What will it take to actually bring manufacturing back to America? In this episode of the Mission Matters podcast, Akhil and I sit down with Cameron Schiller and Aeden Gasser-Brennan, the founders of Rangeview, to discuss the current state manufacturing in the U.S.
We discuss:
What investment casting actually is and how Rangeview is revolutionizing the ancient manufacturing process
Why China dominates much of today’s manufacturing landscape, particularly in the world of consumer electronics
How modern defense manufacturing today differs from the manufacturing we did during WWII
How industrial policy combined with new technologies can help bring manufacturing back to America
You can listen to the podcast on Spotify, Apple, the Shield Capital website, or right here on Substack.
As always, please let us know your thoughts, and please reach out if you or anyone you know is building in the national security domain.
Transcript
Maggie 0:00
In this episode, we’re joined by Cameron Schiller and Aeden Gasser-Brennan, the founders of Rangeview, a startup revolutionizing investment casting through novel 3D printing techniques that are enabling the next generation of autonomous systems and advanced manufacturing. Investment casting is an ancient process—it’s been around for thousands of years—used to create everything from jewelry to jet engine components. But despite its importance in manufacturing critical parts for aerospace, defense, and automotive industries, the fundamental process has not changed much since the 1960s. Meanwhile, the U.S. has lost over 100 investment casting foundries since 2001, creating a strategic vulnerability at a time when our defense systems are becoming more complex and our adversaries are modernizing their manufacturing infrastructure.
Akhil 01:20
What’s particularly striking is how this old-school manufacturing process has become a bottleneck for innovation. As companies like SpaceX, Tesla, and Varda—as well as more established aerospace and defense companies—are pushing the boundaries of what’s possible with modern systems, they’re discovering that traditional manufacturing just can’t keep up. They need components with complex internal geometries that are really impossible to create using conventional casting methods—think turbine blades with intricate cooling channels or missile components that need to be both lighter and stronger than ever before.
Maggie 01:53
Enter Cameron and Aeden, two college students who saw an opportunity to bring this ancient industry into the digital age by combining custom 3D printing with advanced material science. Not only are they making the casting process faster and cheaper, they’re actually enabling an entirely new set of designs that were physically impossible to manufacture before. It’s not just about efficiency—it’s about unlocking new capabilities for America’s most innovative defense companies. Cameron and Aeden, thank you guys so much for joining us.
Akhil 02:22
Awesome, guys. Cameron, Aeden—so excited to be here with you. It’s been awesome to see the journey from your Berkeley lab a couple of years ago to where you are now. It’s been great to see what you’ve built and how you’ve executed. Before we dive into all of that—and honestly, maybe share with a good amount of the listeners what investment casting actually is—let’s actually start geopolitically.
So, over the last couple of weeks, a lot has been happening in terms of geopolitics, tariffs, and the state of the nation’s industrial resilience. To start, where do you think we are in terms of the state of U.S. manufacturing? We can dive beyond generalities since you’re living it on a day-to-day basis. Where are we at? What’s your assessment, and where does your work fit in?
Cameron 03:10
Well, the state that we’re in is—everyone listening, take a look around right now and see if you can find anything you can confidently point to that was made in America. It’s not great. Slowly, over the past 20 years, we’ve let it all slip overseas through globalization and the whole story of how we got here. The truth is, we’ve really hollowed out our industrial base for the promise of more freedom and flexibility—and what we got on the other end was a supply chain that’s more fragile than ever.
The great Americans responsible for making all the stuff underpinning our critical defense systems—and everything else in life—are retiring. Most of them are on their way out or already gone, and there are fewer and fewer people to make the parts we need. Those parts are mostly made overseas. So, for defense and energy applications—where we’re tripling the energy usage needs with the AI boom coming—there’s just no supply chain on the back end of it.
I’m excited to talk about that in more depth today. I think most folks don’t know just how bad it’s gotten, and that could be a real awakening for us if we ever need to pull that lever and start making lots of parts the way we used to.
