ARM CEO Rene Haas on launching the company's first CPU product, AI's compute demand, and the SoftBank relationship

Speaker 8: Oh, my pleasure. Yeah. Thank you. I'm doing well.

Speaker 6: How are

Speaker 8: how are you both doing?

Speaker 2: We're doing Doing great. Fantastically. It's an incredibly exciting

Speaker 1: to your to your team. This this this audio video setup is perfect. We, you know, we talk to a lot of people. People like to go hard visually. Something, you know, that looks great but then they'll miss the audio. Did both quite well. Yes. So I appreciate it.

Speaker 2: Yeah. Fantastic team. We we

Speaker 8: have a great team here. We we hijacked a conference room and turned it into a a little mini studio here. Glad glad to see it's paid off.

Speaker 2: I love it. Perfect. Yeah. I mean, we were just reflecting on the fact that it it I mean, it's an exciting time for us. We cover the news. We cover AI. There's so much going on. You've been with Arm for thirteen years. Is that correct?

Speaker 8: That's right.

Speaker 2: Yep. Is this the most exciting year? Is this the most exciting moment of your career? Is this how else are you contextualizing, like, what it means to be in this particular moment?

Speaker 8: Gosh. Is it I would say, you know, having been in the tech industry my my entire career, it's probably the most exciting time. Yeah. I I would agree. I mean, artificial intelligence artificial intelligence and and the idea of machines that can think like humans, I mean, that's that's Star Trek type stuff. So if you grew up watching science fiction, and I remember I think one the very first movies my dad took me to was 2,001, A Space Odyssey. Yeah. We're kinda living we're kinda living that life now, and it's one of those things I know my compatriots have said the same that you knew it was coming, but you didn't know if you'd had a chance to work on it in your lifetime. So, yeah, so it's it's amazing.

Speaker 2: Yeah. Yeah. Now it's here. Talk to me about how you process the AI boom. There's been a couple of distillations of progress. You have the ChatGPT moment, the reasoning moment, the coding agent moment. As each of those moments happened, how did your business change? How did your perception of the AI boom change? And then at what point did you realize that you needed to expand the strategy, change the strategy? How deliberate were you along the last three years, four years as we've seen these sort of tectonic shifts in the capabilities of AI systems?

Speaker 8: Yeah. One the things, you know, in our industry that has been almost like a force that you cannot stop is the need for more storage or more memory or more compute. Right? There's just something, whether it's an end application, a new product that just drives that need. AI is almost the monster that can eat almost unlimited amount of memory, unlimited amount of storage, and unlimited amount of compute just simply because of the the raw amount of computation that's required, the amount of data that's needed, the amount of intelligence that you're generating in the tokens, etcetera, etcetera. So what we've seen over the last just number of years is as everyone in our industry has is this sudden huge tailwind of how much is enough and we don't seem to be hitting that point yet. And I was at I was in Computex a couple of weeks ago and met with folks in the industry. And I remember talking to some of the guys at TSMC, and they said we've we've never seen a cycle where it's this good for this many years. Something's gotta change. But we've never quite seen artificial intelligence at the level we're seeing it, and it still feels very, very early. I know a lot of people say that in the sense of, to your point, the ChatGeePN moment, the Claude Cudd moment, the co work moment. But still, there's lots and lots of industries that haven't even scratched the surface. So it's still amazingly, even given all the tailwinds, still seems very, very early.

Speaker 2: I can I can intuit or at least guess at what the the crunch at TSMC feels like? It's incredible demand from every supplier, sharp elbows from probably everyone who's trying to get line time and a lot of people working late nights, making sure that the machines are running, that the fabs are delivering at high quality consistently. But what is the shape of a crunch at Arm during a boom like this? Is it different? And then I imagine maybe take me through what the next couple of years will look like and how different pieces of the business will scale to react to different bottlenecks, different opportunities, different supply, just demand increasing broadly?

Speaker 8: Yes. So it's a very, very important question because it's changed a bit for us. When we were seeing all of this growth in demand in the prior years or in previous growth periods, we would feel the crunch in terms of our end customers saying, We're limited by this, we're limited by that. Demand is really strong. Our business model on the IP side is all about a licensing fee upfront, and then we collect the royalty on chips shipped.

Speaker 6: Yeah.

