EP #089: Will AI Take Over Concrete Mix Design?

Seth (00:02.23)
Welcome to another episode of the Concrete Logic Podcast. And today I have Alex Hall, CEO of Concrete AI. Alex was, he actually did episode 33 back in December, 2022. So it's been a long time since we talked to him and I'm sure a lot has changed with AI and what he's doing with that in the concrete industry. So I'm excited to talk to him today. But before we get started,

just want to remind you all how this podcast works. So we, we do not, we don't have any advertisement, we don't have sponsorships or anything like that. And we do that on purpose so we can bring anybody we want on and, and not interrupt these great conversations that we have with our guests, with, with your favorite, deodorant or whatever these people have their average. So some of these podcasts, advertisements crack me up, but anyways,

So, there's three things you can do to support the podcast. The first thing is if you, if you enjoy the podcast and you enjoyed this episode with Alex today, please share it with a colleague in the industry. So they, you know, you share, share the knowledge. the second thing you can do is if you go to concrete logic podcast .com on the homepage, there's a couple of ways you can reach out to me. There's a contact, button you can click on. You hit the contact button. You can shoot me a message. there's also in the bottom right hand corner of.

Alex Hall (01:13.678)
Yeah.

Seth (01:28.15)
the webpage, there's a little microphone. You can actually click on that and it'll, leave me a message. It's just like leaving me a voicemail. Like if you were to call me on my cell phone, it would leave me a message. And what I'm looking for is ideas for guests or topics. so please do that. Cause this podcast is for you all. So I want to, I want to have the topics and the guests that you, you all want to hear from. And then the third way is same thing on the concrete logic podcast, on the homepage.

On the upper right -hand corner, there's a donate button. And what we ask is if you get any kind of value out of the podcast today, you learn something from Alex. Yeah. You hit that donate button and you can give any amount. There's no, there's no set amount. So, we appreciate any amount that you give. And when you do that, we, list you as a producer of the show. So with that, Alex, I, I was within the three minutes. I told you I'm doing good.

Alex Hall (02:27.022)
Thanks for watching.

Seth (02:27.51)
Alex, thank you for joining us today. Like I said, he is the CEO of Concrete AI. Alex, can you give us a, I think people know what AI is at this point, but it's everywhere. Everybody's following it. NVIDIA's rise to, I don't know where they're going. They're going to the moon and beyond, but.

Alex Hall (02:33.134)
Thank you.

Seth (02:55.222)
But anyways, Alex, can you tell us a little bit about yourself, a little bit about your background, because you do have a background in concrete materials, and then a little bit about your company, and then we'll roll right into the questions I got for you.

Alex Hall (02:58.35)
Awesome, thanks. Appreciate having me on again. I enjoyed our last conversation. I'm sure this one will be as good. Yeah, so maybe let me get me out of the way. So I was with Holcim for almost 20 years. I started with him in South Africa and had various roles with him.

moved with them in the US to 2007.

heading up there, really makes concrete performance group. I've been involved in the business side, VP of concrete in the Midwest, basically, and concrete and aggregate. I had a lot of fun, a lot of exposure there. That actually ultimately led to me meeting my two other co -founders here at Concrete AI through the University of California in Los Angeles.

We were always interested in disruptive technologies. And one of the things I came across probably in 2016 was this guy holding up a filter. And it looked like a 3D printed filter made of concrete. And it actually was. And it became what is the precursor to a product called Carbon Built. So Carbon Built's a commercially available block that's out there right now, a concrete block.

the way it's designed or the way it's structured, is it actually absorbs CO2 when you manufacture it. So maybe a really good guest to have on your show one day, a guy called Rahul Shendure, who's the CEO of Carbon Built. But super interesting company. They ended up running the 2021 XPrize for Carbon. And it's really a fantastic story that's born and bred in the US. Along some of the lines, we kind of kept on talking to the guys at the university and...

Alex Hall (05:03.278)
Yeah. Decided how we're going to try and utilize AI to kind of reimagine concrete design more than anything. So once you have enough data around the physical and chemical properties of the raw materials of concrete, you can kind of predict how things are going to perform. And once you accurate enough in your prediction, then you can start figuring out how, how you can model using the same materials, but to get other lowest costs or lowest embodied carbon.