Akhil 05:02
Definitely. And Cameron, it’s not just necessarily the globalization piece. Having watched you over the last couple of years, it’s also about the ability to innovate—the fast feedback cycle between new, novel systems and platforms, both civilian and military. It’s kind of hard to do that when you’re shipping a part every week or two from somewhere else. The cycle time to accelerate innovation isn’t just about production, but the unique design aspect as well.
Cameron 05:24
Yeah, that’s exactly right. That’s a big point that a lot of people are starting to pick up on now. When you’re building a system, the first thing good engineers do is engineer iteration rates. The easiest way to collapse iteration rates and reach higher levels of innovation is to bring manufacturing as close as possible to the final design space. Some of the most innovative folks out there are actually doing their own manufacturing. They understand the constraints of the system they’re building.
There’s this great art that’s been lost in America now, which is called DFM—or “design for manufacturability.” At a cultural level, it means understanding how something actually gets put together. We’ve lost that a bit. And believe it or not, most innovations happen in the design space of manufacturability. You can’t just make something new—you have to make something new that can move its way into the world. Understanding manufacturing is critical to innovation, and the way you understand manufacturing is by doing it. We need to do it so we can collapse those design iteration rates and really bring the best and brightest designs to the end user—the warfighter.
Aeden 06:31
I think there’s a point to be made here. Early on, when we were starting the company, one of the clear indicators that the old paradigm was collapsing—the idea that we’d do all the design here in the U.S., ship the manufacturing drawings overseas, and make all the money because we controlled the designs and the brand—was the rise of Chinese consumer electronics companies killing their American counterparts.
The greatest example is DJI, right? DJI—which is now obviously extremely relevant militarily—started as a consumer electronics company. Why does DJI do so well? It’s a good question. Why were they so much better than all the American drone companies, at least for a while? The answer is that they could manufacture everything down the street. Their iteration rates were incredibly fast. They could do more technically interesting things at lower cost and get those to market faster.
If you’re a manufacturing company—or a company that makes a product—your advantage is the ability to iterate quickly. That was a clear indication that something was wrong. The theory was that with America’s great universities and engineers, we’d build the best products. But that doesn’t work if you can’t manufacture those products in-house or iterate and develop quickly.
Maggie 07:43
And why is it that we lost so much of that capacity? And, you know, what are the levers that we are actually able to pull to bring that back to the U.S.?
Aeden 07:54
So, I have this kind of saying: the right industrial policy is the one that actually changes the unit economics of manufacturing in your country. Everything else is a short-term solution. As long as it’s actually more expensive to manufacture here, new manufacturing won’t come back.
I’m a bit of an outspoken believer in tariffs because tariffs do this, right? Strictly, it is more expensive to manufacture here. But if you—well, I think China’s a great example. It’s really worth looking at what worked so well. Obviously, it started with labor; that’s why we started sending things overseas. But over time, China’s real advantage today is that they are consolidated and have a ton of things in close geographic proximity. They have this ability to get things quickly, build new things quickly, and deploy new technology out to the factories quickly.
Maggie 08:47
They can literally walk down the street and talk to their friend who’s actually doing the castings or forging. And then redesign a part in the course of a couple days or hours.
Aeden 08:56
Exactly—and not only that, they can go down and buy the ceramics feedstock from their friend three streets over, and the mine that’s making the raw iron is only a few miles away. So this real consolidation of all these different causes in close proximity is the real benefit.
I think this is a huge project to build back in America. We had this—America was the king in manufacturing up until around the ’80s or ’90s. I think you have to start in the places that have some sort of DoD significance, where the interest is already there from the government. You need to start with who actually wants the manufacturing back the most and start building capacity there.
There’s a flywheel to get started: if you start in the processes that are most critical for the people who care, those processes will make it cheaper to do all the other processes here. There’s the idea of a learning curve, but there are also compounding learning curves in manufacturing. The more of everything you do, the cheaper it gets. The more things that are adjacent to what you’re doing, the cheaper it gets because you share labor and equipment, and things are amortized over more parts.
