Foundry Robotics raises $100M Series A to build autonomous factories-as-a-service, ten months after founding

projects forward from start to finish. And up next, we have the co-founder and CEO of Foundry Robotics, Arch coming on the show to talk about building stuff. Again, thank you so much for taking the time. Please inter introduce yourself and the company. Tell us about Foundry Robotics.

What's up?

Absolutely. Thanks for having me, guys. Foundry builds autonomous factories. We build factories from the ground up. We orient them towards robotics and autonomy. We actually provide manufacturing as a service to our customers so they don't have to pay attention to how many robots are involved. Uh we are announcing our series A fund raise. It's 100 million bucks. The company's 10 months old. It's what 120. Do I get a

What took you so long? You've been at it for 10 months.

I know, man. And we're we're going as fast as we can. We need more factories.

Uh so so we're also unveiling our 130,000 foot facility uh across the bridge in Emeryville.

We're scaling up people. We're scaling up robots. We are doing everything in between as well.

Uh yeah. So cool.

So needed.

What is the uh the power law product going to be? What is the most manufacturable product? Cars, chips, I don't know. There's so many things that you could manufacture. Everything's manufactured, right? Tables, chairs, diet coke can.

It's a great point. So, fundamentally, right, because we build many things and we're flexible. I can build drones on Monday, batteries on Tuesday, and satellites on Thursday.

That means to answer your question, it's not a power law product, but instead more like a power law set of primitives. And so, we focus on assembly, we do precise placement, we do cabling, we do fastening. Uh, and so that allows you to focus on a ton of different products. So like right now I'm shipping to customers in defense. We do autonomous submarines. We do robot dogs. We do some telecom hardware. We have like six or seven paying customers right now that we work with. It's more about how can I build as many things as possible under one roof that sort of follow the same tenants.

Yeah.

And so for the first time in history, the more products that I build, the better I become at building all products. That's the beauty of it.

Yeah.

Yeah. Do you does do you think that robotics are going to basically break the sort of like laws of lean manufacturing? Right? Like when I hear like we make this on Monday, this on Tuesday, this on Wednesday, my immediate thought is like just because you can doesn't mean that would actually be the most efficient. But maybe maybe that is the future. If you want if we want more manufacturing in the US, it needs to be more malleable and and on demand. And then maybe it's maybe it could be wasteful to have a massive facility that's only set up to make one kind of thing.

Spot on. So there's this fourb block that people talk a lot about, right? There's like high mix, slow mix, and then high volume, low volume, right? And people always talk about, you know, high mix, low volume, or low mix, high high volume, right? Yeah.

But Foundry kind of sits at that like section of the quadrant that's normally grayed out saying it's impossible, which is high mix, high volume, right? And the and the reason that works, the reason it's desirable, right? I get this question a lot like, ah, you know, when a company scales, if if I'm a drone company and I scale to building a million drones a year, why don't I just run my own factory? Well, you're never going to invest the same amount of time, effort, and engineering into the factory as I am cuz the factory is my product. You're never going to have the same unit economics. It's not going to have the same autonomy, and you're not going to have the same data that I do. On top of all that, the underlying sort of capex rigidity that you have to deal with when you build your own facility to build the one thing like I don't have to deal with that. So we think about this a lot. My co-founder Emanuel, he ran the Model Y program at Tesla, right? Bestselling car in the world. And to launch the most recent Model Y, they spent a billion bucks in capex, right? Which is crazy. And so if you think about it, right, like the amount of time flexibility, product flexibility, design flexibility I can offer my customers by actually being a factory company that treats robots in the factory as the core product. It's actually pretty unique. Even as you scale, you still want to work with us.

What uh yeah, what what proof points are you looking for with these early customers that that's actually true, right? Because like I I think uh when you're like the question becomes like yeah I imagine there's a bunch of companies in the Bay Area that want to build a basically want to build a low volume of things and if and if you if they can buy that from you at a fixed price they're like great like you you're solving a bunch of problems for us but the question is like yeah they do start to ramp one of these products and you you you have to be able to retain them.

Here's the good news. Uh, our customers aren't just the startups, right? Like Huntington Engles, HII is one of our customers, right? We do work on autonomous submarines with them, which is crazy. That's like an OG product.

Yeah. No, they make serious.

They're they're not um

that's the right sound for hundreds of angles for sure.

I appreciate that. That's very fitting.

That's spot on.

