Safer, Smarter Construction Sites with Edge AI and Caterpillar Autonomous Machines - Ep. 285

Summary

NVIDIA AI Podcast Ep. 285 Summary

Episode Title: Safer, Smarter Construction Sites with Edge AI and Caterpillar Autonomous Machines
Host: Noah Kravitz (A)
Guest: Brandon Hootman (B), Vice President of Data and Artificial Intelligence at Caterpillar
Date: January 14, 2026

---

Overview

This episode dives into Caterpillar’s journey at the frontier of AI-driven transformation in heavy manufacturing and construction machinery, focusing on the impact of edge AI, advanced digital twins, and autonomous systems. Host Noah Kravitz interviews Brandon Hootman about Caterpillar’s collaboration with NVIDIA, the deployment of AI assistants directly in machines, and how these advancements are changing safety, efficiency, and the future of work on construction sites worldwide.

---

Key Discussion Points & Insights

1. Caterpillar’s Legacy and AI Transformation ([00:41]–[04:43])

• Caterpillar at 100 Years: In 2025, Caterpillar marked its centennial, highlighting its global presence not only in construction machinery but also in mining and energy/power generation industries.
• Brandon Hootman’s Role: Brandon has been with Caterpillar for 27 years, evolving from software engineer to VP of Data and AI, leading digital and AI-driven ecosystems to improve customer outcomes:

  “Our goal…is to just make our customers more efficient, more effective, more productive, more safe with our equipment than any other in the market.” (B, [02:40])

• AI Accelerator Function: Caterpillar now has an AI Accelerator program to speed up AI value and adoption across all business lines.

2. Digital Twins and Data Foundations ([03:01]–[13:54])

• Digital Twins:
  • Essential for blueprinting manufacturing facilities and supply chains.
  • Enable predictive/preventative maintenance and real-time supply chain optimization.
• AI’s Disruptive Power:
  • Allows transformative change without traditional, heavy ERP overhauls.
  • Unlocks efficiency by leveraging historical data across decades.
• Caterpillar + NVIDIA Collaboration:
  • NVIDIA provides a holistic AI ecosystem—not just hardware, but simulation, training, and deployment frameworks (Thor, Omniverse, Riva, Parakeet, QOpt).
  • This collaboration led to rapid rollout of AI in-machine assistants.

3. In-Cab AI Assistants & Operator Experience ([06:15]–[10:40])

• Addressing Labor Shortages:
  • AI assists operators, technicians, and customers, especially when skilled labor is scarce.
  • “By bringing our CAT AI assistant into the cab of our machine, it really allows that operator to get access to our entire digital ecosystem, the hundred years of collective knowledge…” (B, [06:51])
• Operator Experience:
  • Focus is on making AI unobtrusive—helping operators stay focused on the job, not the technology.
  • Interactions are voice-driven, using NVIDIA’s Riva and Parakeet for natural language support, with screen visualization for complex queries.
  • Demonstrated at CES, built in months, not years.

4. Manufacturing Overhaul—Clear to Build Process ([11:17]–[14:14])

• ‘Clear to Build’:
  • Ensures parts, components, and labor are available to fulfill production orders.
  • Historically struggled with disparate systems and asynchronous data.
  • With Omniverse and NVIDIA’s QOpt, Caterpillar now determines 30-day build readiness in 100 milliseconds.
  • “That just shows you the power of, you know, kind of once you have your arms around the data… and working within the Nvidia ecosystem and the AI factory, the power of what you can do.” (B, [13:54])

5. Edge AI for Real-Time, On-Site Autonomy ([15:00]–[18:32])

• Dynamic Job Sites:
  • Construction and mining sites are unpredictable, requiring localized, real-time AI.
• NVIDIA Thor at the Edge:
  • Runs sophisticated models onboard, enabling instant decision-making even when connectivity is spotty.
  • [Example:] “If I set that wall, I want to know that the agent that's responsible for setting that wall on how far my excavator arm can swing out actually sets it… now with the power of Thor, we can do all of that calculation onboard the machine…” (B, [17:37])
• Connectivity:
  • All Caterpillar machines now ship with some form of connectivity (cellular, satellite), but edge compute is essential for isolated locations.

6. Safety and Autonomy in Industrial AI ([19:47]–[22:24])

• From Deterministic to Probabilistic Systems:
  • Past autonomous systems followed pre-coded routes; AI enables robots to interpret and respond to complex, unpredictable scenarios.
• Simulation & Testing:
  • Heavy reliance on Omniverse and digital twins to recreate millions of hours of real-world and edge-case testing virtually before field deployment.
  • “For us, it's priority one. Before we put anything into market, the amount of rigor and the pace that we put it through to ensure safety is second to none.” (B, [20:52])

7. Data as Competitive Advantage ([22:24]–[26:33])

• Decades of Unique Physical-World Data:
  • Caterpillar’s wealth of job site and operational data informs and tailors AI models for domain-specific performance far beyond generic large models.
  • Reinforcement learning and post-training on Caterpillar’s data set proprietary benchmarks in construction and mining AI.
• MIT Study:
  • MIT recently published a case study on Caterpillar’s data infrastructure, signifying academic validation of their approach.

