00:00First, let's start off with that $200 billion, one of the largest Series A's that we've seen
00:03in this space. And more importantly, I guess what I thought was interesting was there's no
00:09co-investor here. This is SoftBank going alone. What is Machi Oshisan whispering in your ear that
00:14he had the confidence to write you a check for $200 million? Yeah, I think SoftBank is absolutely
00:18doubling down in this physical AI space. And they're seeing, I think as you brought up in
00:22the previous segment, there's this big boom going on in AI infrastructure. And where we come in
00:27is we're essentially that first mover on the ground, right? We're bringing in the autonomous
00:31excavators that are accelerating that earth movement and they're accelerating that build
00:35out literally of physical AI infrastructure. But at the same time, we're seeing a spike in demand
00:41across all kinds of infrastructure, right? This is roads, this is quarries, this is the mines
00:45extracting the minerals and the metals that are being used to make the products, right? So where
00:50we're really delivering is productivity across that ecosystem. And I think that broad impact of,
00:56you know, automating essentially things like excavators and wheel loaders, these machines
01:00that you see on construction sites that are, you know, construction itself being 13% of the GDP,
01:05it's a massive impact that we can have by delivering productivity in this space.
01:09And just to be clear for our viewers, I mean, you're basically retrofitting existing
01:13excavators and equipment to sort of be either, if not fully autonomous, semi-autonomous. Is that right?
01:19Exactly. So we're putting a lot of things like LIDAR cameras, computers onto excavators to give
01:25them the perception to kind of understand, for example, the safety of the environment, but also to
01:29do things like surveyors do, like measure using a 3D scanner, the shape of the terrain, to be able to
01:35dig very accurately, but also very efficiently and very productively in those construction environments.
01:41What's the demand though for a retrofit in this space versus an actual, you know, custom, you know,
01:48purpose-built autonomous type machine?
01:51So, you know, excavators are a beast of their own, right? They're a complex hydraulic machine,
01:56and the market is very, very broad. There's a lot of OEMs in it. So, you know, if you look
02:01at,
02:02for example, the top three OEMs, they only account for about a third of the overall market share. So you
02:07have a really long tail of brands doing maybe three, four, five percent market share. And we've
02:12been able to go on a dozen of those different brands, right? And what's really important there
02:16is it's giving construction companies the choice of machine to use. These machines last oftentimes
02:22more than five years. So they're not having to go out and rebuy a new piece of equipment. But at
02:26the
02:27same time, I think you have to consider that a lot of these construction players, they're loyal to
02:31their local service providers who've been able to come out, show that same day they can replace a part,
02:36they can get that machine up and running without delaying the job. And I think if you've spent time
02:41in the heated environment of a construction site, you know that the last thing you want to do is
02:46stop or block productivity. And that's what gives a lot of loyalty to different brands, but also to
02:51our own solution. Well, that's what I'm curious about, though. I mean, we talk about, you're not
02:54naming them, but I assume you're talking about Caterpillar and Komatsu and a few others. I mean,
02:58they've had autonomous mining trucks, I know, for quite some time. So I guess the question is,
03:02what stops them from building, you know, sort of the rack themselves and basically cutting out
03:08Gravis? Yeah. So I think, you know, one important thing to note is we do work directly with a lot
03:13of
03:13OEMs. We work with companies like HD Hyundai and others. And we're really working to bring that
03:19technology out in a very efficient way to customers. So we see, though, that when you mentioned,
03:24for example, the mining trucks, those are vehicles which they have to just drive essentially from point
03:29A to point B. And they have one requirement, which is don't crash into anything, right? If there's a
03:33rock in the road, if there's a pole, if there's a person, you want to hit the brakes. An excavator
03:38or a wheel loader, any machine that has an arm on it, it's actually intentionally crashing into an
03:43environment. It's changing that environment as it navigates through it. And very importantly,
03:48the real ground is rocky, it's frozen, it changes in an instant from maybe soft loan to solid bedrock.
03:56And what we're able to do is use learning based control to essentially teach these machines to
04:01feel the earth, which lets them dig very productively, lets them take a full scoop of dirt, which when
04:06they dump that full scoop of dirt, using a very repetitive, robotic, you know, fast motion cycle,
04:11we're able to actually accelerate the amount of earth that can be moved in an hour on these, you know,
04:15heavy, repetitive tasks. Well, in addition to sort of the competition, though, from the established OEMs,
04:20I am curious, Ryan, just about this idea of some of the newcomers on the same day that we learned
04:25about the SoftBank's investment, Bedrock Robotics, which was built up by a couple of
04:29veterans of Waymo. It's backed by NVIDIA and I believe a capital G over at Google.
04:34They're saying their excavators are fully autonomous. Is that and live right now at several
04:39sites? Is that a direct competitor to you? So I think what we're seeing right now is
04:45the demand in construction has never been higher, right? So what we're very excited about is to have
04:51multiple competitive players, whether that be OEMs or other startups working in this space,
04:56who are going to work to deliver a competitive, safe and productive product to customers. And
05:02that's just going to accelerate the entire ecosystem. You know, I've spent my life, you know,
05:06committed to the built environment, and I'm really excited to see players coming in and being able to
05:11push. I think where we stand out is a couple of things. One is that we have a product which
05:16delivers
05:16full autonomy, but it also exposes all that information that the robot uses to operators.
05:22So it can be used like a co-pilot to show them where to dig, where not to dig, if
05:26there are people
05:27coming into their working environment. So we really let operators gradually adopt autonomy at their own
05:32pace. The other thing is that because of a learning-based system that is trained heavily in simulation,
05:37we're literally training in lifetimes of construction operators experience. So we're putting
05:42in essentially billions of cubic yards of earth movement in those simulation-based models. And
05:48that's letting us exceed an average operator when we deploy these systems. So we're seeing, you know,
05:53up to 30 percent or even sometimes 40 percent of expectations exceeded when we're doing that
05:58excavation. And then that last part, as I mentioned, this fragmentation of the market,
06:03you've got a lot of different OEMs. And we're the only player who's really been able to show that we
06:08can go on pretty much any machine and we can deploy that same learning-based control on a 10-ton
06:14Yanmar as we can on a 50-ton Catapult. All right. Well, I do have to end this by just
06:18circling back to
06:18SoftBank and really, and more importantly, Masayoshi's son. He doesn't cut checks like this just willy-nilly.
06:24And his expectations for growth are pretty ambitious. And we've seen that in the past. I mean,
06:29they expect big scale or bust. And I guess the big question is, I mean, are you prepared for that?
06:34Are you going to be like the next Arm? Are you going to be the next WeWork? I mean,
06:38where do you fall in all this? We're, of course, really excited for a very big scale. And I think
06:43what we've shown is a commitment to our customers to prove this out in real environments. We're
06:47operating in four continents right now on the biggest variety. We've got customers, you know,
06:51with half a dozen machines. We've got customers coming back to us for more. So we're really,
06:56really excited. And we're doubling down also on distribution relationships. We just announced
07:01last week an $8 million project to bring actually a turnkey rental solution together with the largest
07:07operated rental company in the UK out to customers. So we're really using this opportunity with SoftBank
07:13to deliver autonomy and use also this AI build out to slingshot our capabilities to be able to deliver
07:20to all the other kind of construction domains which need this, right? Build out housing faster,
07:25build out renewable and resilient infrastructure faster. We see this kind of AI boom as a
07:30massive opportunity for autonomous construction.
Comments