Industrial future: China owns >60% of all humanoid robot patents; U.S. at 21.8%; Europe facing marginalization in another field of engineering

Context: At the start of the year, ip fray explained why Europe is effectively being colonized for lack of innovative capacity and entrepeneurial dynamics (January 1, 2026 ip fray article). Nothing has developed more favorably for Europe since that article. If anything, AI progress in the U.S. and China has been impressive in 2026, while Europe is accelerating on a downbound spiral as we speak.

What’s new: It has apparently gone unnoticed by IP media that a South Korean parliamentarian obtained useful data from his country’s Intellectual Property Office based on data from the world’s five major patent offices (September 28, 2026 Korea Herald article):

  • Chinese companies have grown their share of humanoid robot patents from 51.8% in 2024 to 60.8% in the first half of 2026. This growth came at the expense of all other regions: U.S. companies’ share of the patent pie declined from 27.3% to 21.8%, and South Korean companies went from 6.3% to 4.6%, which is still very strong compared to Europe when considering population size. Japan, Europe, and the rest of the world collectively own only 12.8% (down from 14.6% in 2024). Given that Japan is traditionally very strong in that field, it means Europe is not even an also-ran: its share is very in the low to mid single digits.
  • Robot Foundation Models (RFMs) are a newly evaluated category. Just like large language models (LLMs) are the foundation models of AI chatbots, RFMs are the software operating system of robots. China’s lead in that field is even greater: 66.3% versus the U.S. ar 14.1%, South Korea at 8.8%, and the rest of the world (Japan, Europe etc.) at 10.8%.

Direct impact & wider ramifications:

  • For now there is little patent litigation in this field, but those patent counts indicate where innovation is happening (even if direct proportionality cannot be presumed) and who will have IP leverage going forward.
  • The economic consequences for Europe are brutal:
    • For a long time, European politicians used to point to hardware-intensive product categories as an opportunity for Europe. That was incompetent (alternatively, just propaganda) and incorrect at the same time: incompetent because software is taking over more and more functions (the automotive sector is a good example), and incorrect because Europe has lost ground even in next-generation hardware products.
    • The EU’s weakness in technology, especially anything related to AI, accelerates the continent’s relative impoverishment. Latin America and even parts of Africa are better positioned for a future in which only tech, natural resources, and low energy costs determine competitiveness. Due to Europe’s tech weakness (“that’s what’s changed in these last 10 years”), even former British Prime Minister Tony Blair, a former opponent of Brexit who for some time advocated that the UK should rejoin the EU, now opposes the idea of rejoining the EU (November 21, 2026 BBC article).

Opinion and macroeconomic perspective

For us in the IP sector, Europe’s decline is either irrelevant because our business is international (we pride ourselves on multijurisdictional coverage) or there is, at least for the foreseeable future, still a significant opportunity for Europe to serve as a patent battleground for mostly American and Asian innovators. That said, even the UPC (whose member states are deeply divided: November 13, 2026 ip fray article) would have better growth prospects if the European economy was growing.

But for the overall economy and public finances, patent litigation is obviously not a viable business model. What then?

In 2026, AI has made enormous progress. I personally consider Google Gemini’s progress most impressive. Anthropic is getting a lot of publicity for its IPO, and it has revolutionized software development. But if you look for a multipurpose AI system that can analyze text, numbers, and images, I consider Gemini the best system, by far and away.

Jobs are already getting lost (“junior hiring freeze” etc.). Unlike former industrial revolutions where enough new jobs were being created, that will not be the case now. Elon Musk exaggerates when he says that human beings have only three years left to earn money with their work because then they will be replaced in almost every field. There will be some delays in development as well as deployment. But unemployment rates will rise, and inexorably so. Unlike in past transformational situations, new jobs for human beings will not be created by the economy. It will require a major societal restructuring, possibly involving measures such as disallowing people from working too many hours and creating many jobs for which there is no sufficient market demand (such as hiring more teachers). That’s a means of giving more people the opportunity to work and, therefore, a sense of purpose, but somewhere the money must be made to pay for that.