And really, that's exactly what this business is all about. We typically will connect to a producer's quality management system, get data from there as to the actual raw materials. So think of all the, you know, whether it's the, you know, the aggregate, the sand, the cement, the cement substitutes, the admixtures, all of these products are supplied and they supplied in accordance with an ASTM specification.

So if you think about the aggregate, the aggregate always has to have a cert that basically says this is the ASTM C33 certification. And within that, there's a whole lot of really important data, not only around the grading, but physical properties, things like moisture absorption, fineness, modulus, the density, all these different kinds of things. But yeah, let's just aggregate. Think of C494, C150 for cement, all these different things. And you know,

because it's applied in accordance with an ASTM requirement, we can harvest this data and have a lot of trust in the quality of the data that's being provided. And so we take that information, we take the actual recipes that the clients have, and then we take a whole lot of test results, feed that into a set of neural networks or models that we've built specifically for this. And then are actually able to predict how things are going to perform. Now,

The reality is bad data is not good. So there's a lot of scrubbing that has to happen on the front end and make sure that the data is ready, accurate and relevant. But the bigger the amount of data we have, the more accurate we can be in terms of the prediction. So the business actually bifurcates into two different kinds of artificial intelligence. The one is the first that I mentioned, which is the predictive artificial intelligence that tells us, these are the materials that have gone into

Alex Hall (07:30.03)
a truck and based on that, this is how we think it's going to perform. So we'll give you things like set time, slump, you know, compressive strength at one, three, seven, whatever interval you want. Right. And that's all well and good because it's going to tell you how things are going to effectively work. But then if you kind of reverse engineer it and you say, it really makes producers go anywhere between say 15 to 26 different raw materials available at their operation.

including all the chemical admixtures. And now you go to the AI and you say, all right, let's take a different approach and let's use a different type of artificial intelligence. It's still within the same family of machine learning. But we're not just going to predict. Now what we want you to do is kind of do what a chat GPT does. And let's reimagine this. Let's give us a new mix design.

given the same constraints that we had from the previous design and the same engineering criteria that we're looking for and tell us what the optimal mix design for that looks like. And that's a different branch of AR, which then allows us to actually totally reformulate the mixes. You know, we've got to be really careful about this because AR is phenomenal at crunching numbers, right? And taking huge amounts of data and computing that.

It's like it's a really great combination of big data and artificial intelligence that you have all in one place. Concrete being as prevalent as it is across the planet, second most consumed planet in the world, there's a whole lot of data around it. And so you can get an output. But what's really important is that you also need a human being to provide some judgment to this. Because...

AI doesn't have judgment. It'll give you output of a calculation based on an algorithm that it has, but it's not going to be able to make a judgment call. That's where a human being comes in. And you know, as well as I do, I mean, think about how some guys, when they paste concrete, they're looking for a certain kind of frituability. And the AI is not necessarily going to know, that's kind of where I want this thing to be within the Shilstown curve, for example. Right?

Alex Hall (09:49.934)
That's where the human being comes in. So, you know, ultimately where we've gone is we've built this product, we call it concrete co -pilot. And the co -pilot is there ready as a co -pilot to the human. The human still ultimately the person in charge are going to be driving it, but the co -pilot really just helps them with some of these, you know, mass computation and number crunching, frankly, that folks don't have the time to do.

Seth (10:20.214)
Yeah. And this, we talked about this in the last episode you were in that, you know, this is not a hands -off thing that it's not going to spit out you the answer. And cause that was one of my questions. I think I asked you as, Hey, are we going to get rid of engineers now? And you said, absolutely not. so yeah, that, that, that, that touches on that. So is, is all the information that you're feeding, concrete AI.

Is that all, is that information coming from the just the ready -made suppliers or how deep into the materials are you going?

Alex Hall (10:58.254)
So we'll go all the way up to supply chains. So anything we can get, even millserts, that kind of stuff from the cement producers, like I said, the C33 documentation from the aggregate guys. Any information that's got relevant information with respect to the constituents and the properties of those constituents, we're going to use and populate within the models.

