The owner of the biggest number of AI patent applications, IBM, has a portfolio of 8,290 inventions. They are closely followed by the company Microsoft with 5,930. Both companies have patent registrations and applications from various fields and covering many techniques, thus proving that they are not limited to a specific industry. The other top applicants are Toshiba (with 5,223 applications), Samsung (with 5,102) and NEC (with 4,406). Leaping into the top 20 is The State Grid Corporation of China, by annually increasing its patent filings by an average of 70 percent, since 2013, specifically in the machine learning techniques of bio-inspired approaches, born out of from observations of nature, and support vector machines, a type of supervised learning.
Certain fields of expertise have the highest numbers of patent applications generated by companies with a high degree of specialization and expertise, such as Baidu, (ranks at the top for deep learning), Bosch and Toyota, (that stand out in transportation), as well as companies such as Siemens, Philips and Samsung in life and medical sciences. Interestingly, some well-known companies are not included among the top overall players in AI patents; a worthy mention here would be Facebook and Tencent in networks and social networks. This also comes to show that access to specialized data explains why certain companies lead in specific industries.