China may be winning the AI patent race in terms of sheer volume, with almost 13,000 granted patents, but the U.S. (8,609 patents) dominates in terms of impact. American AI patents are cited nearly seven times more often than Chinese patents (13.18 vs 1.90 average citations). Many Chinese firms like the Industrial and Commercial Bank of China and State Grid Corporation file exclusively within China, while larger firms such as Huawei and Tencent have a more diversified AI patenting strategy. Similarly, companies such as Intel Corporation and South Korean tech giant Samsung Electronics Co., Ltd. file across 10 or more jurisdictions.
China leads in overall GenAI patents, too
This pattern extends to generative AI (GenAI) patents as well. In a study covering 2014-2023, WIPO found that China-based inventors filed 38,210 GenAI inventions, six times more than the U.S.’s 6,276 inventions. The surge in GenAI innovation has been particularly dramatic, with patents increasing eightfold since 2017’s introduction of deep neural networks behind Large Language Models. More than 25% of all GenAI patents were published in 2023 alone, though GenAI still represents just 6% of all AI patents globally in WIPO’s analysis.
Japan, South Korea and Germany had strong showings
In the first ten months of 2024, South Korea and Japan, each surpassing 1,500 AI patents, represent the next tier of innovation hubs. Japanese companies show stronger citation impact with 6.26 citations per patent compared to South Korea’s 3.12, though both significantly outperform China-based organizations in terms of citation rates.
Next in line is Germany, which continues to be a leader in AI. With 784 total AI patents in the dataset and an average citation rate of 6.12 (more than triple China’s 1.90), German companies demonstrate strong patent quality. The German AI patent landscape spans multiple sectors: industrial giant Siemens AG leads with 268 patents, while automotive supplier Robert Bosch GmbH holds 197 patents across two entities, and enterprise software leader SAP SE contributes 101 patents with an especially high citation rate of 15.59.
Rounding out the German landscape are the research institution Fraunhofer (23 patents, 11.35 citations), Bayer AG (16 patents, 11.44 citations), and a strong automotive cluster including Continental, Mercedes-Benz, Volkswagen, and BMW. German companies typically file across about four to seven jurisdictions, with Siemens and Bosch showing especially broad international coverage across AU, CA, CN, EP, GB, JP, US, and WO markets.
Interactive heat-density map of the 2024 AI patent landscape to date
Geographical breakdown of AI patents in 2024
China and the U.S. stand far above other countries in terms of overall AI patents. Japan and South Korea show identical patent volumes of 1,537 each in the dataset, though Japan achieves this with slightly fewer companies (41 vs. 43) resulting in a higher patents-per-company ratio of 37.5 compared to South Korea’s 35.7.
Rank | Country | Total AI Patents (2024*) | Number of Companies in Dataset | Patents per Company | Key Jurisdictions |
---|---|---|---|---|---|
1 | China | 12,945 | 523 | 24.8 | CN, US, EP |
2 | United States | 8,609 | 257 | 33.5 | US, EP, CN, JP |
3 | South Korea | 1,537 | 43 | 35.7 | KR, US, CN, EP |
4 | Japan | 1,537 | 41 | 37.5 | JP, US, CN, EP |
5 | Germany | 784 | 18 | 43.6 | EP, US, CN |
6 | United Kingdom | 369 | 11 | 33.5 | GB, EP, US |
7 | Netherlands | 249 | 7 | 35.6 | EP, US, CN |
8 | Sweden | 243 | 4 | 60.8 | EP, US |
9 | Finland | 180 | 1 | 180.0 | EP, US |
10 | Taiwan | 156 | 11 | 14.2 | TW, US, CN |
*Data for the first roughly 10 months of 2024. “AI” broadly defined here using all relevant WIPO PATENTSCOPE Artificial Intelligence Index CPC codes. The table draws data from more than 930 independent organizations.
Mohamed Ahmed says
This is super helpful as I am trying to understand patents and innovation.
Brian Buntz says
Hi Mohamed,
Thanks for the feedback. Any particular focus you have on understanding patents and innovation? Might be able to provide more on the subject.