Speaker 8: That all kind of changed for us in at the end of March when we announced our first product. So now we are in that soup as well. We have huge, huge demand for the product, which is ARM AGI CPU driven by somewhat selfishly, it's a great product. It's two x to performance at the same power as the competition. But also with all this takeoff of Vagintic demand and all these agents that essentially spawn off workloads, CPUs do a lot of that work. There's been a huge demand for CPUs. So we were fortunate in terms of timing of announcing that product. But yes, just like everyone else now in that world, we're looking for all kinds of avenues to find increased supply because demand is extremely strong. So, yeah, to your question, we feel it on both sides. We feel it from our end customers on the IP side who are trying to ship as much as possible and then ourselves because we've got our own end customers that are really, really pushing us for supply.

Speaker 2: So when you're getting into your new own chip business, what changes for you? Are you a fan of founder mode? Is this something where you're clearing your schedule so that you can spend more time on that new growing division, that new opportunity? Are you bringing in a new team? Are you shuffling folks around in the organization who you know can deliver on this new initiative? Like, what is your managerial philosophy for delivering on that?

Speaker 8: Yeah. I've had a fortunate opportunity in my career to work for a lot of founders directly. Probably, if I just think back now over the last twenty five years, I've kind of I've not been a founder myself, but I've worked directly for founders, whether it was early startup companies, then a bunch of time at NVIDIA working for Jensen, and then at SoftBank who owns Arm, know, Moss is a founder.

Speaker 2: Yeah.

Speaker 8: And what I found for myself personally is that is probably the style that resonates best with me personally. Mhmm. So what does that mean in terms of data activity? Am I in the meetings with the operations guys figuring out where the supply is coming from? Absolutely. Am I thinking really, really hard about what we wanna do next in terms of product and can we get the supply and can we be strategic to a number of different customers? Absolutely. Did we bring in new people with new skill sets and new capabilities to help us there? Yes, we had to. Building chips is not the same as building IP. You need people who can do back end design, customer validation, supporting customers. You need operations folks that can work downstream. So we knew we needed to do all that. I had done a lot of that in my past and previous jobs and careers. So I had a good sense of what was required and as importantly, had a really good team around me, but also folks who came from that who came from that world. So yes, high degree of focus and I don't know if I would say I've gone into full founder mode, but that's kind of how I operate. I think intrinsically and as as you get on in your career, you you start to find out what are the environments that that not only resonate with you, but you can thrive in. And then for me, that's definitely one.

Speaker 2: We we we talked about, like, this interesting this interesting phenomenon in AI where with with twenty twenty hindsight, you can look back and see that the ChatGPT moment was a turning point. Claude Coat, as you mentioned, reasoning models. There's these unlocks that feel binary in hindsight and they sort of blur together as they're happening. But a couple months later, you can see that there's some sort of takeoff. Are you do you have a vision of what we're waiting for for the next one? Or are you more zoomed out looking at broader trends, of straight lines on log graphs, and you see that capability will come as compute scales?

Speaker 8: I think it's a little bit the latter. I mean, people like to look at the chat GPT moment or the Claude code moment, and maybe that's mile marker six in a marathon or mile marker eight. But actually, what they are is proof points Mhmm. Of scaling.

Speaker 2: Mhmm.

Speaker 8: That the more compute you can apply to the problem, the more sophisticated the level of problem that you can solve.

Speaker 2: Mhmm.

Speaker 8: And that's what I think we are seeing on on all levels. So I don't think there's another moment. You know, people talk about AGI or ASI or things things of that nature. I think what you're starting to see is for some of these models, for example, or some of these applications, we don't have the datasets and the Internet to solve them. So these are like science problems. Right? Engineering problems. Yet using the compute and models to create synthetic data

Speaker 1: Yeah.

Speaker 8: To help you solve these complex engineering problems that you can't find on the Internet, that could be the next thing, you know, if you will. So I I don't know that there's another end quote moment ahead of us, but I think what we have seen so far is that the law of scaling has not been defied. And the more compute you can apply to these problems and the more power that's required for that and everything around it, we get greater and greater benefits.

Speaker 2: Can you I I'm I'm interested to hear how AI is changing chip design specifically, like the actual work that your team does on a day to day basis. But I'd also like a little bit of a primer on just computer aided design through your tenure because it's not like it was pencil and paper three years ago and now it's a single prompt make a great CPU. So I imagine we're on a it feels probably a little bit smoother to you, but there's a lot of excitement about AI actually designing chips. Where are we on that trend? Where were we ten years ago?