Seth (11:19.35)
Okay. Cause I, we just talked about this, I think a couple of episodes ago and we've talked about before is alkalinity and what it does to concrete. And the, the mystery, I guess is the best way to put it. The mystery is the, what those aggregates have attached to them. What, where, where's the cement coming from? What, what things that they're using to.

to help them make the cement. I'm trying to make this as layman as I can. All those materials go into this. And then this is sent over to the ReadyMix supplier. And the ReadyMix supplier is banking on what these milserts tell them. They mix this all together at the plant. And then it's on them to deliver this product to the job site and say, hey, this is how the concrete's going to act.

Alex Hall (11:53.166)
Well, I know you're right. I mean...

Seth (12:18.998)
This is what its strength is going to be. This is how long it's going to take to cure all these things. But really, you know, they're dependent on what the aggregate supplier is telling them. They're dependent on what the cement supplier is telling on them. I imagine since AI can crunch all this data better than a human can, I think we could do a better job at doing that is at...

part of what y 'all are doing.

Alex Hall (12:51.054)
Yeah, no, so I think what you're bringing up is super important because it's it's it's one of the reasons why there's very few companies doing what we do If you have a look at the AI space right in its entirety and You look at what they call kind of artificial intelligence for for material optimization You'll find a lot of different companies. There's there's robotics. There's mobile AI

detring, there's like literally a list as long as my arm of these different companies that are able to do things with materials and enhancing. But the reality though is, you know, think about this, like, concrete is the second most consumed product on the planet. Okay. But think about the number of ingredients that go into concrete. And I'm not talking about, you know, just ingredient categories. I think, think across the planet.

Think of how many different types of rock they are. Think about how many different types of chemical admixture and admixture suppliers they are. Every cement is slightly different chemically. We know this. Every SCM is also slightly different. So you've got like literally potentially millions of different input. And obviously the reality of all of these things is every one of them is a commodity. I'm sorry to the cement folks. They never like carrying that cement to commodity, but.

Ultimately, the distance you are from the cement plant is going to determine the cost of distribution. And that's going to have an impact on the cost of the powder in your silo. That's just the reality. So what do we know about commodities? Well, typically they don't travel very well. So the further you'll see cement go from a plant is where you got to plant on a waterway. And it can maybe travel five or six hundred miles down to a port somewhere.

That's as cheap as you're going to get it because you're moving tens of thousands of tons of product on barges. And that's how it goes. Aggregate, maybe a hundred miles if it's already good. Well, actually, let me take that back. That's also not true because here in Los Angeles, we get aggregate from British Columbia, for example. But again, coming on water, so the cheapest way of distribution. But if you're just moving rock on the road, maybe a hundred miles, I'm guessing.

Alex Hall (15:15.47)
And then, same thing, think about how far readiness goes. So this is a challenge, right? All the products are essentially pretty local, pretty centralized, and they're so diverse. You've got a range of different products that goes in. So modeling concrete is more than complicated. It's complex. It's a challenge to get this right. And just to add a little bit more complexity, think about the fact that you're using a series of...

admixtures with your concrete product, but how compatible is that with the grinding aids that the guys have used upstream in the cement? No one ever wants to jump into that, but that's something that I've seen that happen in my life. So it's a reality you got to be able to deal with. It's kind of one of the good things about when you model the stuff is that what you always want to try and do is you want to take into account the potential variability of these things.

And so, you know, for instance, with, with, with us, you'll, you'll always see the output of a prediction as an error range. So we'll say, okay, this thing is going to, let's say you've got like a 5 ,500 PSI target, right? And we predicting that the reference mix is going to get to 5 ,600, but it's got an error range of 300 PSI up or down. You know, so it could be as low as 5 ,300, but it could be as high as 5 ,900. Let's just say. Well.

When we model for mixes that we're going to verify or back up, we will always take that error range into account. And so we become really conservative to start with. So we'll actually slightly overdesign, but even with that overdesign and taking into account that error range that accounts for the variability of these different things, you can actually still have an optimized mix and still have your performance guarantees as well. So.

It really depends on how you want to model, but we've been pretty successful in doing that. One of the benefits is that over time, like I said earlier, the more data you get, the smarter things you become because what we employ is a process known as active learning. So as we generate more mixes and more of this stuff gets tested and we get more test result data as feedback.

Alex Hall (17:38.67)
So we continue to train these models and these models actually get smarter over time so that the degree of variability starts changing. But yeah, you know, this is a reality. The one thing you're never going to change is 80 % of the materials in concrete are naturally occurring between the sand and the aggregate. It's not a manufacturing process, right? This is a, you know, this is commutative. You're just changing the shape and you're sorting it into different gradations. That's what you're doing. And.

