Let’s connect
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
VettaFi recently sat down with Morten Paulsen, head of research for robotics & machinery at CLSA, to discuss the transition of physical AI into a tech-driven industrial up-cycle. Paulsen projects that persistent U.S. labor shortages will drive domestic robot shipments toward a historical high of 40,000 units in 2026.
Paulsen is a strategic advisor for VettaFi’s ROBO Global Indexes, which underlie the ROBO Global Artificial Intelligence ETF (THNQ) and the ROBO Global Robotics and Automation Index ETF (ROBO).
Paulsen: Yes, when we look at robots globally, “Physical AI” was clearly the big buzzword in 2025. It started at CES in January when Jensen Huang named robotics as an enabler of AI in the physical world. Humanoid robot builders were very quick to point out that their robots were the answer to that vision.
I think it changed a bit in October 2025 when SoftBank Group acquired ABB Robotics for $5.4 billion. That was a turning point where the broader market realized physical AI wasn’t limited to robots with a human form factor, but applied to a much broader range of automation equipment. That triggered a rally in other industrial robotic names in Japan — Fanuc and Yaskawa, for example.
Regarding investor interest in Asia… In China, Hong Kong, and Singapore, the investor base has been very focused on the humanoid form factor, whereas I see less interest in that specifically in Japan. Japan, the U.S., and Europe are more similar in that sense.
Paulsen: AI has the power to make robot applications more powerful, and I also think that we can start to see the development of new business models.
When it comes to developments — and underappreciated developments heading into 2026 — I do think that we’re going to see more focus on the robot hand, especially things like tactile sensing. If you look at the robots right now, we have machine vision that lets robots “see.” With cobots, the robots have been enabled to work in proximity to humans. And then we have AMRs (autonomous mobile robots) that give the robots mobility. Also, LLMs have had an impact on the robot’s ability to understand human language and so on. But still, the missing link here is to get hands that are working properly.
Paulsen: I continue to believe that the labor shortage is going to be one of the fundamental underlying drivers for robotics and automation demand. Regarding the correlation between job openings and robot shipments [that I’ve highlighted in my research], what we saw now in the temporary numbers in 2025 is that robot shipments are still slightly under-indexed compared to the labor shortage that we saw in 2024, but the big gap that we saw during COVID has narrowed.
Job openings in manufacturing have also come down, but we’re still looking at a pretty big shortage. On average in 2025, we’re looking at about 400,000 job openings in manufacturing. If this historical link holds, we should see about 40,000 robots shipped in the U.S. in 2026. That would be very close to an all-time historical high. It would indicate about 30% year-over-year growth.
Paulsen: China as an end market performed a lot better than initially expected in 2025. The automotive and electronics sectors remained very active. With the Trump tariffs — many of those specifically targeting China — we were more worried at the beginning of the year, but China held up incredibly well. One reason for that is replacement demand, but also because they were still investing in new automotive technologies. I think it’s worth highlighting that foreign companies actually reported fairly good growth rates in China in 2025, including FANUC and Yaskawa.
Regarding Chinese robot makers looking to increase abroad and the extent they represent a competitive threat… I do think that they are a competitive threat within China, but as we can see on the growth rates in ’25, the foreign companies still have a lot to offer.
Chinese robot companies started to show up at trade shows in Europe and Japan [in 2025]. But so far, there has been a limited market impact. I do think that could change, but it’s not an easy market to break into. You have high entry barriers, especially on the distribution side in terms of safety certifications. It’s usually not enough just to have a low price point; the quality requirements are very, very high. Manufacturers would be very cautious about picking systems. Automation equipment is part of a bigger system, and the system is only as strong as the weakest link.
Paulsen: I think it is more interesting to talk about robots doing things humans can’t replicate — looking beyond what humans can do. [One example is] nano-robots that can go into the body. That is still in the research phase, but really interesting.
We can also think of robot types that don’t exist. Why shouldn’t we put effort into making a, let’s say, rare earth mining robot? Everyone can instantly see that could be a very good idea. But we should start to think about what is needed — the best form factor, mechanical structure, and software package needed to move into these new areas. Once you start with that way of thinking, it’s almost unlimited where robots can go.
Morten Paulsen: I did a presentation a year ago where I had predictions on ’25. I think accountability is important if I make a prediction — I like to look back and see what went wrong and right. A year ago, I presented a fairly optimistic view on the cycle. ’25 turned out to be a bit more volatile than that.
It started off very well — manufacturing PMIs above 50 in the first quarter — but then the second quarter came with a lot of volatility with “Liberation Day,” which brought a lot of uncertainty. But then things picked up in the third and the fourth quarter. One year ago, I said we were about to enter an up-cycle. Today, I would say we are already in an up-cycle.
That was largely ignored by the market, and it was really the thematic angle that brought investors back. But from a cyclical perspective, I do believe we are in an up-cycle. I believe this initially is going to be a tech-driven up-cycle, but I do expect automotive capex to come back towards the end of the year. We are going into what we can look back at as a pretty long up-cycle, but at a slower pace than the “sugar rush” after COVID in ’21 and ’22.
This article was originally published January 26th, 2026 on ETF Trends.