But they're they're not they're they're not playing around, right? Like this this is this is a serious engagement. and as they think about scaling up their autonomous surface vehicle and autonomous undersea vehicle USVS and UUVs, right? Like they need a real production solution. Uh and so a lot of our customers, as it turns out, are not actually that small. Uh we do have some, don't get me wrong, we we have some startups that we work with, but they come from all over the place. Yeah.

Uh and so the very like real technical proof points we look at is like are we getting faster with each incremental unit? Are we able to see the improvements in the models across different base architectures and data collection paradigms across different kinds of products? Are we seeing the amount of overlap across the different paradigms? Like there's always this technical activity that goes into like our data collection program in conjunction with our model fine-tuning program which is actually what allows me to turn around and then on the manufacturing side go scale up. How

what do you uh if Foundry is successful, does the idea that like hardware has like more moes like go away? Because historic like for basically the last couple years, I think a lot of investors have been investing against this idea of like, well, I want to back things that have some hardware component because hardware is hard and you have to figure out manufacturing and your supply chain and all these different things. But it feels like what you're building towards is like somebody can go to their favorite agent and say like I want to make this sort of complex product and a smart enough model SL agent or harness will be able to like produce a file that you could then take and theoretically manufacture at small scale quite rapidly at at which point like it's unclear if the sort of legacy factory it be is still like you know a moat in in the way that it's And today

I think um the moat for a great hardware company is a mixture of your manufacturing prowess and your design excellence right and I think that will skew towards more of the design excellence. I don't build or design drones or batteries or satellites or any of those things. There are amazing companies out there designing amazing products. The way the way I think the future is going to go is when a company like Foundry is ultra successful and we have thousands of factories across tens of millions of square feet is that the speed of hardware design innovation and new products being conceived of prototyped produced mass produced distributed will skyrocket the way the pace of software innovation also skyrocketed when we saw the cloud sort of take over as the main form of how compute is handled. Right? Like no company sort of AI aside coming out of YC is like, "Oh, cool. How do I go buy more servers?" Like that's not like a thing, right? But today in hardware, if you want to build three of something, it's a total nightmare.

And so I think what'll end up happening is the economy will build so many more things and the moat will come from designing an excellent product, right? Because that's still not what I do. I don't design your product for you. That's your thing. I can advise, I can make it more manufacturable. We have this thing called design for robotic assembly that we've coined. We can talk about making it more manufacturable, but like there are drone experts, there are robot dog experts, there are satellite experts. They're going to keep doing their thing. My job is to just build a million of whichever one we need a million of. In the short term, is the factory mostly like cucka robot arms, spot welding, soldering, pick and place, be like putting in screws, or are there like CNC machines, 3D printers,

or or or are the parts assembled elsewhere and then you're taking in the parts? It's not raw materials in on day one, right? It's not like aluminum gets dropped off and I get a drone out the other side. Right.

Are you thinking what I'm thinking, John? We got we have a little internal hardware project we're working on that's kind of a pretty difficult challenge.

I actually am blinding on what you're talking about.

Send it my way. Whatever it is, we can probably crush it.

I I acquired the base.

Okay. Yes. Yes. Yes. We do have some

We have We have a challenge that if you can do this, I'm going to probably beg you to let let me invest in the company.

True.

I am almost certain we can do it just fine. Okay. So um let's uh uh basically today we focus on assembly because if you look at manufacturing there's two categories right there's fabrication and there's assembly.

Yeah. Yeah. Yeah. That makes sense. So like winding a drone motor that's going to happen somewhere else. You're you take the motor but get seated in the right place. Put the screws in. Make sure it's at the right tolerances. Get it all out the door which makes sense.

Exactly. Exactly. Yeah.

Especially cuz like if you look at the world now in defense as an example, right? We do a ton of defense. We no longer need like five fighter jets a year. We need like 10 million drones. Sure. Which is very different. Totally. And because that's very different. Uh the the combinatory problem of scaling assembly is what we spend most of our time doing.

Yeah. That's so that's so exciting. Well, congratulations and thank you so much for coming on the show.

Yeah, I love I love I love that you're working on this. This is like this is the thing that I I care most about is like going from from digital to physical. And I think that's where people are going to feel like an excel like a real like feel agile abundance is like the most impactful possible thing you could work on in my humble opinion but I'm not biased

physical industrial abundance. Oh my god. We can't even begin to imagine the implications of that in society. Like it's going to be awesome.

Yeah.

I love it, dude. Well, come back on soon. This is great. And and I will follow up after this with the with with

challenge TVPN bench. You can do it. You're goated. Have a good one.

Great stuff, dude.

We'll talk to you later. Let me tell you about Cisco. Critical infrastructure for