Speaker 8: Yeah. I I remember as I was smiling when you said that. I I think I remember my first job out of TI. We actually had, I think, four foot by four foot tables where we would actually lay out printed copies of the the actual chip itself Yeah. To look at the layouts from a transistor level, trying to figure out exactly what we had done from a design standpoint. And that was before the workstations really had a lot of sophistication to do automated design. People have talked about CAD and CAM of that nature. Now now when you get to the the the billions of transistors are on a chip, there's there's no way you're printing that stuff that stuff out, etcetera, etcetera. You know, chip design is pretty interesting. You know, one level, it's code. In other words, when write a piece of software to describe what a circuit is doing, that's a piece of software we use or the industry uses something called Verilog Yeah. To describe what RTL is, which is a different version of a Python or C if you will. But chip design is a multidimensional problem. Right? You can write RTL and Verilog and it can be functionally correct. I can do an adder or a multiplier or a certain level of circuit, but to get that to operate at a certain frequency, consuming a certain level of power, and then occupying the smallest set of millimeters on a piece of silicon, that's still a very hard problem.

Speaker 1: Mhmm.

Speaker 8: So the industry isn't there yet in terms of being able to use those tools. Essentially, the the nirvana would be, I come up with an idea, I press a button, and I'm I can go tape it out at TSMC. Right? That is sort of sort of the nirvana. And will we get there at some point in time? Most likely, we will. We're a ways away because the datasets aren't quite correct and we don't have the good closed loop systems, but we'll get there. But where it's helping in chip design are things like so let me back up and just say, in chip design, a lot of the work actually is not in designing the circuits, it's actually in verifying the circuits, it's in bug checks, It's in validate validation. And AI helps a lot there. AI can do a great job in terms of triaging bugs. Engineer will do a run over a weekend and come back in them on a Monday and see where the bugs are and triage them and fix them. In the AI world, that can all happen automatically. AI can not only order the bugs in terms of the high priority bugs, but in some cases it can fix them. So we are seeing that today. We are absolutely seeing the benefit of AI helping us design chips faster, to get bugs out quicker, to help us in the validation process, which is super important because these chips are getting bigger and bigger. And the bigger they get, the more complicated they are to build, the longer it takes. So it's a little bit like running uphill against a treadmill. We're running fast, but the chips are getting hard. So it's not like we've shaved chip development times in half yet. But without the tools, I think they would be it'd be longer to develop, so we're certainly seeing benefit.

Speaker 2: A bit of a more anodyne question, but where else is AI permeating within ARM Holdings? I'm interested to know, like, you have license fees. Is there are there AI agents making sure that invoices get paid appropriately, like customer support, even just, you know, you have you probably have a ton of documentation. Have you wrapped that in an LLM? Like, there's some there's a bunch of, like, you know, very basic use cases for AI that you see propagate through all sorts of businesses. How have you thought about adopting AI outside of the core context?

Speaker 8: All over the place. I would say probably 85 to 90% of the company is using it in some way, shape, or form. We use it in legal, for example, to look at contracts

Speaker 6: Mhmm.

Speaker 8: To do inputs and checks against our standard terms and conditions. It's very, very fast. Finance teams uses it for all different kind of functions. One of the more interesting ones that we use it for is so on our IP business, we collect a royalty for every chip shipped. Yeah. We have a quirk in our financial model in that we have to every quarter accrue for the royalties that we think will come in the next quarter. So it's a forecasting tool. Mhmm. And that forecasting tool is almost like a a GDP hedge fund where we're trying to look at macroeconomic indicators, what's going on in terms of certain markets. We use AI to help us rationalize what our predictions are, and it's very, very good.

Speaker 2: Yep. Yep.

Speaker 8: When we when we report the royalties at the end of a quarter and then the following quarter, we do something which called a true up. In other words, where we were wrong and and where and where the error bars were. And I tell you AI makes the forecast really really well. And then just something anecdotally, you might be sitting in a in a in a in a meeting where you're brainstorming about different ideas and you're trying to figure out what's the size of this market or who shipped this much volume into this area or is someone using this type of substrate. You know, in the past, that would be we'll go look this up offline and get some answers. Mhmm. People are using ChatGPT or Claude or any of those tools real time in a meeting, and you just get answers much more quickly. Mhmm. It's, by the way, it's not necessarily that it's a 100% right answer each time, but it's it's 80 to 90% close enough that you can start making decisions. So we use it everywhere, and I use it all the time.