I think the policies the government has pursued have been a little misguided in a few key ways. One is that the government is really good at funding development—they love funding development—but there’s an issue: a lot of things get developed with government money and then never go to production. There aren’t enough incentives to actually build capacity.
If I were to change our industrial policy, it would be less about development contracts or qualification contracts and more about: if you build this amount of installed base capacity, we’ll fund your CapEx, or this becomes a tax write-off, or something to that effect. That actually changes the economics, as opposed to funding a million early-stage development contracts proving out new technology.
This happens again and again—America develops amazing new technology, and China implements it at scale. America invented the CNC machine, and yet, I don’t know what percentage of CNC machines are in the U.S. This will keep happening so long as the government just funds development.
China is a great example: funding CapEx and providing cheap government financing really does drive growth in manufacturing because it’s geopolitically relevant for them. It’s geopolitically relevant for us too. I think we’ll start to see this happen more and more, but historically we’ve focused too much on funding development instead of scaling. What we should really be doing is scaling more similar processes that have less of a technical moat—less difficult to develop—and focusing on building lots of capacity.
Akhil 11:30
Aeden, here we are talking about manufacturing — your focus on investment casting. There are probably folks listening in, or even, frankly, when I first met you, I had to learn a ton about what investment casting was and what you’re doing with it, what’s novel?
Aeden 11:42
Yeah. So I’m going to walk through each step of the process and explain what we do, what an official foundry does, kind of explain it from the customer’s view, right? Because that’s who matters at the end of the day.
So, you’re a customer. You want to make a part. You come with a drawing, CAD file, budget, requirements — density requirements, mechanical properties of the part, what kind of processing is allowed, what shape it is, how we measure that shape — all these things, right?
And you start out at a foundry. We do the same thing. We look at this and say, “Okay, how are we going to make this part? How are we going to set it up? What’s the right process to make this?”
For us, it’s largely similar across a lot of parts. Foundries, though, are actually very different. So, they’re making some tool to start — a big block of metal — and that block of metal has the shape of the part cut into it. Usually it’s actually tweaked a little, because the part might deform or change shape during the process.
The foundry starts by making a wax pattern. Rangeview doesn’t do this. The foundry makes the wax pattern, then assembles a bunch of these patterns together.
If you want to make a spoon, you have a wax spoon. You may take a bunch of spoons and put them together — they’ll have a wax tree feeding them. You’ll see this big wax tree that holds a bunch of wax spoons.
Next, you dip that in a ceramic coating — usually dip, dry, dip, dry — it takes a few weeks. What you’re doing is building up a layer, like a clay layer, on the surface. Then you melt all the wax out, burn the remaining wax again, and finally fire the ceramic shell.
You’ve got basically a big, complex mug — a strong ceramic mold.
What Rangeview does is make the mold directly. We print direct molds, so we skip all those steps — all the manual labor — and, really importantly, we skip all the variability in the final part that comes from human decisions. You want the technical characteristics of the part to be driven by software and machines, because then it’ll do the same thing every time. Your tolerances are much tighter, and you can rely on the process more.
At the end of that whole process, you’ve got a part inside a big ceramic cavity. Then you pour metal into it. There are a variety of different casting processes — Rangeview works with one that produces reactive or very high-temperature alloys. You pour the metal into this mold, remove the ceramic after it cools, and you end up with a single part.
You throw away the ceramic each time and remake it for every new mold. That’s how you make the part.
Then you cut the part off the tree and do a ton of post-processing — you might take X-rays of the part to look for porosity, coat it with a penetrant to show cracks, cut some parts apart to pull mechanical tensile bars and test strength. Then you’ll heat treat, machine, and coat them. Generally, at that point, you’ll have a final part.
So, the whole process traditionally involves all this fixed tooling to make the part. For us, it’s the same process every time — we directly print the ceramic mold. It’s a process we’ve built in-house, and it allows us to make different parts without changing any tooling.
Today, the U.S. has offshored almost all of its investment casting capability. Investment casting is used in enough of these “commodity” capabilities that people have said, “Well, this isn’t defense-critical, maybe we can push this overseas.”