Rock is naturally occurring. The last time I looked, no one manufactures rock. Well, I better take that back. There's some that are about to come online soon, but anyway. But it's naturally occurring and like you can be in a quarry and like literally move 20 yards in one direction as a fisher and you can have a slightly different chemical reality which you're dealing with. And as a consequence, have slightly different physical properties. And so those are the kinds of things you always try to model and incorporate into the data.

Seth (18:37.559)
Yeah. It's, it's ever changing. I mean, we've talked about before is just talking about aggregates is, you know, if you look back three or four decades ago, we were using round river rock and now you don't see that anymore in concrete. So it's, it's like you said, when you're out there digging for rock, it changes as you go down, changes as you move over. And.

Alex Hall (18:40.878)
always.

Alex Hall (19:01.198)
Yeah, for sure. So a big ore body like a quarry is probably the best from a consistency perspective. But if you get into the Midwest, basically think of Detroit all the way through even St. Paul, those kind of places. Minneapolis, the reality is that at least 50 % of your material is coming from sand and gravel deposits.

Right. There's some limestone deposits, but that's basically an entirely sedimentary basin. There's not that many hard rock quarries there. And salmon gravel by its nature is significantly variable. I mean, you know, you've got some benefit and some of the particles are really somewhat rounded. So from a packing perspective, there's some potential benefit in there. But, you know, again, this doesn't always play nicely into the aggregate producer's hands, because I continue like...

2008, 2009 in Chicago, we had an operation out there and we had, I don't know, more than a million, five tons of sand and stockpile just in one operation. And you can only cap that to your sale. So you've probably got 1 .2 million tons of sand that in theory is on inventory and you've got to spend that cost over the rest of your products. And that becomes a challenge. I mean, that becomes a real big deal.

Seth (20:32.502)
Can you share what inputs you're using that you're putting into your model? I assume like compressive strength. Are you doing like flexural strength, temperature? Can you share a little bit what you're?

Alex Hall (20:43.182)
Yeah.

Alex Hall (20:47.95)
All those different things. That's kind of the beauty with artificial intelligence is that you're not limited to the number of dimensions you can model. So I would say kind of the vanilla stuff is exactly what you'd expect. F prime C, give me 28, that's what everyone needs. That's kind of typical. Give me a set time, give me a slot. So I mean, I think that's happening. Everyone wants that. But...

You can, you can, you can go even further. And the reality is as long as you have enough data to start training the models on, you can start going down into any rabbit hole you want. So, you know, whether it's temperature, whether it's flexural strength, you know, if you want to go as low as 20 hours, we've been able to do that as well in terms of predicting actual strength at 20 hours or a day even for sort of early stripping. I mean, you can do that.

Seth (21:44.758)
Yeah, yeah, we're something interesting that we talked about a couple of episodes ago was actually going beyond the 28 day was going, you know, some of these, you know, our type, our favorite type one else submit, you know, they're allowing us now to do 56 day breaks. But we were looking at we had an interesting chart where the chart was showing based on the temperature of the of the concrete that, you know,

Alex Hall (21:52.702)
yeah.

Alex Hall (21:58.926)
Mm -hmm.

Alex Hall (22:04.526)
90.

Seth (22:14.774)
when it was curing, that you might meet the 28 day strength, but the strength of the concrete actually becomes less over time. It had breaks at 365 days and it actually showed that the concrete got weaker because of the temperatures were not optimum for curing conditions. So the podcast was all about focusing on the importance of curing, but the...

Alex Hall (22:31.278)
Good.

Seth (22:41.59)
the chart that the gentleman shared with us, Robert Higgins, he was showing that if you're not at the right temperature, that you can actually have concrete that breaks at a less or compressive strength further down the line.

Alex Hall (22:57.774)
Yeah, well, look, full disclosure, we don't go past 91 days at the moment. That's where we've taken the model to in terms of the limit. I mean, I think that brings up an interesting point of the conversation because the one thing we are definitely starting to see is some more demand. It's not massive, but it's definitely increasing, especially with companies looking at both data centers and stuff. There's a real big focus on lowest embodied carbon concrete. And, you know,

Seth (23:03.062)
Okay.