8. Working with AI—Team Structure & Mindset Shifts ([27:21]–[29:07])

• Rapid Prototyping Culture:
  • Developing AI experiences requires breaking from legacy, monolithic approaches.
  • Small, multidisciplinary teams can prototype and validate new AI features in weeks.

  “…challenge yourself to think different in the way that you've built software and be willing to step away from convention a little bit, try some things, and then kind of see how those prototypes work and then rapidly scale from there.” (B, [28:37])

9. The Future: Autonomous Physical-Cognitive Convergence ([29:07]–[33:19])

• Accelerating Change:
  • The intersection of AI agents/cognitive work and physical robotics is happening faster than predicted.
  • Example: Instead of retrofitting old machinery with modern sensors, send AI-powered robots with sensors to interact with and update digital twins in real time.
  • Data "flywheel" effect: New data continually reinforces models and digital environments.
• Transformative Tech on the Horizon:
  • Quadruped robots with thermal, acoustic, and visual sensors; AI copilots; ongoing enhancements to digital twins.

10. Skills and Mindsets for the Next-Gen Workforce ([33:19]–[37:21])

• Ubiquity of AI Assistance:
  • All roles (engineering or otherwise) will interface with AI tools.
  • Prompt engineering—knowing how to interact with AI—will be a baseline skill.

  “So much of people's ability to be effective in the future, I think, is going to depend on your ability to be able to leverage those tools and interact with those tools and understand to a certain extent how they think and work.” (B, [34:13])

• Practical Advice:
  • Experimentation and learning-by-doing are strongly encouraged—AI tools are “the most approachable” new tech seen in decades.

11. Resources for Further Exploration ([37:21]–[38:44])

• To Learn More:
  • Visit caterpillar.com CES landing page for updates and press releases.
  • Reference the MIT CISR Caterpillar Study for details on Caterpillar’s data-driven transformation.

---

Notable Quotes & Highlights

• On in-cab AI:
  “It's not AI instead of people, it's AI with people to help them be more efficient at the work that they're doing and be better at the work that we're doing. And we're really, we're really excited about that.” (B, [10:40])

• On manufacturing evolution:
  “We were able to take the Clear to build process and calculate that for a 30 day window in 100 milliseconds.” (B, [13:39])

• On AI-powered robots:
  “Now you can bring the sensors to the machine. By the way, while they’re walking your factory, they’re also going to give you an updated version of, you know, data to reinforce that digital twin…” (B, [31:38])

• On workforce preparation:
  “There are no excuses for people not finding a way to work with these tools and interact. A lot of them are free…very approachable in terms of understanding them.” (B, [34:13])

• On the future of the industry:
  “This knowledge aspect of what AI is doing...is going to get injected into the physical world in ways that are really going to surprise all of us.” (B, [32:12])

---

Memorable Moments & Timestamps

• [06:51] – "By bringing our CAT AI assistant into the cab of our machine...the ability to get real time interaction with that machine in a natural way..." (B)
• [13:39] – “Calculate that for a 30 day window in 100 milliseconds.” (B)
• [17:37] – “We can do all of that calculation onboard the machine and invoke that machine command to be able to set those parameters…” (B)
• [20:52] – “Safety. For us, it's priority one. Before we put anything into market, the amount of rigor and the pace that we put it through to ensure safety is second to none.” (B)
• [31:38] – “Now you can bring the sensors to the machine. …these were problems that felt more like kind of it and ot before, and now they're going to be physically solved in a way that...is super interesting.” (B)
• [34:13] – “So much of people's ability to be effective in the future...is going to depend on your ability to be able to leverage those tools and interact with those tools...” (B)

---

Conclusion

This episode demonstrates Caterpillar’s commitment to integrating AI deeply into every level of its business, not for the sake of hype but to enhance safety, efficiency, and the quality of human work. The combination of edge AI, massive proprietary data, rapid prototyping, and a people-centered approach stands as a model for digital transformation in heavy industry. As AI and robotics merge more rapidly than expected, both technology and the workforce are being reshaped, with Caterpillar (and its partners like NVIDIA) leading the revolution.

Transcript

[A]

Foreign. Welcome to the Nvidia AI Podcast. I'm Noah Kravitz. With me today is Brandon Hootman. Brandon is Vice President of Data and Artificial Intelligence at Caterpillar, and we're Speaking just after CES week 2026. Happy New Year, everybody. It was a big week for Caterpillar, so we're going to get into all the exciting stuff that's been going on this week, but leading up to it, certainly. But before we do that, Brandon, welcome to the pod. Thank you so much for taking the time to join us.

[B]

No, awesome to be with you. Big week for Caterpillar. We were excited to be on stage at CES and also excited to announce some things in partnership with Nvidia.