But there are a lot of cases where people actually got exemptions to manufacture critical parts overseas because it was cheaper. Investment casting is full of labor — the traditional process has hours and hours of manual work. You might have 40 hours of labor going into a part that fits in your hand. So, in that case, it really matters if you can save another $20 or $30 per hour of labor. When you have a very unautomated process — which is the traditional method — that’s a huge factor.
Akhil 15:35
That’s awesome. It sounds like the previous instance hasn’t changed for thousands of years — some sort of wax base to ceramic, you pour metal into it, and… great.
Aeden 15:46
Yeah, it started with beeswax and river sand — that’s how this process began.
Akhil 15:50
And critically, now you’re able to, because you’re 3D printing and designing this in a digital foundry way, actually make a whole new set of complex geometries that you can pour whatever you want into, correct?
Aeden 16:01
Yeah, and we’re enabling a whole bunch of designs there that you couldn’t make in the traditional process. Those are some of the most exciting parts we make. And it’s also worth saying that almost every one of those post-processing steps I described are generally done by people.
So, the way you cut the casting off in a traditional foundry is a guy sitting on a saw. This is a horrible job, right? It’s a super dangerous job. It’s super loud, the saws break and send bits flying—you know, people die doing this. It’s really a horrible place to work. It’s not somewhere I’d want to work. It’s not somewhere I’d want my kids to work.
And it’s the same thing with everything else—the person grinding the part, the person heat treating—it’s a bunch of people stacking parts in a big furnace, closing the furnace, opening the furnace. In our factory, almost all of this is done by a machine. The person is programming the machine, setting it up to do a new part, but they’re not the one actually cutting the part or grinding it. That’s all done by a robot, really, at the end of the day.
The traditional process to make any of these parts—and any of the traditional competitors—has a very inflexible process that cannot be changed without huge budgets. You’re making a part, you want to make it slightly differently, you want to iterate on a design—it could take six months, a year, two years.
The second piece is there’s a huge amount of extremely skilled labor that goes into this, and it’s really hard to scale. So, if you’re an investment casting foundry and you want to get bigger, you want to double your capacity—there are no people you can hire. These people have all retired.
So, the things that are built are factories that are not dependent on specialized human labor to make components and are not dependent on tooling to make parts. What these foundries look like is—you want everything to be made by computer. How can you give the computer the tools to make your casting?
We use a variety of different processes that make sure each manufacturing process in our factory is software-defined. This starts with making molds. You want to make sure the molds can be made in a software-controlled way. We’ve done a huge amount of innovation on the process there—in the equipment and the materials—to make sure that when we make a mold, there’s no fixed point associated with it, and you’re forming it all digitally. The same thing applies to every single piece of equipment and process that comes after.
Then you’ve got eight or nine steps in our process internally to do that, and for each one, we have a digital definition. The machine is running with telemetry going back to the system and running to the prescriptive definition that we’ve given it for that process. This includes things like pouring—what temperatures you pour the metal at, how fast you pour it, what kind of environment around the metal, the vacuum levels, and things like that.
The big thing is: our factory is not running on highly specialized labor; it’s running on software-controlled equipment. I can form different parts with just changes in software and without retraining people. So that gives you the high level—I can dive into all kinds of details, and I’m sure Cam can talk to that.
Akhil 18:51
Super, super helpful. Maybe one more high-level—just walk through, if you can, even just a couple of the steps from design to 3D printing to cast. What does that sort of look like?
Aeden 19:01
So, traditional process here—and what an average aerospace customer does—is they want to make a turbine blade. They have some preliminary design, and they want to get into production. First, they have to get to design validation. The goal is they want to be making turbine engines in six years, at rate.
Generally, what will happen is they’ll go out to as many foundries as they can and try to find someone who will bid the part. The first surprising thing is that most of the parts get no-bid most of the time, because most foundries are overbooked and don’t have capacity—and also can’t scale. They’re stuck because of their labor.
So, the first process of design, at least for you as a customer, is to try to find someone who will bid on your part. You might not have a finished design at this point, so you’re going to tell them, “Hey, look, I want to make a few of these. I want to do some prototyping.”
