TIPO releases analysis of global AI patent technology landscape
E251023Y1 Dec. 2025（E305）
Artificial intelligence (AI) is transforming the world, and AI technology has been a key priority to the governments around the world. To encourage enterprises to strengthen AI technology development and application and to perfect AI patenting strategies, TIPO has analyzed the recent trends in AI patenting and released its findings in terms of AI patent technology, AI hardware, knowledge processing, machine learning, evolutionary computation, vision, natural language processing, speech, and planning and control, by using data from its Global Patent Search System (GPSS). TIPO’s analysis aims to assist businesses in identifying AI future potential development opportunities and challenges.
1. Countries worldwide race to secure core AI patents and vigorously invest in AI application and development.
JPO released a “Survey on Filing Trends for AI-Related Inventions” in June 2025, outlining the status of global competition in AI patent technologies. According to the survey report, although China far outnumbers other countries in the overall volume of AI patent applications, the U.S. still leads the world in AI core patents, including the patents regarding image processing, natural language processing, large language model, etc., with its vast R&D investment and financial resources. Such precedence, however, slowed down in 2020, as indicated by the said survey report, because some businesses may consider shifting to “closed strategy” to preserve its technological advantages by trade secrets protection. South Korea sees its steady growth in AI-related patent applications, with most of them filed for smart phones, medical images, and retailing applications, while Japan has experienced a gradual decline in recent years, with its filings concentrated on medical diagnostics and medical image applications.
2. Machine learning and machine vision are the focus of AI patent portfolios.
TIPO also has identified similar findings. An analysis of the global filing trends of AI patent applications indicates a major growth in the overall number of AI-related patent applications in the past decade, with the number of filings rising above 30,000 applications in 2020. From the perspective of subfields, machine learning and machine vision emerge as major focus areas for global AI patent filings. In Taiwan, the number of overall patent applications mirrors the global pattern, showing sustained and significant growth since 2017. A comparison with the number of patent filings in 2014 reveals that the number of applications in 2021 increased by nearly ninefold, and machine learning and machine vision also account for the largest share of applications. When aggregating the number of patent applications across all AI subfields within the statistical period, the prominence of machine learning and machine vision becomes even more evident. Both globally and in Taiwan, these two subfields together account for more than half of all AI-related patent applications. Overall, with AI technologies becoming mature, AI technologies have moved out of laboratories and penetrated into all kinds of industries and will further integrate into daily human life to satisfy human needs and provide assistance in problem solving.
3. Rising AI patent filings strengthen Taiwan’s growing significance.
Among the top ten patent offices for overall AI technologies as well as the subfields of machine learning and machine vision, the differences are not significant. Notably, the patent offices ranked first to third and sixth to seventh are identical across these categories. Taiwan ranks 9th in terms of the number of patent applications for overall AI technologies, while it ranks 8th and 9th, respectively in machine learning and machine vision. This demonstrates that Taiwan is also an important jurisdiction in the global AI patent landscape. In recent years, the number of AI-related patent applications filed in Taiwan has continued to go up, and it is expected that Taiwan will become an increasingly important venue for AI patent portfolio strategy.
4. IBM leads global AI-related patent filings, while Hon Hai Precision Industry tops the list of AI-related patent applicants in Taiwan.
Among the global participants in AI-related technologies, five of them are U.S.-based companies and finish 1st, 3rd, 5th, 9th, and 10th, respectively, and three of them are China-based ones in the 2nd, 6th, and 7th positions. This ranking reflects both the U.S.’s and China’s strong overall competitiveness in AI-related patent technologies. For Taiwan, the principal participants are the five companies or legal entities, including Hon Hai Precision Industry, the Industrial Technology Research Institute (ITRI), Chunghwa Telecom, Institute for Information Industry (III), and Inventec, which secure the 1st, 2nd, 4th, 7th, and 9th places, respectively. This demonstrates that Taiwanese companies or research institutes have devoted substantial attention and investment in AI-related patent technologies, which constitutes a driving force for Taiwan industry’s continued advancement.
As to the national participants in machine learning, the U.S.-based applicants include 6 companies, ranking 1st, 3rd, 4th, 7th, 9th, and 10th, respectively. Notably, IBM, ranking 1st, has a substantial lead over the second-ranked applicant. The U.S. still retains its continued dominance in AI model development and computing capabilities. For Taiwan, the principal participants are ITRI, Chunghwa Telecom, and Hon Hai, which rank 1st, 6th, and 8th, respectively. This makes it clear that ITRI, Chunghwa Telecom, and Hon Hai have actively engaged in AI technology research and development in recent years and presented visible results.
In terms of national applicants in the field of machine vision, Chinese applicants comprise a total of nine companies and collectively rank first in terms of application volume, while South Korea has also secured a position within the top four. As for Taiwan, the principal participants include Hon Hai, ITRI, Chunghwa Telecom, National Chiao Tung University, and the Institute for Information Industry (III), ranking 1st, 2nd, 6th, 7th, and 8th, respectively, which reveals that machine vision, particularly its hardware-oriented applications, is a key focus area for Taiwan’s future development.
5. The research and development should continue to strengthen Taiwan’s technology advantages.
Taiwan possesses the world-leading capabilities to manufacture semiconductor manufacturing and the high-quality system integration ability. Notably, the list of partners displayed on the backdrop of NVIDIA CEO, Jensen Jen-hsun Huang’s keynote speech, spanning manufacturers or suppliers engaging in semiconductor manufacturing, system assembly and integration, packaging and testing, component and parts, underscores Taiwan’s pivotal role in the global AI supply chain. Moreover, the government has designated the “Five Trusted Industries”, including semiconductor industry, artificial intelligence industry, military industry, security and surveillance industry, next-generation communications industry, as the key industries for priority investment. In addition to continuing to strengthening Taiwan’s leadership in the global semiconductor industry, the government aims to drive the high-value AI applications by responding to demands across industries, build new industrial chains and ecosystems based on AI technologies, promote both AI-enabled industries and the growth of the AI industry itself, and accelerate and accomplish Taiwan’s digital transformation and industrial upgrading.
6. AI-related patent application guideline is released.
TIPO has been consistently refining and strengthening its examination capacity for AI-related patent applications amid the surge of artificial intelligence. In 2025, TIPO compiled a collection titled “AI-Related Invention Cases in Taiwan”, which provides in-depth analysis of examination standards for patentability requirements through relevant AI-related invention applications filed in Taiwan. This collection compilation was revised and completed based on the discussions and exchanges by and among experts and scholars from the industry, the academia, and research institutions engaged in invention activities at the 2025 Workshop on AI-Related Invention Cases held on June 20, 2025. TIPO also held a promotional session on September 18, 2025 to further promote awareness and understanding of the compilation. These efforts are expected to help the industry better understand examination practices for AI-related inventions, improve the quality of patent specifications for AI innovations, and ultimately strengthen global patent portfolio development. (Released 2025.10.23)
/CCS