[A]

Fantastic. We'll get into all that in a second, but first, if you would, can you tell us a little bit about Caterpillar? I would think most people listening know the brand, have seen your machines, have seen the logo, but if you can give us a little bit of background about the company and then also maybe a little bit about your role, just kind of to set up the conversation.

[B]

Be happy to, Noah. So 2025 was a big year for Caterpillar. We celebrated our 100 year anniversary. So we've been helping build a better world in multiple areas for the past century and really proud of that. And, you know, I think a lot of the audience would recognize Caterpillar from just the drives down the interstate and you see our machines doing the work and, you know, that's one area of our business. But we've also got a pretty significant footprint in other areas that individuals may not know about. We do a lot of work in heavy mining, and then we also do a significant amount of work in energy and power generation too, which with AI now really taking off, as you can imagine, has been a huge part of our business. We're really excited by that.

[A]

Yeah, for sure. And so what about your role? How long have you been at Caterpillar? If you want, you know, a little bit of backstory of how you landed here, but then maybe you can just kind of describe what, what you do as VP of AI and Data.

[B]

Yeah, yeah, we'll do. We'll do. So I'm a software engineer by my education and by most of most of my time in the workforce. I've been with Caterpillar for 27 years and doesn't. Doesn't really feel like that, but I've been here now for 27 years. And over the course of my 27 years, while it feels like you Know, you could look at that and say, well, you've been, been in one spot for quite a while. Like I've probably had six or seven different careers over the course of that 27 years. So it's been pretty cool to do. Yeah, I've spent time in our, you know, supporting our dealer network and the systems that they use to run their business and then our manufacturing systems. And now of course, I'm in our digital organization where we're building an ecosystem of digital capabilities that complement the machines that we sell to market. And our goal with that is just make our customers more efficient, more effective, more productive, more safe with our equipment than any other in the market.

[A]

So how is Caterpillar thinking about AI when it comes to reshaping modern manufacturing? We'll start there.

[B]

Yeah, that's a great question. So in addition to my role that I play in CAT Digital, I also serve another role for the enterprise. It's a recently established function called AI accelerator. And it truly is about accelerating the adoption and the value that we get from AI across the enterprise. You can imagine with our business, NOAA manufacturing and supply chain is a huge component of that. With Caterpillar having been around for a hundred years, we've had, you know, probably for 60 of those hundred years we've had systems and data produced and you know, the concept of transforming your enterprise in those areas used to be that you go through a big systems deployment, ERP deployment. What we found with AI is it's, it's kind of breaking through the traditional barriers of accelerating, getting to, getting to those big transformations without the heavy lift that you'd have of kind of ripping out the core of your systems to be able to do that. We're looking at AI from, and I would say, Noah, we're starting with the foundation, the concept of a digital twin that can truly blueprint your manufacturing facilities. Each of the work centers that you have, machining centers, your supply chain that complements that. That first basis, then reinforced with real world data, opens up a world of opportunities. Predictive and preventative maintenance on key assets on your shop floor. Being able to simulate in real time supply chain constraints and then optimize your build schedule for what you're going to be building in the next day, next week, next month. All of those scenarios that felt like they were really hard to go achieve, we're finding that AI is allowing us to, to break through those walls pretty quickly.

[A]

That's fantastic. Yeah. We've spoken in the past months to folks using digital twins. As you said in Factory manufacturing situations, automotive comes to mind. You guys are in the business of creating heavy machinery and as you said, you know, have, have your reach into other industries and kind of areas as well. You're collaborating with Nvidia on AI transformation. Can you dig in a little bit more and kind of talk about what that looks like day to day and whether that's for CAT customers, dealers of your equipment, or maybe even, you know, your own employees?

[B]

Yeah, yeah, I'd be happy to. I guess the first thing I would highlight is that the reason that we're working closely with Nvidia is the, you know, the comprehensive approach to a AI ecosystem. Not just hardware, not just some libraries and frameworks, but it's a, it's a really well thought out ecosystem that allows you to simulate, train and deploy.

[A]

Right?

[B]

Those are the core fundamentals of getting AI right, irrespective of where you're at in an enterprise like Caterpillar. Whether we're doing AI to make our customers more efficient in the way they work with our machines, or whether we're making our supply chains more efficient, or whether we're looking at core processes like, you know, finance, transformation, et cetera. So those, those concepts in and of themselves are almost immutable, Noah. But very specifically, you know, when we, at the work that we're doing, and I'll just, I'll just zoom in a little bit on specifically our customers and some of what we announced this week as CEO.

[A]

Great.