Generally, they’ll give you back two really important quotes. One is your upfront quote, and the other is your piece price. The upfront engineering and tool making might be $300,000 for a part and might last six to eight months. Then you’ll go into that and also provide a huge amount of technical documentation.
For each of these parts, if you’re making a turbine engine, you might have 100–200 pages of technical documentation for a single small part. That’s a contract between the design engineer and their requirements and the foundry. What the foundry is agreeing to is: “I’m going to meet all of these requirements.” It might take many iterations to meet those requirements.
For the foundry, every single one of those iterations means iterating on tooling, retraining people to work on the new parts, and trying to hit those requirements—and that’s what stretches it out.
For us, when we receive something, we’ll ingest it, do some design, and essentially just put it straight onto our printers, which make the molds for the casting—and then get results quite quickly.
Akhil 20:47
And these are 3D printers, right? Standard 3D printers.
Aeden 20:51
So yes, we use a bunch of in-house developed materials and processes to make these castings and molds. Essentially, we’re forming the geometry that’s going to form the part with a printer. The end part is actually just a normal casting. So, unlike a lot of the additive processes—the metal 3D printing processes—there’s no requalification, there’s no retesting. The material properties are the same as these casting alloys that have been developed for 50 years, and then we get a part straight out of that process. Maybe that gives us a high level; I can dive into more details.
Cameron 21:28
Thanks, Cameron. I can come back to you on the discussion we were having around the iteration cycle. Can you walk through an example—whether it’s a real or hypothetical customer—how you would have done here, both Digital Foundry and the novel casting process? The full lifecycle has just allowed for that incredible acceleration in both speed and scale.
Yeah, absolutely. And I’m not able to use direct customer names in the line of work that we do, but I’ll give an example of something that we see happen and how ultimately it results in, you know, American supremacy on the battlefield.
So, as Aeden was speaking about a little bit earlier, when you’re making a part, the end designer is actually just delivering the capability. And that capability may be, you know, let’s build the engine that can fly the furthest. Let’s build an engine that can fly the fastest. And ultimately, that comes into a balance with reality and physics. When you try to put those metal particles through their paces in a jet engine, it’s going to be how hot you can get it.
A really common one that we see—because we do a lot of turbine blade work—is: what is the exact shape of that turbine blade? We’ve gotten good at simulating physics to some level to help us figure out how the air might flow through that turbine blade. But the real way to figure it out is to put it on the test bed and let it run. The simulations are never perfect. Simulations are only as good as the data you feed them, and so there may be a few variations of how that blade should be shaped that a designer is looking to go through. What they’re really trying to figure out is what shape is going to be the best for production on this engine.
Every form of manufacturing has some inflexibility associated with it. The way to think about this in manufacturing—the term is called tooling. Tooling can be hundreds of thousands to millions of dollars to set up a set of machines to do a set of instructions. And this tooling is hardcoded—it’s coded in the actual cavities of the molds that you’re producing. So when you ask that tooling to do something a little bit different—to figure out if that little bit different shape on the blade results in a little bit more or a little bit less performance—you have to pay hundreds of thousands or millions to change it. But once you’re in, you’re good to go.
The qualification process of a lot of these components often involves changes. And when you move that tooling from a physical world to a digital world, you unlock a whole other world of possibilities—not just for cost reduction, but for achieving the optimal design.
Now, there’s this other kind of intricate layer that’s starting to come into play as America becomes more serious about producing. This is maybe a bit more general of a comment on the state of the DoD world right now, but lots of folks are delivering to kind of low-rate initial production. But a lot happens when you go into full production. We’re talking about what is the capability—not just to make engines that fly as far as possible, but how many engines can you really make a year if you had to make every single engine that you could? It’s a different scenario when you have to hit full production.
In many cases, when people are designing today for these flexible manufacturing technologies, they’re using a different manufacturing technology. You might metal 3D print something, because then your tooling is digital. But when it comes time to make full production, you’re moving to a traditional form of manufacturing—you’re moving to casting, you’re moving to stuff that just scales. So one thing a designer has to look at is how serious they are about production. And if you’re serious about production, you’re going to design to the constraints of the manufacturing technology that you use to scale production.