VettaFi recently sat down with Morten Paulsen, head of research for robotics & machinery at CLSA, to discuss the transition of physical AI into a tech-driven industrial up-cycle. Paulsen projects that persistent U.S. labor shortages will drive domestic robot shipments toward a historical high of 40,000 units in 2026.
Paulsen is a strategic advisor for VettaFi’s ROBO Global Indexes, which underlie the ROBO Global Artificial Intelligence ETF (THNQ) and the ROBO Global Robotics and Automation Index ETF (ROBO).
Paulsen: Yes, when we look at robots globally, “Physical AI” was clearly the big buzzword in 2025. It started at CES in January when Jensen Huang named robotics as an enabler of AI in the physical world. Humanoid robot builders were very quick to point out that their robots were the answer to that vision.
I think it changed a bit in October 2025 when SoftBank Group acquired ABB Robotics for $5.4 billion. That was a turning point where the broader market realized physical AI wasn’t limited to robots with a human form factor, but applied to a much broader range of automation equipment. That triggered a rally in other industrial robotic names in Japan — Fanuc and Yaskawa, for example.
Regarding investor interest in Asia… In China, Hong Kong, and Singapore, the investor base has been very focused on the humanoid form factor, whereas I see less interest in that specifically in Japan. Japan, the U.S., and Europe are more similar in that sense.
Paulsen: AI has the power to make robot applications more powerful, and I also think that we can start to see the development of new business models.
When it comes to developments — and underappreciated developments heading into 2026 — I do think that we’re going to see more focus on the robot hand, especially things like tactile sensing. If you look at the robots right now, we have machine vision that lets robots “see.” With cobots, the robots have been enabled to work in proximity to humans. And then we have AMRs (autonomous mobile robots) that give the robots mobility. Also, LLMs have had an impact on the robot’s ability to understand human language and so on. But still, the missing link here is to get hands that are working properly.
Paulsen: I continue to believe that the labor shortage is going to be one of the fundamental underlying drivers for robotics and automation demand. Regarding the correlation between job openings and robot shipments [that I’ve highlighted in my research], what we saw now in the temporary numbers in 2025 is that robot shipments are still slightly under-indexed compared to the labor shortage that we saw in 2024, but the big gap that we saw during COVID has narrowed.
Job openings in manufacturing have also come down, but we’re still looking at a pretty big shortage. On average in 2025, we’re looking at about 400,000 job openings in manufacturing. If this historical link holds, we should see about 40,000 robots shipped in the U.S. in 2026. That would be very close to an all-time historical high. It would indicate about 30% year-over-year growth.
Paulsen: China as an end market performed a lot better than initially expected in 2025. The automotive and electronics sectors remained very active. With the Trump tariffs — many of those specifically targeting China — we were more worried at the beginning of the year, but China held up incredibly well. One reason for that is replacement demand, but also because they were still investing in new automotive technologies. I think it’s worth highlighting that foreign companies actually reported fairly good growth rates in China in 2025, including FANUC and Yaskawa.
Regarding Chinese robot makers looking to increase abroad and the extent they represent a competitive threat… I do think that they are a competitive threat within China, but as we can see on the growth rates in ’25, the foreign companies still have a lot to offer.
Chinese robot companies started to show up at trade shows in Europe and Japan [in 2025]. But so far, there has been a limited market impact. I do think that could change, but it’s not an easy market to break into. You have high entry barriers, especially on the distribution side in terms of safety certifications. It’s usually not enough just to have a low price point; the quality requirements are very, very high. Manufacturers would be very cautious about picking systems. Automation equipment is part of a bigger system, and the system is only as strong as the weakest link.
Paulsen: I think it is more interesting to talk about robots doing things humans can’t replicate — looking beyond what humans can do. [One example is] nano-robots that can go into the body. That is still in the research phase, but really interesting.
We can also think of robot types that don’t exist. Why shouldn’t we put effort into making a, let’s say, rare earth mining robot? Everyone can instantly see that could be a very good idea. But we should start to think about what is needed — the best form factor, mechanical structure, and software package needed to move into these new areas. Once you start with that way of thinking, it’s almost unlimited where robots can go.
Morten Paulsen: I did a presentation a year ago where I had predictions on ’25. I think accountability is important if I make a prediction — I like to look back and see what went wrong and right. A year ago, I presented a fairly optimistic view on the cycle. ’25 turned out to be a bit more volatile than that.
It started off very well — manufacturing PMIs above 50 in the first quarter — but then the second quarter came with a lot of volatility with “Liberation Day,” which brought a lot of uncertainty. But then things picked up in the third and the fourth quarter. One year ago, I said we were about to enter an up-cycle. Today, I would say we are already in an up-cycle.
That was largely ignored by the market, and it was really the thematic angle that brought investors back. But from a cyclical perspective, I do believe we are in an up-cycle. I believe this initially is going to be a tech-driven up-cycle, but I do expect automotive capex to come back towards the end of the year. We are going into what we can look back at as a pretty long up-cycle, but at a slower pace than the “sugar rush” after COVID in ’21 and ’22.
This article was originally published January 26th, 2026 on ETF Trends.