[B]

There's a lot of, especially when you get into, you know, what we would traditionally call blue collar work there. There truly is a shortage of skilled and qualified labor in the marketplace. And it really doesn't matter whether for us, it might be within our factories, operators, people that actually work within our factory, you go to our dealers, it will be service technicians, and you go to our customers as people to operate the machines. So we're thinking about AI as a very pragmatic way to solve those problems in a very tangible, very outcome based way and starting with our customers. As simple as making sure that when an operator steps into the cab of our machine, they can understand the operations of the machine, they can get feedback on how they're operating the machine, they can get coaching. The advanced technology features that we have on our machines are very accessible in a natural language way. So by bringing our CAT AI assistant into the cab of our machine, it really allows that operator to get access to our entire digital ecosystem, the hundred years of collective knowledge that we've built, and then the, you know, the ability to get Real time interaction with that machine in a natural way where you're not having to open a, a cell phone or you know, look at a laptop screen to go do these things. It's just there, it gets out of the way and it kind of helps them, helps them do their work. And we're excited by that, a big enabler of that. And you know, part of the reason that we're working with Nvidia is the edge compute platforms that Nvidia provides. The Thor ecosystem is class leading both in terms of compute power, but also the software frameworks, the libraries and the ability to quickly deploy what we saw at CES that we were able to demonstrate with AI in the cab of an excavator. No, that wasn't a five year research project that we jointly did. That culmination of that work would be measured in months, not years. So it just shows our ability to take these technologies at scale and rapidly integrate them with some of the best machines in the world. And we're really excited by that.

[A]

To go back to what you were describing about, to me it sounded like this co pilot scenario or metaphor that's used a lot when we're talking about, you know, humans using AI tools in whatever line of work or whatever they're doing to kind of get that assist. Yeah. What's it like in the cab of a cat machine? What is the operator? And I'm sure they're all different, but when you're talking about, you know, the AI systems being there to help, what is the operator seeing?

[B]

Yeah, so. And we're still working on what that, what that final experience will be like. You know, we had the opportunity to demonstrate at Cesar shortly after, you know, coming up here quickly will be Con Expo, Construction Expo. It's a really big industry event for the construction industry that'll be in Las Vegas where our customers will get an opportunity to go more hands on with that and get to see that. So we're looking for feedback on what that final experience will be. But one of the things that we wanted to make sure that we did with at Noah is that this was never about the AI being front and center. It was about helping our customers get the work done.

[A]

Yeah.

[B]

And allowing them to get to all of the advanced capabilities that we've either built with on the machine or across our digital ecosystem in a way that felt natural for them. So we want them focused on digging, moving, hauling, not playing with another screen. And we do have a screen that, you know, in our concept machine that we've showed where Things come up. So as an example, if you know, you have questions about operating procedures, that screen will visualize those operating procedures and show them to you. But other than that, you could have a voice interaction with the assistant and get everything done that you need to and never have to worry about taking your eyes off of the, off of the job site. Again, I would. My compliments to Nvidia. We were very quickly able to do that with Nvidia's Reva voice services and Parakeet models, et cetera.

[A]

Well, glad to hear that. But it's always about the collaboration. Right? And as you said, it's, it's for all of the advances, all of the speeds and feeds and benchmarks and everything we talk about when it comes to AI and data and everything that goes into it, as you said, it's about getting the work done. And really, the more the system can stay out of the way, but give the operator what they need to focus. Yeah, that's what it's all about. That's great to hear.

[B]

You know, and that's such a, such a salient point. No, and as I would say, as we've really hit stride with AI, that's what we've found. Its ability to naturally help you just get things done in a way that isn't intrusive, that doesn't require you to have deep knowledge on systems and menus and where data is at. It just allows you to naturally get to that outcome. That's the game changer. And it's not AI instead of people, it's AI with people to help them be more efficient at the work that they're doing and be better at the work that we're doing. And we're really, we're really excited about that.

[A]

Absolutely. Kat's using Nvidia's AI factory and your manufacturing digital data platform to automate things like forecasting and scheduling. Can you speak maybe to a manufacturing process that's really been overhauled that just looks completely different now because of AI?

[B]

Yeah. And I had hinted at this a little bit earlier, Noah, but let me double click on it. And for your audience, I'll apologize. I'm going to go a little bit deep into how factories are ran here, so just bear with me for a minute.

[A]

Go deep. That's why we're here. Yeah.

[B]

But if you think about the complexities of running an operation like what we do at Caterpillar across a really diverse product line, it's a multi layered supply chain where you may see final assembly happening in a factory, but There are a lot of predecessors that are feeding into that. Either parts coming from suppliers or other facilities that we've got that are doing work on components. So as we are getting our orders coming in from our dealers and customers, you want to be able to ensure that you have everything available to be able to produce those orders. So that process, and I think this is probably an industry term, but we use it pretty heavily at Caterpillar. It's called Clear to build. It's very simple. In a 30 day window, in a 14 day window, in a 5 day window, am I clear to build? Do I have all of the parts and components, the supply that I need, either in the network or available at the factory? Do I have the labor that I need to be able to do it? Now think about Caterpillar at the scale that we are and all of the data and the disparate systems and the signals from your supply chain that have to go into calculating that. It's really hard.

[A]

Yeah.