Alex Hall (23:28.142)
What does that normally mean? Well, normally means you're going to put as much SCM as you can into the product. Just take it as far as you can and then let me understand what my, what my compressive strength is going to be at 56 and 91 days. And we've done that. It's, I think it's going to be interesting because as that market grows, I think we're going to see a little bit more, a different kind of pressure that.

folks aren't used to and they're gonna have to figure out how they can actually start supplying these products because you know, Slag and Ash are definitely increasing at prices that seem significantly higher from them. Just direct inflation in comparison to cement, they're getting more pricey and I mean there's some regions where Slag is more expensive than cement. So why would you use it? And this is just the reality, this is what we're gonna have to face. So...

Seth (24:15.414)
Yeah.

Alex Hall (24:21.518)
Producers are going to have to think a little bit differently, but probably more importantly is, clients are going to have to figure out that there's a green premium that's coming. And they're going to have to, that's just a reality people are going to have to deal with. And it's not easy. People don't want to pay more for what they can see to be the same product, but it's not.

Seth (24:42.87)
Yeah, they're going to have to prioritize. I mean, yeah, but just in 2020 fly ash mixes were a discount. And now, like you said, they're, they're premium now. So.

Alex Hall (24:52.846)
Yeah, but it's interesting because there's a flip side to this as well. It really all depends on the client. I mean, I was talking about data centers, for example, but if you have a look at any of the tech companies, I'm not quite sure where Amazon is on this train at the moment, but they've got a really excellent report in terms of where they're going on the decarbonization journey. But certainly, Apple, Meta, Google all have

net zero targets at 2030. And that includes the Scope 3 emissions. So Scope 3 effectively is anything not directly associated with the business, but it still relates to things like the assets. So if you own a data center, you probably got a big plot of land, a couple of acres, and somewhere in the order of hundreds of 250 ,000 cubic yards of concrete in some cases. And...

that's got a carbon footprint. And so there's real incentive for these guys to say, okay, if I've got a massive deficit that I've got to take into account, maybe it's, you know, I think Apple published an 8 million ton deficit, I stand to be correct on that, at 2030. But you know, the cheapest carbon they're gonna have to deal with is the carbon, they don't have to.

you know, remove from atmosphere. So if you can buy a low carbon concrete and it costs 20 or 30 bucks more a yard, it's actually cheap. Because if you look at the cost of a removal credit per tonne is today, it's not, it's not pretty. It probably averages out around $450, but I've seen it as high as $1 ,400. And, but, you know, I think the challenge we've got is that, you know, there's the technologies.

Seth (26:38.806)
Mm -hmm.

Alex Hall (26:50.926)
are there, but are they there at scale of 20, 30 to actually give someone 800 ,000 tons worth of CO2 removal credits? I just, good luck. Yeah, that's gonna be massive.

Seth (27:02.378)
Yeah. Are, are y 'all, are you putting in the cost data too as well? So if I said, Hey, can I have a mixed design that doesn't exceed $170 a yard and I want it to break at 5 ,500. You all have something like that?

Alex Hall (27:24.046)
That's actually part of what we're building right now, where you can actually just do everything from scratch. So at the moment right now, what we do is we typically will use a reference mix that a customer has and we'll compare it to that reference mix because we know that's been active in the market. And one of the things that you want to be able to do is a client needs to be able to take the product and have a like -a -like substitute, especially in commercial products. The customer still wants to feel like they're getting the same thing and not something new.

Yeah, that's definitely coming. That's currently something we're busy with. So when you move away in our world from kind of a subscription, you know, prescription mix, you can then start going to places where we can provide exactly that kind of service. So you give me the constraints, you know, you want a 6 ,000 PSR with a 1300 over design and a four inch lump and you got to pump it through a four inch hose and you want to set time of

four and a half hours and you don't want it to be more than 170 bucks, we'll go ahead and try and figure out what that looks like and give you an option and see what that looks like. I think, you know, one of the challenges with this is that every time you want something new and like a new add -on, it's additional development, right? It doesn't, it takes time. It's not like you can just flick a switch and the next thing is available. But.

