differ from fully open-source AI systems, where the training code, datasets, and documentation are also released. See OECD (2025). AI openness: Balancing innovation, transparency and risk in open-weight models. Available at: https://oecd.ai/en/wonk/balancing-innovation-transparency-and-risk-in-open-weight-models (accessed June 05, 2026) – most notably Meta’s Llama family ⁽⁸⁾Meta (2024). Introducing Llama 3.1: Our most capable models to date. Available at: https://ai.meta.com/blog/meta-llama-3-1 (accessed May 29, 2026) – democratized access to frontier-level GenAI, enabling a much wider ecosystem of companies, research institutions and individual developers to build and innovate on top of foundation models. This trend was further accelerated by the emergence in early 2025 of highly efficient models from DeepSeek, ⁽⁹⁾DeepSeek-AI et al. (2025). DeepSeek-R1: Incentivizing reasoning capability in LLMs via reinforcement learning. Available at: https://arxiv.org/abs/2501.12948 (accessed May 29, 2026) which demonstrated that competitive reasoning performance could be achieved at a fraction of the costs previously assumed for training models.

Perhaps the most widely discussed GenAI trend of 2025 has been the rise of agentic artificial intelligence (agentic AI) – systems that can autonomously plan, execute multi-step tasks and adapt their behavior based on real-time feedback loops, with minimal human oversight. Unlike earlier GenAI tools that respond to individual prompts, agentic systems can deconstruct complex goals into sequences of actions, use external tools and data sources, and iteratively refine their outputs. ⁽¹⁰⁾Abou Ali, M., Dornaika, F. and Charafeddine, J. (2026). Agentic AI: a comprehensive survey of architectures, applications, and future directions. Artificial Intelligence Review, 59(11). Available at: https://link.springer.com/article/10.1007/s10462-025-11422-4 (accessed May 29, 2026) While agentic AI is still a nascent field, early patenting activity is already visible, with companies such as Alphabet (Google) and Nvidia among the first movers. As these systems mature, they are likely to become an increasingly important category within the GenAI patent landscape.

This WIPO Technology SPARK Report on GenAI serves as an update of the 2024 WIPO Patent Landscape Report. It is based on the same analytical framework, examining GenAI patents from different perspectives by analyzing global trends, top inventor countries and top patent owners, as well as major GenAI models and modes. Where the original report covered patent data from 2014 to 2023, this update extends the analysis through to 2025, incorporating the most recent patent publication data available. The focus is on what is new: how rankings have shifted, which trends have accelerated or moderated, and what the data for 2024 and 2025 reveal about the direction in which the field is heading. ⁽¹¹⁾For a comprehensive overview of the GenAI patent landscape, including detailed methodology and background on the technology, please refer to the full WIPO Patent Landscape Report on GenAI published in 2024. See World Intellectual Property Organization (WIPO) (2024). Generative Artificial Intellligence. Patent Landscape Report. Geneva: WIPO. https://www.wipo.int/web-publications/patent-landscape-report-generative-artificial-intelligence-genai/en/index.html