[B]

And without, you know, everybody being on a singular system or a singular platform, you're dealing with, you know, kind of an asynchronous environment of multiple signals coming in and being able to parse and process those at scale is really, really difficult and has required a lot of really smart people to do in the past. And even with that, it's been difficult to calculate on a, on a long term window. We have found that with, you know, first with Omniverse, building a good digital twin that shows an accurate representation of your factory, the capacity of the factory and the supply chain, and then taking some of the Nvidia models and the inference services, QOPT as an example, we were able to take the Clear to build process and calculate that for a 30 day window in 100 milliseconds.

[A]

Wow.

[B]

Yeah.

[A]

That's remarkable.

[B]

Yeah. Now granted these are, these are, you know, we're doing it at a facility level right now and we're working through the rest of the enterprise to be able to do that. But that just shows you the power of, you know, kind of once you have your arms around the data.

[A]

Yeah.

[B]

Once you understand the platforms and working within the Nvidia, the Nvidia ecosystem and the AI factory, the power of what you can do. And you know, these, these aren't models that we had to go develop on our own. You know, compliments to Nvidia, things like Co Opt or. They're really good and they work really well. So with a little bit of tuning on our side, we were able to able to make those things Work at a pretty impressive scale.

[A]

I'm speaking with Brandon Hootman. Brandon is Vice President of Data and Artificial Intelligence at Caterpillar Cat, as the logo says. And people know you, as you said for, you know, you're driving down the highway and you see the machines, but, but so much more than that as we've been discussing. Speaking of the machines, when operators are out, they're working on sites and they're using the machinery and the CAD systems that now have or are starting to have these AI guidance, you know, sort of co pilots as you were describing. How does AI at the edge come into play? How do edge AI systems enable real time decision making and kind of just adapting to what must be incredibly unpredictable environments when you're out on a construction site or in a mine?

[B]

Yeah, yeah, yeah, you nailed it. No, and if you think about the dynamic nature of a construction site, day to day, they're never the same. Things are moving around, different work is being done, things are being completed, materials being brought to the job site. The very nature of it isn't. It's not like our factories where things are relatively predictable and you've got a line that things are moving down and it looks the same day in and day out. So if you think about what it takes to do some level of automation and bringing AI to a job site, even as performant as networks and cloud systems are today, there is a level of processing that has to occur at the edge. And that level of processing doesn't just require CPU or GPU compute, it requires a level of storage, it requires a level of memory to be able to process these things in as close to real time as you can get. Situations with safety can sometimes be just a matter of seconds between something being avoided or something really serious happening. So for us, as we're thinking about what we bring to the machines and what we bring to the job site, this partnership with Nvidia and the architecture and capabilities that Thor brings is really what we see is the game changer that's going to allow us to bring the capabilities that we've had in pockets for a while to now start to bring those to customers and machines en masse. And the way that we're doing this now is Thor allows us to bring a lot of the sophisticated AI and ML models that we had built that run across our data platform that powers digital today. We can now bring those to the machine. And in the event that we have network connectivity, in the event that the scenario supports it, we might invoke a cloud agent but if there's too much network latency, we've now got the ability to revert back to something that can run on the machine at a fidelity and a level of insight that still allows the, still allows the customer, the operator to be effective in the tasks that they're doing.

[A]

Could I ask you, Brandon, to dig in a little to what some of those capabilities are kind of at the machine level in, you know, whatever scenario, construction, mining, whatever it is that, that makes the most sense.

[B]

Yeah, you bet. So just think about a perception based warning of needing to understand maybe an unsafe scenario on a machine or thinking about a machine operating. Let's just say we've got a machine operating on the side of a road.

[A]

Okay.

[B]

Think about the dynamic nature of traffic going by, power lines being above. And let's say this machine's an excavator. So if you need to set a wall on how far that excavator arm might swing out, you want to know that if I set that wall, I want to know that the agent that's responsible for setting that wall on how far my excavator arm can swing out actually sets it. Because the last thing we want to happen is an operator be in the machine and all of a sudden that arm swings out in oncoming traffic for an excavator. So that example, NOAA would be one where, you know, now with the power of Thor and the ability to run these AI agents locally, we can do all of that calculation onboard the machine and invoke that machine command to be able to set those parameters on the machine to perform those functions.

[A]

Kind of a naive question on my part, but what's network connectivity generally like in. And again, I know there's a huge diversity of scenarios that equipment is working in, but is it generally pretty unstable? Is it like akin to using cell phones to modern construction sites usually have network set up? What does that look like these days?

[B]

Yeah, I would say it's, you know, it really depends on the area where our machines are operating. One of the, one of the decisions that Caterpillar made over a decade ago was that every machine that we produce will come with connectivity.

[A]

Okay.

[B]

So that, that is there and that connectivity can be cellular connectivity. It can be, it can be satellite connectivity. We're also looking at the next generation of connectivity of high speed satellite connectivity. How 5G is going to play into it. So that infrastructure is there, but then also too our machines end up operating in areas where the cell network just isn't great and there's not great Coverage. And what we want to do ensure is that when we bring these AI capabilities to the machine, that when the customers need to count on them, they can count on them. And by now having the storage capacity, the compute, the memory bandwidth on machine, we're now able to take a lot of what we've had to do in the back office and now bring that local to the machine, which is super exciting.