Speaker 2: Of course. How are you thinking about more farther out opportunities or technologies? I'm interested to know if you have a view on these wafer scale ASICs or even quantum computing. Is there something where you're starting an investigatory R and D process to see if there's at least an integration in the future, if not an indigenous, like, r and d effort.

Speaker 8: Yeah. No. That that you we were talking earlier about, you know, is this the most exciting time in the industry? And and there was a time, you know, twenty years ago when semiconductors were not seen as the next growth engine and and a lot of the investment and growth went into into other other areas. Yeah. Not only is there amazing opportunity for what you just described, whether it's wafer scale compute or in memory compute or different computer architectures. But everything around material sciences, different components, whether it's co package optics, whether it's different type of copper, the the compute the system is the computer.

Speaker 4: Mhmm.

Speaker 8: And you have to think about everything that goes into it relative to how you maximize performance. So, even things, you know, believe it or not, like efficiency of turbines and efficiency of gas engines and everything goes into building out power plants because everything in the supply chain, you know, suddenly matters. And the resiliency of the power matters. Everything can come off the grid if you're using alternate power sources. What does that mean in terms of interruption of supply? What does that mean in terms of reliability and sustainability of the systems? So it's probably a long way of saying that there's investment opportunities not just in the chip space, but in everything that surrounds it because now the entire system becomes an opportunity for innovation. And at ARM, yes, we have an investment profile. We also work very closely with our parent company Softbank in terms of looking at the investment opportunities there.

Speaker 1: On Softbank, what was your first meeting with Masa like?

Speaker 8: That was back in 2016 when they just had announced the purchase. So gosh, that was ten plus years ago. And his vision for ARM at that time was a little different than what we're doing now, but he had a big belief back then in terms of the world is going to need a trillion connected devices and everything needs to be connected on security and software and such. But my my first impression of him, and I had not met him before, but I actually had known and heard of him, was he has a lot of big ideas, and he's been consistent on that for the last ten years.

Speaker 2: So is your is is part of your role, like, the translation layer between that unbridled optimism and the pragmatism that needs to keep a business running? Like, how do you see your your your role in the SoftBank organization?

Speaker 8: That's a good description of it, think, at at a high level. I I think the reason it it works well is that I share Mas' ambition. I admire the big bold views. I like big bets myself. Mhmm. He's really really good at looking around corners and thinking about what are the big trends ten years from now and then how do we act on it quickly. So where I think I fit in is I align with that, but I'm also an operator. I'm on the ground running a company and have put out products and done solved engineering problems, etcetera, etcetera. So it's thinking about, okay, we take those and make those into realistic plans that we can execute on?

Speaker 2: Where should we go next? Jordy, do you have anything to move on to, please? I'm interested to hear more about your path to becoming CEO. Yeah. It's it's just like was it always on your mood board? What were the key steps? Like just advice for people who want to be in this position one day.

Speaker 8: Yeah. I I you know, we we meet a lot with with new employees, and and I I do get that question a lot. Was was the CEO job something that you you always aspire to? And and I would say that was really probably not the case. You know, again, having in the last twenty five years worked really closely with founders Mhmm. If I kinda look back in terms of what that experience was and how it got me here, first off, I I was always interested in in not only doing something that made a big big difference, but I always had an intellectual curiosity to doing lots of different things.

Speaker 2: Mhmm.

Speaker 8: So I've done probably almost every single job that exists in kind of the semiconductor world. I've done I've been in FAE. I've been a design engineer. I've been in sales. I've been a GM. I've been in marketing. I've done kind of all these jobs, not because I was trying to fill out a checkbox and said, okay, here are the eight things I need to become a CEO. It was more a function of I'm really curious to learn about these disciplines and learning more about how things all kind of come together. So what turned out is I was gaining more and more experience and more of those skills and working around great people like Masa and Jensen, I just started to learn a lot and became a sponge for for all of that. So when the time came for me to take this role, I felt very ready for it because I had watched CEOs do the work. I had done a lot of different work. Clear clearly, it's a different job. Mhmm. When when you're the CEO, your jokes are a lot funnier. People always say hello to you in the halls and stuff. Sure. And and you feel a different sense of responsibility, accountability for what's going on. But Yeah. More broadly speaking, I think I was very very lucky in that everything I had done up to my becoming a CEO had really kinda helped me get me there. And then what I tell young folks is get as much exposure as you can to as many things as you can. The one thing that I've always found that was a trait in in great people was curiosity

Speaker 2: Yeah.

Speaker 8: And and and and never stopping to learn. Because if you're never stopping to learn and you're quite curious, you'll probably land in a pretty good place.