It's kind of fun because everything that we thought was going to be important to folks in terms of getting things like what's going to add value, what's going to create greater savings for a client. It's actually maybe not so much of a priority. I think I've been really surprised at how a lot of the producers are, especially the quality control folks there. I was always hoping my QC folks would like.

always there to try and think about the cost optimization as a benefit for the client, but for the client for us, but also so that the product gives exactly what the client's looking for. But the reality is that the quality control people are already hands on. So many of them are busy in the field every day attending to details of clients that this kind of becomes something that goes on the back burner. And that's part of what we've been trying to do with the businesses, makers.

Alex Hall (29:51.95)
Low key as possible as light to touch as possible. So everything kind of works in the background and You know, he has a suggested optimization for you You can maybe save another five or six bucks a yard. Do you want to put this into a trial mix and see how it goes? So you want that stuff to be happening in the background so that folks can keep it keep on doing what they're doing because I you know, I mean, I'm sure you've seen this but Ultimately, I've seen a lot a lot of clients as much as they like talking to the production guys

They believe the QC guy when he comes to the job site and explains, hey, this is what I've seen. This is where we think the strengths are going to go. This is why you can trust the, you know, the ash has got a delayed reaction. It's waiting for the calcium hydroxide to be formed by the, you know, by the hydration of the cement. It's coming, you know, that kind of conversation gets people comfortable. And, and yeah, yeah. I think what we try to do is give people more time to attend to that part of their job. And.

Let them not worry about optimization. Let this happen in the back.

Seth (30:56.214)
So I imagine as they're punching in their mixed designs, your company's just gathering that information in the background. And then the results that they put in, again, you're just gathering that, collecting the data. And then that's what you meant by in the background. And it's a tool for them if they want to call up the information.

Alex Hall (31:16.302)
Exactly.

Alex Hall (31:23.374)
You know, and that's why it's a co -pilot. That's why we call it co -pilot is that what we want to do is we want to be able to push information to them as well. So like on a regular cadence to say, hey, here's a list of opportunities that are potentially available for you right now. Which one's the most interesting? Which is the one that's going to be the biggest bang for your buck? Or ultimately, hey, you've got this project coming up.

you know, they're ones that are calling for very specific requirements in terms of low carbon concrete, how can we reimagine it? What what you know, is there a price premium? And it you know, in terms of that price premium, what's the benefit for the client? So otherwise, what you want to also be able to do for them is to say, listen, it's going to cost, you know, $1 20 more yard, okay, but if you saving 30 % CO2 by paying $1 20 more,

This is what that means from an embodied carbon perspective. This is the cost of a removal credit that you're not going to have to purchase in the future. And, you know, this is the net saving of 30 % of reduction in CO2 will give you per cubic yard. So there's logic behind this green premium as well. You've got to be able to explain. And I think it's actually interesting. It's definitely tougher in the US than it is in Europe, Frankie, because in Europe you already have carbon tax.

So by virtue of the fact that you have a carbon tax, there's this kind of pseudo economy, well, it's not a pseudo economy, you've got the whole cap and trade system that's running there. But you can very tangibly understand what the cost of a ton of carbon is by understanding what you're paying in the form of carbon tax. And that makes things really clear. And whether we want to, I don't want to...

see any more tax, that's not what I want. But there is a cost to carbon and there's a cost to carbon avoidance. And people that are buying low carbon concrete fundamentally know that and understand it. It's something additive that they're looking for. They're trying to do an environmental benefit. But the reality is there has to be a cost to that. It doesn't come at a discount.

Seth (33:48.566)
Yeah. Well, I think, you know, whatever side of the fence you are on the, on the carbon and the carbon taxes and all that. I think the, the, what's really appealing to this technology to me is the, building efficiency and making and designing the mixes. that's, I mean, that's, I have said this numerous times on this podcast, but concrete got its start because it was a, efficient economical.

Alex Hall (34:07.406)
Absolutely.

Seth (34:18.134)
building material. That's why it pushed steel to the side. And it was easy to work with. And it seems like lately we're getting away from that. We're trying to figure out reasons to charge people more for the concrete. And what I'm afraid of is, yes, there is a premium there for a better product that we're providing the customer, but we're going to price ourselves out of some of this work.

that was designed for concrete and they're gonna go, concrete costs this much now, now I'm gonna go look and see if steel works or maybe, god forsaken wood. I don't wanna see any wood buildings. But yeah.

Alex Hall (34:48.078)
Mm.