[A]

Yeah, no, that's fantastic. Yeah. Well, how do you think about safety when you're building autonomous equipment? When you're building, you know, heavy construction machinery and mining machinery in the age of AI? Like, how do AI and safety go together? How do you guys think about it? Kind of, you know, starting at the beginning of the design process and working through.

[B]

Yeah, and this is, this corner worked, running really quickly. If you think about autonomy systems of the past, they were very, very interesting in the way that they worked, but they were also much more deterministic than they were probabilistic.

[A]

Right.

[B]

So, you know, what that means is that, you know, if you're going to put autonomous machines on a job site, you would almost pre program routes into them. You would then have a complimentary suite of sensors on them that were able to detect an object moving in front of them and then appropriately, you know, take action to avoid a safety incident. Now with AI, you're kind of going away from pre programming things in a deterministic autonomy and now you're going to giving a robot a task.

[A]

Right.

[B]

And it now is using various subsystems and agents and, you know, visual processing models to then determine the action that it needs to take safety. For us, it's priority one. Before we put anything into market, the amount of rigor and the pace that we put it through to ensure safety is second to none. Doing this in a physical world is very time consuming, as you can imagine.

[A]

Yeah.

[B]

So for us, the pivot to AI based autonomy and machines is predicated by a robust simulation environment. The ability to take real world data. A digital twin of a job site, a digital twin of the machine. The models that we would have driving that AI subsystem where the, you know, the machine's making decisions and ensure that we run that through millions of hours of simulation and test scenarios and honestly no worst case scenarios in our simulation environment, then obviously that's followed by a, you know, a series of physical tests too. But now that that simulation environment, now when you're getting into, you know, Omniverse and the combination of that, you know, Isaac and Cosmos, and, you know, we now have the Capability of being, being able to, to do that at scale. And I think that's what's going to make the, you know, make it feasible for companies like Caterpillar now to make this, this pivot pretty quickly from the way we've been doing autonomy for the last 20 years to what the future of it's going to look like.

[A]

Right. You must have an incredible amount of data on just the different types of job sites and the different, you know, natural materials and building materials and all of, just all of the different physical world, you know, components that go into the work that these machines are used for.

[B]

We do. We do. And that, that, that data, when you think about what it takes to make AI really work, is the foundation models that are out there today are, they're good foundation models, but those foundation models haven't been built with the specificity of the way a construction site works.

[A]

Sure.

[B]

Or the way a mining site works or the way a query operations work. So we have the opportunity to take those foundation models now and do some level of reinforcement learning and post training on those models with the specificity of, of kind of the way our customers work and operate. We do think that's an area of differentiation for Caterpillar is that we want our customers to know that when they get an AI assisted or an AI autonomous machine from Caterpillar, it's backed by the quality of Caterpillar, but it's also backed by the hundred years of experience and the decades of data that we've got to be able to do this the right way.

[A]

No, absolutely. So you mentioned Obviously Caterpillar turned 100 last year. Congratulations to the company, to everyone.

[B]

Thank you.

[A]

And you've been around at the company a little over a quarter of a.

[B]

Century, but when you say it that way, it sounds like a really long time.

[A]

I mean it in the most celebratory manner possible. Believe me, I'm too old to make age jokes myself. How long? I think you said this at the beginning, so forgive me, but how long have you been working specifically on sort of data and AI related stuff?

[B]

It's been an interesting journey for me. As I said, I started as a software engineer and I was super privileged that my first assignment was working with our dealers. Caterpillar's got an independent network, a network of dealers of roughly 155 globally. So got to work on the systems that they use to run their business and got the opportunity to just see kind of what that process looked like in terms of the edge of where they interact with the Customers and that so super, super valuable. And then, you know, as I, as I kind of transitioned through my career, spent some time overseas in Grenoble, France at a manufacturing facility there and then shortly after that in Gasolis, Belgium, which is really close to, to Brussels and another manufacturing facility there. And I think one of the things that that's given me the privilege of seeing NOAA is the kind of just the federated footprint of how things move and interact and what the data looks like across our, our value chain. So when we started our work in Cat Digital of thinking about how do we utilize all of the data that we've got, our machines operating in the field to build a digital ecosystem where we can, you know, do fleet management, do predictive and preventative maintenance. I don't know whether it was by accident or good planning on somebody's part, but I kind of feel like, you know, those things prepared me pretty well to be able to step into what we're doing in digital. We've been, we've been, I'd say really, while we've been doing digital, I think we really stepped our game up in the last six or seven years in terms of how we thought about it. The platform that I cited for you, Noah, our team started building that about six years ago to be able to build it at that scale. I'm super proud of the work that our teams did. Oh for sure, roughly about two years for them to build that data infrastructure. It's cloud, native event driven polyglot, As I said, 16 terabytes of data and counting that we've got in that. And was also pretty proud too. If you don't mind, I'd share that. MIT recently completed a, a pretty interesting study on Caterpillar that's available if you just search MIT center for Information Systems Research Caterpillar, your audience can go out and it's about a 25 page article that they wrote on how we approach data. So I do also think, Noah, that that data foundation, you know, we're moving pretty quickly with AI. I really do believe that Data foundation has been a game changer for us. It is the foundation where we know that we've got data that we can trust. AI working with.