Speaker 1: What what is your approach to creating your own sort of model for what the world will look like in in two to three years? The pace of of progress over the last three years has been, you know, really aggressive. But at the same time, you can consistently go back and look at people's predictions, a handful of people's predictions in 2023 and let's say 2024 that have turned out to be almost perfectly accurate. And so I'm curious how you kind of curate information sources to try to build an accurate model of what the industry will look like in two, three, four years. Beyond that, it's probably much harder. But what's the approach? Because everyone's sort of like selling something, right, that you're that you're interacting with. And your job as CEO is to try to actually figure out what the world will look like so that you can build against it.

Speaker 8: We're in a very fortunate place in the in the business that we run-in that we're sort of at the center of the universe. And what what I what do I mean by that? ARM's core business is is compute. We do we do CPUs. And CPUs are table stakes. Every digital product needs a CPU no matter what it is, whether it's the smallest device that fits in your ear or wearable or these giant data centers. And ARM is the most ubiquitous CPU platform ever invented. You you guys were chatting to being a platform, you know, over 350,000,000,000 350,000,000,000 devices have shipped based on ARM. So what we have inside our company is almost like a control tower view

Speaker 2: Mhmm.

Speaker 8: Of the entire industry.

Speaker 2: Mhmm.

Speaker 8: Right? We we just we see we see everything. We talk to everyone. And as a result, we have kind of an amazing perspective in terms of where the the direction of travel is going. And I I think back maybe to your question in terms of how do we how do we do and think about that. Because we have that view and vision and the products that we are working on and the stuff that's in the roadmap, it's not unusual to be sitting in a meeting and we're talking about the product that's going to be coming out in 2030, 2031, 2032. It's just the nature of what we have to do in terms of the business. So we have a really good view on that. And what we tend to try to look at is these gigantic macro trends. But some of the things always tend to stick, right? In terms of people efficiency matters, Low power matters. Compute matters. And when I say low power matters, it doesn't mean everything needs to run on a battery. But ultimately, you can't continue to have all these parameters go up into the right because the costs get into play, thermodynamics get into play, etcetera, etcetera. So

Speaker 1: It's almost like Bezos approach where he has some line. He can't predict the future, but he knows that customers are gonna want things, you know, faster and cheaper. Yeah.

Speaker 8: Yeah. Interesting. I I think that's mean, it's it's a good way to describe it. Right? Yeah. Those macro trends don't change. If we were having this discussion in 2000, that's still a valid comment. Right? Now, 2000, we'd be talking about, oh, the Internet is gonna explode and we're gonna have to eat about build out this fiber. We need all these lasers, etcetera, etcetera. It's subsided, but the Internet obviously is is foundationally everything we do today. Yeah. And AI is kind of the same. I mean, it's almost a utility. Right? It it it will be something that will be used by everything, everywhere, by everyone. But gravity will still apply. Right? People are gonna wanna make sure it's efficient. They're gonna wanna make sure it's economical, and they're gonna wanna make sure they're getting benefit from it. So when you think about product development, you have to have that kind of sensibility that makes sure that you're delivering on that.

Speaker 1: What what are what are you expecting from robotics broadly over the next two to three years? Like, I imagine you have an interesting you're able to look in some of these customer segments and maybe see, you know, areas that represent less than 1% of revenue starting to maybe grow at a faster clip. But I think people are asking like when are we gonna have the the chat GBT of robotics which I think people just bucket like a humanoid that can be economically valuable or have some like utility. But it's probably much more likely that it's possible it's already happened like in an industrial setting and it just it's gonna look different. But what's your what what are you kinda predicting in the in you know, before 2030, let's say, in robotics broadly based on what you're seeing?

Speaker 8: 2030 might be a good tipping point. I think I think it's gonna take a little longer than we think, and it'll be bigger than we think. By that, I mean the business models I think will need to be worked out. I think robots as a service is probably more viable than buying fixed one robots. 100,000 The bill of materials on those will be will come down, but at the end of the day, there's just a lot of parts. There's motors. There's actuators. There's computers. You're you're not gonna be able to make one. I think that's gonna be $4,000, for example.

Speaker 6: Yeah.