China and the U.S. will face a major redistributive challenge, but at least they will have substantial income from those new technologies to redistribute: “spreading the wealth” as Barack Obama once called it. Latin America, parts of Africa, probably also Australia, and maybe to some extent Canada have natural resources (including agriculture) to offer. Europe has none of that. Even its agriculture is actually weak compared to population size. Export numbers are inflated by high-priced alcoholic beverages, chocolate etc., but Europe already needs to import almost 75% of the protein-rich oilseeds and legumes used as animal feed. In other words, Europe can’t even feed its cattle.

What exacerbates the situation in Europe is that the continent failed to reform its pension system decades ago. Major European countries state their public debt in a narrow sense, leading the uninformed to believe that debt-to-GDP ratios are “low” or “acceptable” compared to that of the U.S., but if those countries had to produce balance sheets like corporations, the reality is that their debt load is a multiple of their GDP. Even a European Commission web page compares pensions obligations to GDP. Unlike public debt in a narrow sense, there are no interest payments, but those obligations have an interest-like effect when money loses value.

In the most negative sense, the combination of that system and Europe’s tech weakness is dynamite. It’s a system that depends on contributions by coming generations of the workforce. But high levels of automation are approaching fast, reducing the number of net payers. If anything, shrinking populations would actually be best-equipped for that new era, and a pension system relying on future contributions is a ticking time bomb. It is already ticking louder and louder. Look at France for example. That is the country that will most likely start the collapse of the euro. There are also some other candidates and Germany is catching up fast. The last euro crisis was triggered by Greece with 300 billion euros of debt. France is ten times more, and a country like Germany is far weaker now than it was when Greece and some others needed a bailout.

Taxing robots and AI systems would work for highly competitive economies. If Europe tried that, it would reduce its competitiveness to a greater extent than it can afford.

Europe’s mainstream media don’t talk about this. Most journalists don’t even understand this. Some do (I’ve had private conversations with a few), but can’t tell their readers the whole truth. Nor can governments be forthcoming about it. There is no potential for a turnaround due to network effects and economies of scale (for example, gigawatt data centers operate far more efficiently, and Europe’s energy policy is the result of ideological irresponsibility). The only way for Europe’s economy is down. In my opinion, the euro is still grossly overvalued at $1.10, but forex is a highly liquid market in which they all believe they will get out in time if they have to.

One lie that propaganda-oriented economists in Europe tell is that there is a huge opportunity for Europe at the application level. We in the IP space know that this is not how it works. We understand that the one who controls the wizardry can ultimately impose ad valorem pricing. European companies are not just going to use AI services at a low cost and then reap the profits of automation. Realistically, half or more of what they save will go to those who own the IP and the infrastructure.

Foundation models are not “commodity”. “Commodity” is an applicable term when there are easily substitutable offerings from many vendors, none of whom has leverage. AI is an oligopoly in a differentiated market. Even “open source” is not a free lunch. Android is a good example. There are ways in which the actual innovators and right holders can still extract value.

Most people don’t even understand the limits of trade balances. When it comes to physical goods shipped from A to B, it doesn’t look as bad for Europe as the situation actually is. Once one factors is in IP, and the IP-based businesses in Ireland that are formally intra-EU trade but just subsidiaries of U.S. companies, it’s clear that Europe has already been colonized to quite some extent and that this colonization will advance further. Every single year.

Those who own technological foundations are in a far stronger position than those who try to build applications on top of them. This year alone, countless AI “applications” have seen their value reduced to zero because of foundation models and AI chatbots becoming ever more powerful and multifunctional. That’s the equivalent of smartphones having replaced cameras, alarm clocks etc. for the largest part.

IP matters, especially in the technologies that represent the future of the economy. Humanoid robots and robot foundation models are fields worth watching. And sooner or later there will be some interesting patent infringement cases in that field.