[A]

Yeah, no, 100% and as you said, that's it. You have to trust the data you're working with to be able to have trust in the rest of the rest of the system, rest of the process, let alone in your case, having people out there in the physical world relying on it through the process, through building the digital ecosystem and working with the AI models and working with the physical AI and getting out into the real world. Is there anything that jumps to mind that, you know, either maybe surprised you along the way, a learning or something that, you know, a way things went that worked, that you weren't expecting, or maybe a challenge, you know, a challenge that you wanted your teams had to overcome? That, as you look back on, was kind of like a, you know, a big watershed moment.

[B]

Yeah. Yeah, that's a great question. I've got. I've got several. But one of the things we learned, and it was almost kind of by accident, a little bit. You know, the traditional way that software is built is. I don't want to call it monolithic because there are agile approaches to building it. But if you. If you think about that process of going from I've got an idea to I want to build. Build a software product, it can be a little monolithic and it can. It can drag out a little bit. We found that when you're building AI capabilities and AI native experiences, it does really require you to think differently, that you don't need to have necessarily a completely formed thought. What you need is a notion on what you want to go do. And then you put a small team together of a couple of data engineers, a prompt engineer, a context engineer, and a data scientist, and someone that might be able to throw a quick user interface on top of it. And then you spend three weeks building a prototype, and then you validate that prototype against the hypothesis, and then you go scale rapidly. You know, when we. The cat AI assistant that we showed at ces, that notion of developing that. No. Is something we only started about 12, 14 months ago.

[A]

Yeah.

[B]

And the pace of developing that, you know, that's backed by a library of 20 to 30 agents with a knowledge graph and, you know, context that passes across all of those. And, you know, we were able to bring it into the cab of the machine. And my counsel would be, as you go into AI, challenge yourself to think different in the way that you've built software and be willing to step away from convention a little bit, try some things, and then kind of see how those prototypes work and then rapidly scale from there.

[A]

So, Brennan, that's kind of a great segue to my next question, which is kind of looking ahead a little bit, talking about the future of AI and heavy industry. What do you see? If I could ask you to forecast a little. Right. What do you see over the next five to 10 years as being the most transformative technologies that are going to affect manufacturing and heavy equipment.

[B]

That's a really great question, Noah. And this has happened much quicker than I expected. The first iterations of, you know, if you really think about late 2021, 2022, when OpenAI really made a splash with, you know, with generative AI and large language models, up to this point, it's largely been cognitive, it's been knowledge workers, it's been, you know, getting people information or, you know, now with, you know, the concept of MCP and agents, you know, now actually performing tasks and doing work. The intersection of the cognitive world and the physical world are converging quicker, I think, than anybody, anybody expected. And they're, they're opening doors that have been closed for a really long time. Let me just, let me give you a real example of that in our, in our factories, you know, we've got some factories that have been stalwarts and Caterpillar for decades and they're really important to Caterpillar. And if you think about the, you know, the capital of building a factory that builds our machines that, you know, you don't stand one of those up just kind of on a whim you're really serious about. And once you build that factory, you know, it's usually there, usually there for a while. That can mean that you have critical assets in there that might be a little bit aged sometime. They may not have sensors on them, they may, you may not have the ability to get data off of them. In the past, the answer would have been, well, find a time, shut that factory down for a couple of months, go put sensors on all of these machining centers or work centers to be able to get the data off of them, you know, to do predictive and preventative maintenance.

[A]

Yeah.

[B]

Now what we're seeing there are quadruped robots that are AI based that have thermal sensors on them, acoustic sensors on them, the ability to visually interpret gauges on a machining center where these robots now, instead of having to, you know, put the sensors on the machine, now you can bring the sensors to the machine.

[A]

To the machine.

[B]

By the way, while they're walking your factory, they're also going to give you an updated version of, you know, data to reinforce that digital twin that you've got of the factory. You know, these were problems that felt more like kind of it and ot before, and now they're, they're going to be physically solved in a way that, you know, is super interesting, but way different than anyone anyone ever thought. And I think that's just that took the iceberg no. More and more we're going to see like this knowledge aspect of what AI is doing. It's going to get injected into the physical world in ways that are really going to surprise all of us.

