00:00How is AI being infused?
00:02I mean, Jensen Huang and everyone else here from Qualcomm CEO
00:05all talking about agentic AI as the next new springboard
00:10for the adoption and investment and profitability of AI.
00:13How is agentic AI being infused into a traditional industrial tool
00:18like you have that's been around since the 1970s?
00:21How is it being improved?
00:23So, I mean, ABB started in robotics in 1974.
00:27I think we were the first to introduce microprocessor controls
00:32into the robotics space.
00:34And over the next 50 years, we've come up with a very wide range
00:38of industrial, collaborative, and mobile robots.
00:42And we spent probably the last 20 years evolving that system
00:46now into robots that can operate more autonomously
00:50and in a more versatile manner.
00:53So think of robots with intelligent eyes.
00:56Think of robots with hands and ability to manipulate.
01:01And also think of robots that now have a brain.
01:04But these are evolving to a point where ultimately simulation
01:09becomes more and more important.
01:11Because you're merging so many different technologies,
01:14the ability to create hyper-realistic synthetic data
01:18to recreate these environments is truly bringing these technologies
01:23together and is really the next step that we're working on.
01:25So how are you then as well partnering with NVIDIA?
01:29Because NVIDIA stole all the headlines yesterday, no doubt,
01:32with their new laptop to take on the likes of Intel and AMD.
01:35But, of course, we all know that they have the architecture
01:38to go into those big data centers.
01:40And they have the new infrastructure with Vera Rubin
01:43in full production right now.
01:45But you also use NVIDIA in a way to sort of virtually game out a scenario
01:52for a client, virtually, so they don't have to necessarily have a physical prototype.
01:57Tell us about that.
01:58Yeah.
01:58So if you look up until now, synthetic data generation hasn't been accurate enough
02:04for robotics.
02:06Many of these applications are 50% to 60% accurate.
02:09And when it comes to systems that need to operate safely at speed and at scale
02:14in an industrial-hardened environment, it hasn't been achievable.
02:19Over the last year and a bit, and we announced in March our collaboration with NVIDIA,
02:23we've really merged their capabilities in accelerated compute
02:28and their abilities with wide world simulation with our industrial knowledge,
02:34our digital twin capability, and, of course, our robotics
02:37to address the sim-to-real gap.
02:41Yeah.
02:41And in a number of applications, most notably in the electronic space,
02:46we managed to shrink that sim-to-real gap to a very, very close correlation.
02:52So we can, with a 99% accuracy now, recreate robot motion in the virtual space
03:00when compared to the real space.
03:02So what kind of cost savings are we talking about?
03:04And also, more importantly, perhaps, is time to market.
03:07Oh, absolutely.
03:08You can now reimagine how you bring product to market.
03:11So think if you're sitting in Taiwan right now, you're designing a product here,
03:15you're having to apply it in China, apply it in India, apply it in Vietnam.
03:20You need to be able to, in a concurrent way, generate those production systems
03:25but also train them for all these manufacturing edge cases
03:29so that you can then concurrently engineer,
03:32taking probably about 50% out of your time to market,
03:37but also at the same time take about 40% of your total cost out
03:41because from day one, you've pre-trained for all those manufacturing edge cases
03:45so you can hit productivity much more earlier than traditional systems.
03:50So through this, or just the big uptick, obviously, in physical AI,
03:55how are new installations for you guys looking this year?
03:57I mean, the electronic space specifically is going through a super cycle.
04:03I think there's been a lot of activity on the consumer electronics side
04:08but also more recently, the AI compute side has got to volume levels
04:15that have justified robotization.
04:17Many of these systems up until now were only partially automated
04:21because of the lower volumes compared to the consumer electronics
04:24but I think you've all seen the headlines on data centers
04:27and how that market is growing exponentially
04:29and with that comes the robotization of that space.
04:32What's your strategy for China?
04:34Because they have their own infrastructure
04:36and their own indigenous innovation in robotics.
04:40Yes, I've seen a lot of automobile factories with ABB robots, no doubt.
04:44They're going dark factories and the like.
04:45But again, I think the cost basis because of their stem-to-stern supply chain
04:50is about 40% of what...
04:52I mean, you're at the high end, really, ABB.
04:54But what is your strategy for China
04:55as they go up the value chain on their industrialization?
04:59Oh, absolutely.
05:00I think China is 50% of the world's robot market if you look at the IFR data.
05:05So they are obviously the dominant player in robotic automation from a market perspective.
05:11And we've been there for the last 20 years.
05:14We've built one of our main production facilities in China
05:18and we've got a long history of helping Chinese system integrators
05:23and the ecosystem there evolve and become more robots capable.
05:28Now, there's no doubt it's a competitive space.
05:31But at the same time, we are deeply present in China
05:35and we are also using, I think, the deeply integrated supply chain on the cost side,
05:40but also on the system deployment side,
05:44this ecosystem of fantastic value providers to come up with many of these solutions.
05:48So we're active there.
05:50A number of different parts makers in the whole IT supply chain
05:55are talking about key component crunch right now.
05:58Yeah.
05:58And prices, obviously, for memory is going way up.
06:01How is that impacting you?
06:03What are you seeing?
06:04No, I think in the global market of robotics, actually, we've seen for the last few years,
06:10we haven't seen exponential growth in the robotic space.
06:13So we haven't quite seen the same pressures on supply chain
06:17that you might have been seeing coming from, say, the compute side in electronics.
06:21Yet.
06:22Yet.
06:23Now, definitely, it's something to keep a close eye on.
06:26The good part is, as we went through the COVID time,
06:29we retooled many of our supply chains to become more resilient to possible supply chain shortages.
06:36So this resilience, certainly, we believe, is going to stand us in good stead
06:40should there become a tighter supply market.
06:42Could that be exacerbated by the exponential interest in humanoid robots?
06:46I know you do industrial robots primarily.
06:48You do have a division, I think, called You, Me, which is looking at sort of co-bots.
06:54Absolutely.
06:55Yeah.
06:55So we are active in the co-bot space.
06:58But also, we are very active in trying to make our robots more intelligent
07:03and essentially utilize many of the characteristics that you see out of a humanoid.
07:08We just haven't seen the volume application of the humanoid form factor in the industrial space.
07:14We've seen a slightly different form factor.
07:16Right.
07:17We see...
07:17It's not as necessary, is it?
07:19Unless you need dexterous hands for sorting and things like that in warehouses, right?
07:24Yeah, it's not.
07:26But you can still achieve that dexterity by putting your robot platform on wheels.
07:31And you can come up with much higher payloads.
07:34Now, some of these technologies from humanoids are great.
07:37But from a supply chain perspective, we're not seeing those volumes
07:40and those volume use cases yet impact our supply chain.
Comments