Alex Hall (34:54.542)
Yeah.

Alex Hall (35:00.082)
No, that's not me, me too. I mean, you know, even, you know, I've seen these massive CLT structures that they're starting to put up in Australia. That's kind of interesting to me, but I had a really interesting experience recently. We were in India and so we've done our, we're midway through our first pilot in India at the moment. And absolutely fascinating, you know, just to give you context, I mean,

The one company we're working with there on its own produces around 25 % more cement in the entire US. One company. Okay, and it's like, okay, now you start looking at scale and you start scratching your head and thinking about this. And again, it's one of the interesting conversations I've had a lot with folks that want to talk about climate and impact the climate is that everyone thinks that what we do here locally is all that matters. Well, no.

because it's kind of like China putting a new 500 megawatt power station on the grid every two weeks. That's a global impact, no matter what you want to think. That's significant. Between India, Indonesia, Malaysia, and Sub -Saharan Africa, the growth rates are just unbelievable. But the one thing every single one of those countries is going to be using for the next decade, and to the extent that there's...

So I don't know. I'm going to find this place where I found the number. But basically cement and concrete consumption in the next 20 years is going to be equivalent to what's happened in the whole of the 20th century.

So in 20 years, you're going to see literally an entire century's worth of concrete placed. And the reasons are pretty simple. And you said it really well earlier. It's a great product. It's cheap. And it's beneficial. And this is really the key part of our mission is that how do we help concrete stay that product? Think about it. I mean, I don't think.

Alex Hall (37:11.79)
Any product has done more for humanity than concrete. Like you can, people tell, you know, penicillin, no, no, think about it. How do you have access to water? Dams, concrete dams. Now look, that's just, that's the reality, right? Think about the infrastructure we live in, okay? People don't like urbanizations. You don't have cities without concrete.

You just, you cannot have a steel and glass skyscraper. You couldn't do it at the scale that we have it, right? You don't have access to use the infrastructure to manage wastewater and storage without concrete. It doesn't happen. There is no substitute. And yeah, maybe the ingredients are going to change. You know, I saw a really interesting presentation by Sublime Systems, for example.

And they're talking about a megaton production. So basically effectively almost a, I wouldn't say carbon neutral, but like a global warming potential around around 100, 120 kilograms per ton of this product. But even in that scenario, we're talking about a million tons in 2029. That's assuming everything ramps up and scales up at the same. This, you know, the...

cement and concrete, yeah, there may be less of it, but you've got a 5 % compound annual growth rate in this industry globally. Again, I'm not going to just speak about, if you look at it in its entirety, this thing's going nowhere. This is yeah, and humans need this product to thrive. And that's the thing I love about it. Like, okay, now we're going to figure out is, how do we do that without pumping, you know,

8 % of the world's carbon that's going to increase to 12 % of the world's carbon in 20 years time.

Seth (39:09.365)
Yeah. Well, it sounds like as concrete changes and the composition changes, I'll keep you busy. Alex. You can keep, keep on.

Alex Hall (39:17.966)
Well, that's the idea. And honestly, that's part of why we love the platform so useful is that these are the products that are going to come online. You know, I think you got to love America. If you have a look at even the DOE funding that came out recently, all these literally tens of hundreds of millions of dollars that went out into new technology. So I know Sublime got some of that money.

companies like Summit Materials, all these companies are being incentivized by the government to go and start looking and deploying these new technologies. I mean, think of the way like these companies like Brimstone, Fortera, all these interesting guys that are coming online. And that's what we think so exciting is that we don't care where the material comes from. We're absolutely agnostic. Just let us have at it. If you get enough data, we can start predicting how this is going to do. So,

We want to be a compliment to those folks. We want to help show them exactly how those, and show producers how that technology can potentially fit in as a substitute for a product somebody don't know.

Seth (40:32.278)
Yeah. Yeah. Well, I think this is a good spot to end today, Alex. Thank you for coming on. I will make sure I put Alex's contact information, the website concrete .ai. Check it out. And so yeah, I'll put the stuff on there so you can get ahold of Alex if you want to work with Alex. Alex, thanks for coming back on the show. It's always good to always good talking with you.

Alex Hall (40:55.854)
I need to toss it.

Seth (40:58.774)
And folks until next time let's keep it concrete