AI for Onshore Manufacturing - An IP Analysis
There are a few key factors to enable onshore, near-market manufacturing - this can mean US-based manufacturing or simply more efficient manufacturing in any location. It seems media headlines prefer to pit major economic powers against each other... Regardless of how you look at it, efficient, future-looking manufacturing systems benefit from advancements in and implementation of AI and flexible manufacturing technologies. Artificial intelligence and automation focus on smarter machines. These advancements reduce the time and labor needed to perform the most menial tasks, allowing for a leaner workforce that is generally more skilled. Labor focus can be re prioritized to process management, technology upkeep & configuration and other higher value tasks. Flexible manufacturing generally refers to technologies to improve process, control manufacturing costs, reduce inventory on hand, and support quick changeovers and customization.
There are plenty of discussion about the benefits (and limitations) of these technologies. This article series will look at the IP landscape and see what insights can be gleaned:
Artificial Intelligence
Modern Materials Handling recently published an article entitled "Manufacturing in the USA vs. China: AI could be the tipping point"
The first two US companies mentioned, Microsoft and Amazon are no surprise to see in the discussion. Certainly these two companies are building enabling technologies to provide to industrial customers. While Microsoft finds itself in the top 5 globally for AI patent holdings, Amazon comes in around 70th with an AI portfolio about 10% the size of Microsoft. As far as US service providing companies go, although not mentioned in the above article, Google and IBM need to be in the conversation, joining Microsoft in the top 5.
Looking specifically into industrial-AI focused companies, the article identifies a handful of interesting startups in the space. These companies have technologies ranging from software algorithms to edge operation to IoT applications. Here is some data on these startups:
100-200 patents:
Uptake - Predominately US, with additional filings in China, Canada, Australia, Europe and Singapore (in that order).
50-99 patents
Zymergen - Predominately US, with additional filings in Canada, Europe, Korea and China (in that order).
C3 - Predominately US, with additional filings in Canada, Europe, China and Australia (in that order).
10-49 patents
Sight Machine - Global filing approach focusing on US, China, Canada and Europe.
FogHorn - Global filing approach focusing on US, China, Australia, Europe, Japan and Korea.
SparkCognition - Very domestic US focus.
Falkonry - Predominately US, with additional filings in Europe, Canada and Japan.
Patent filings for this group have been increasing 20%-30% year-over-year. The above charts show the combined global distribution for the group.
In comparison, major industrial solution providers such as ABB, GE, Siemens, and Honeywell are not sitting idle, relying on technology development by others. In fact, Siemens itself has a spot in the top 5 global patent holders in the AI space (with nearly 3000 patents), while GE, Honeywell and ABB all claim spots in the top 50 or so. Other industrial players that should be in the conversation include Hitachi, and Bosch.
Looking now to China, Huawei, Tencent, and Baidu make up three of the top five AI IP portfolios for Chinese companies. This group falls around the top 50 globally. Alibaba is on the list, just a bit further down with a smaller AI portfolio. One thing to keep in mind in China is the prevalence of technology and IP creation at universities. A full thirteen of the top 25 Chinese AI patent holders are universities. Not surprisingly, Tsinghua and CAS top that list. Incidentally, the combined holdings of all of the universities in China rival the top 5 global patent holders in this space.
What is interesting is that within this cache of patents, the vast majority are filed only in China. The pie chart to the right shows in red the percentage filed in China, in blue those filed in the US, and grey represents everywhere else. This likely means that the focus of these inventions is on the Chinese domestic market. One takeaway here is that there is a high volume of research-driven AI innovation that is not leaving China's borders. Another, more accurate, way to look at this is that there is a high volume of research-driven AI innovation that is not protected outside China's borders. That difference is actually big, especially when these universities collectively would fall into the top 5 patent holders globally. The implication is that those inventions protected only in China become public domain to the rest of the world - something that may be especially useful for manufacturers specifically trying to operate outside of China.