[A]

Yep. No, it's great to hear you talk about that, you know, because I have the context of, you know, humanoid robots and we've had conversations with folks, you know, building humanoid robots to be in people's houses, houses and obviously delivery robots and self driving cars and agricultural equipment. But to think about the concept of, as you said, bringing the sensors to the equipment and then along the way gathering all that extra data and validating your digital twin and all of that, it's just the data flywheel just goes and goes.

[B]

It does, and it's mind boggling when you think about it. Just to go execute a, a scan of one of our, you know, millions of square feet on a manufacturing facility. I mean, you're talking about months to go do that now. It's, hey, I can, I can have a robot walk in a line and inspecting machines and getting an updated version of that digital twin of the factory almost in, in kind of real time.

[A]

Yeah.

[B]

You know, and then re reinforcing your models and your simulation environments getting better. So it's just these things are, they're, they're busting wide open in a hurry. Yeah.

[A]

You spoke just a minute ago about kind of the shift in mindset to work with AI and kind of moving from, I mean all the things you talked about really kind of came together. Thinking about deterministic versus probabilistic and being a software engineer and working for so long and now kind of having in relatively short order this kind of shift in the way that you, you know, everything from how big of a team that you assemble for a new project to how do you, how do you start and you don't have to have it all mapped out and having a notion. Can I ask you for listeners, what other skills and mindsets do you think the next generation of workers are going to need to have to succeed in these AI accelerated industrial environments?

[B]

Yeah, I think that's a amazingly complex question to answer.

[A]

Right. It's the trillion dollar question, right?

[B]

Yeah, it is, it is.

[A]

Solve the future economy for us. Brandon. Gotcha.

[B]

I think it's going to, I think it's going to depend on really where you set and what function you set in. You know, for my team as an example, we're very heavy on engineering. Software engineering, data engineering, AI engineering context engineering, prompt engineering. And that work has rapidly become AI assisted. Not AI completed, but AI assisted. And our best engineers know how to use those tools in a very effective way. You know, I think, Noah, we're moving to a world where I would struggle to cite a job where you wouldn't need to have some level of prompt engineering. If you think about what prompt engineering is, it's how do I talk to AI to be able to get it to give me the right answer. And so much of people's ability to be effective in the future, I think, is going to depend on your ability to be able to leverage those tools and interact with those tools and understand to a certain extent how they think and work. And for me, that starts with the very basics of just interacting with them and working with them. And that's always the advice. Irrespective of the function that I talk to. This technology is the most approachable technology that I have seen in my couple of decades of being. There are no excuses for people not finding a way to work with these tools and interact. A lot of them are free, they're available, they're there for you to start working with and learning upon, and very approachable in terms of understanding them. As I said, I just don't know that there's any one area that I could point to that I think is going to be immune from it.

[A]

Noah Yep. No, it sounds about right to me.

[B]

I would qualify. The way you interact with it might be different though. And I mean, let's think about this, right? My software engineers as an example, you know, they're probably going to have a copilot or a coding assistant that is working shoulder to shoulder with them as they're doing their work, right? I mean, they might want to simulate a code change and it's looking upstream to say, hey, if you make this change, these are the integration impacts offering some recommendations on some, some changes to your, your, to the code you're getting ready to, to submit. On the other side of it, if we look at the shop floor, we're probably not going to have a laptop sitting in front of somebody that's doing torque operations on a, on a shop floor, right? But what we might have is we might have a, you know, a AI enabled torque machine that tells you if your, your torque into the right level or what we might have is if you're doing complex operations, you might have a visual assistant that's helping look at the work that's being done and providing real time work instructions to you as you're doing it or offering you advice as you're doing it. So I think that's another thing that people are going to have to get comfortable with is irrespective of your job, it's probably going to be somewhere next to you, helping you understand the tasks that you're doing. And, you know, ideally taking some of that menial work or the mundane and the things that you have to go look up or find and just make you better at doing the work that you're doing.

[A]

Yeah. Brandon, for listeners who want to learn more, there's so much great stuff that you touched upon during the conversation, all of the technical and data and AI related stuff. But then how it translates into the real world and the physical world in particular is just so much to dig into there. And you mentioned the MIT study as well. But for folks who want to learn more, where's the best place for people to start online?

[B]

Well, thanks for that, Noah. Yeah, the best place for listeners to go is caterpillar.com CES We've got a great landing page that really encapsulates everything that we announced at CES with, you know, also some more details and some direct linkage to some of the specific press releases and things that we had made at ces.

[A]

Terrific. Well, again, thank you so much for taking the time to come on the show. It's been a great, really fascinating conversation and so much to think about. The next time I drive by a construction site. Right. You know, it's going to mean so much more to me and when I see the cat logo.

[B]

So, yeah, when you, when you drive by and see that machine, Noah just visualized the data flowing off of that, the AI running machine helping the operator. So thanks again, Noah, for the, for the opportunity to, to be with you here. It was, it was a privilege to talk a little bit more about the work that Caterpillar is doing and I'm really excited about what the future holds.

[A]

Absolutely. Look forward to tracking your progress and best of luck with everything.

[B]

Thank you.