How patent data can provide intelligence for other markets
Patents are an interesting phenomenon. Believe it or not, the number one reason why patent systems exist is to promote innovation and the sharing of ideas. Simply put, a patent is really a trade. A government has the ability to give a limited monopoly to an inventor. In exchange for this exclusivity, the inventor provides a detailed description of their invention. The application for a patent needs to include enough detail about the technology that a peer in the field could pick it up and understand how to make or practice that invention. Next, this description gets published to the world so others can read and learn from it. That's the exchange. Disclose how your invention is made, in enough detail to replicate, and you can get a patent.
It gets really interesting when you consider that the patent carries additional metadata with it. This additional data is above and beyond the technical description of the invention. Included in this data are the inventor names, addresses, the companies they work for (the patent owner), the date of the patent filing, a list of related patents/applications, and more. This metadata and the technical description of the invention make up an amazing set of data identifying research and development activity across the world. Also, since patents are issued by governments, they are inherently geographic. This means that an inventor has to apply for a patent in every country where they want protection. Add the fact that patents are quite expensive, and we are left with a set of ideas that have at least passed some minimal value threshold. That willingness to spend money signals a value in the technology, specifically a value of that technology in the country/market where each patent is filed. In many ways, if you want to analyze a technology space, patent data can be better than analyzing products. The technology is described in substantial detail and, in many cases, identifies tech that has not hit the market yet.
Breaking down a patent dataset
Black Hills IP has amassed a dataset of over 100 million patent and patent application records from across the world. We not only use data published by patent offices, but we also run proprietary algorithms on that data to create additional patent records and metadata. This means we have billions of data points to use in analysis, and likely have the largest consolidated patent dataset in the world. In the artificial intelligence (AI) space alone, we have an identified set of between one hundred thousand and two hundred thousand patent records. This has been fertile ground for analysis and insight.
Breaking down this dataset, we can see ownership and trends around foundational and implementational technologies. For example, several of the large US players are covering their bases with patent filings in multiple jurisdictions, including China. Interestingly enough, the inverse is not necessarily shown. Many of the inventions from Chinese companies have their patent filings (and thus protection) limited to just China. While many large US companies in the field tend to have their patent portfolios split roughly 50/50 between US and international patent filings, the top players in China have a combined distribution with well over 75% domestic and only the remainder in international jurisdictions. This means that there is a plethora of technology protected only within the borders of China, and the implications could be significant given the push for AI technology development in China and the wealth of resources available to fuel that development.
Why does all this matter? When patents are filed in a single jurisdiction only, they are visible to the world and open the door to free use outside the country of filing. In years past, we have seen Chinese companies repurpose Silicon Valley technologies for the China domestic market. With more of a historical patent thicket in the US than in China, this strategy made sense. When development and patent protection have been strong in the US, repurposing that technology in a less protected Chinese market is not only possible, but a viable business model. What we’re seeing now in the emerging field of AI technology, specifically the implementation of such technologies, is the pendulum starting to swing back.
In an interesting reversal of roles, the publication of Chinese patents on technologies not concurrently protected in the US has the potential to drive a copying of Chinese-originated AI tech in the US market. We may see some rapid growth of implementational AI technologies in the US or other western countries, fueled by Chinese development and domestic-focused IP strategy. Of course, there are many other insights to glean out of this wealth of patent data. The use of these patent analytics in the technology space will only increase as the patent offices across the world improve their data reporting and availability. Thanks to advances by some of the major patent offices, visibility into new developments is getting easier and easier. Technology and business intelligence programs stand to gain substantially from the insights hidden in IP data.
Author: Tom Marlow