Speaker 8: But can you can you provide them at a cost where it can be a very very good replacement in certain areas for human labor in terms of what you're paying per year and have it subsidized? Yeah, I think that is that's quite real. And I and I do think, you know, it's a hard it's a hard problem to solve. Right? Because one of the one of the hardest problems around robots with AI is our fingers. Right? Our our fingers are unbelievably good at what they do in terms of sensing, applying the right level of pressure, being able to pick up a phone or pick up a 20 pound weight or picking up a needle to thread to thread a needle. That that for a robot is a lot of work relative to sensors. It's a lot of work relative to motors. It's a lot of work relative to ambidextry, but it it will be a solved problem. And when it becomes a solved problem, it's gonna be transformational. I mean, you think about infrastructure projects. Right? Just looking at the amount of time it takes for roadwork or putting up bridges and and anything around those spaces. You know, just look in The United States, the amount of infrastructure that needs replacement. If you had and you can't find the people to do the work, right? Those are very, very hard skills to go off and do that. Once you start to put real automation in place around that, it's going to be hugely transformational. So I think it'll take a little longer than than people think. We see a lot of, you know, amazing stuff with humanoids and dancing robots, etcetera, etcetera. But it will be bigger than than people anticipate. I think it will be gigantic.

Speaker 2: Staying in the control tower of the AI build out, the global economy, where are you focused on outside of your core business in terms of potential bottlenecks? We've been going back and forth on, is it fabrication, is it energy, is it permitting, powered shells, memory? There's been so many different bottlenecks that have sort of popped up and there's been a lot of excitement. And that feels like an opportunity for America or the world to focus on on alleviating a bottleneck when it comes out. You're helping a lot with what you're doing. But what else is top of mind for you as you view the potential bottlenecks to just scaling AI?

Speaker 8: Those are you hit on all the big ones. Okay. And they're all they're all they're all challenging. Right? I I think logic is a is a challenge. Memory is certainly a challenge. All the things that you mentioned. I think it is a fantastic opportunity on the flip side for The United States to be able to invest in manufacturing and support of these infrastructures. I know there's been a lot of lot of backlash, for example, about these giant data centers being built and and communities not wanting data centers in in their area. And then some people look at the data centers and say, these data centers are are so giant. They they don't actually employ a lot of workers. They're just giant Costcos with whirring motors inside. But I would argue that while that may seem to be true in terms of the when the operation is completed, the ecosystem that gets created by building out these data centers is immense. Right? And if you can have the entire supply chain, whether it's again around the motors, the turbines, the air conditioning, the chillers, the chips. Yeah. If you can if if The United States can put that all in in such a way that it's gonna be all vertically integrated, that's huge huge job opportunities for us. You know, I I I I read some of this in in in an op ed about 71% of folks are, you know, are kind of against the against data centers. And I was thinking to myself, gosh, this reminds me of almost back when fab when The US had most of the semiconductor fabs in the world, you know, thirty, forty years ago. And people weren't complaining about fabs per se,

Speaker 2: but there were a lot

Speaker 8: of complaints about the chemicals and things of that nature. And then you had, you know, labor shortages, and next thing you know, thirty five years later, every advanced fab, whether it's memory or whether it's logic, is not inside The United States. And and I would say just on these supply shortages, which we have now, which are real, I think it's a great opportunity for The United States to really boost up domestic manufacturing. And not just, again, like I said, around the data centers, but the ecosystem around it. If you think about the automobile industry, one of the powers that that drove for for The US industry was the huge supply chain around automobiles Yeah. That created lots and lots of jobs. I I think data centers are similar.

Speaker 2: Yeah. Yeah. It's always been Yeah.

Speaker 6: Good point.

Speaker 2: Point to Flint, Michigan as a place where water was contaminated. But if you trace back through the Flint, Michigan story, it is the departure of the General Motors plant in Michigan that actually led to all the renegotiations about where the water would come from and ultimately the degradation of the local the local water supply. It's sort of the flip of once the businesses pull out and the and the local economy stagnates, then you wind up with the environmental problems in an odd And

Speaker 8: it's and it's super hard to bring super hard to bring them back once they go.

Speaker 1: Yep. Yeah.

Speaker 2: 100%. I like it. Well, thank you so much for taking the time. We went overtime, but we this is a great conversation.

Speaker 1: Yeah. Thanks for

Speaker 2: We really appreciate you taking the time. Have a great my pleasure. Have a great rest of your week. We'll talk to you soon, Renee.

Speaker 1: Cheers.

Speaker 8: Thank you

Speaker 2: so much. Goodbye. Let me tell you about console.com. Console builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution for access requests and password resets. And without further ado, we are going to reindustrialize.