This is one of the core beliefs that we coded into our technology stack. Our technology is capable of iterating on quick-turn qualification to allow that part designer to get the shape and the blade that they want. But at the flip of a switch, they can go to full-range production with the same constraints—the same casting constraints. We don’t have to change our blade thicknesses. We don’t have to change our overhang angles. It’s directly the same manufacturing technology.
And then I’ll give you another one, which is on the tail end of this. We’ve seen this with past sustainment issues before, but at the end of a lifecycle, you store that tooling. And you store that tooling on a wall while you’re making the next, newest, and greatest part. That tooling, believe it or not, will get lost, rust, or wear down. The tools wear down over time. And when you go to pull that part back off the shelf—when you need to make 14 joysticks because a plane is down and that joystick has to be remade to put it back on—you have to build up that $100,000 tool and that million-dollar tooling all over again, just so that you can serve a hilariously low quantity that the industry calls alien castings. They literally call them alien castings, or these onesies, twosies that they need just to sustain stuff.
So you have this other flexibility request on the tail end of these programs. And once again, this is what we’ve designed our production technology to do. And I think the bigger picture, when you zoom out from all of this, is you’re looking at manufacturing holistically. You look at the problem—it’s not just about making one part, it’s about making a process that can consistently make things at scale. And that’s what Rangeview is about.
Akhil 27:35
About. That’s right, and that’s your sort of idea to speak of the digital or cyber founder, Cameron. What’s the Cyber Foundry? What’s the vision for it?
Cameron 27:43
Yeah, Cyber Foundry is Rangeview’s first product. And the Cyber Foundry is simply a factory that is making America’s most critical componentry with the workforce that America has and with the technology that we invent here. So it’s a collection of custom-built processes, materials, and equipment to make a digital file of a part—which we’ve gotten really good at building in America—into an actual part, a real part that can be certified and clicked into a plane, ship, or vessel, and unlock the capability of those systems that we’ve designed so well but have been hampered by not being able to be made. You know, unlock that problem for the U.S., DoD, and our other partners.
Akhil 28:43
Awesome. That’s from digital design iteration, digital manufacturing, and then actually the real hands-on production.
Cameron 28:51
Every stage. I mean, our factory is just as much of a computer as it is a metal-shaping device.
Aeden 29:00
Yeah, I think it’s worth saying a big part of our philosophy is that the next generation of American manufacturers has to have a huge amount of engineering expertise in-house. You’re not going to innovate on process if you don’t have the expertise on how your process works. And it’s really important that the manufacturers—not just the designers, but the actual people designing the process, building the equipment, setting up the factory—know what’s going on and are incredible engineers.
Maggie 29:23
What role do you see for new technologies in accelerating reshoring of manufacturing capabilities to the United States? And what are those technologies that are actually going to drive those trends?
Akhil 29:39
Yeah, and maybe, Maggie, in addition to that, to your earlier answer, Aeden, we can’t necessarily replicate what other nations are entirely doing, right? We have certain advantages here. And you guys, among many other things, are innovating from a technological standpoint—from both a digital and physical standpoint—the software-defined aspect. So, a little bit of an addition to Maggie’s question: how much can we actually rely on technological innovation as part of it, and then what else needs to be done?
Cameron 30:07
So, technology is the way forward. It is the only way forward for us. But that’s a great American trait—America knows how to do that, and it’s not that complicated. The first thing you’ve got to do is pull your manufacturing technology into the 21st century. I mean, it would blow you away if you walked into a foundry today. All the folks—the heat-treat shops, the anody shops—all those folks need to pull their furnaces into the internet. They need to bring laptops into the work and just run it the way a modern company is run. And so I don’t think that’s reinventing the wheel; I think that’s getting more focused into the industry.
So, you know, it’s tough because the factory owners—that was a bit of our origin story. We were just two robotics guys, gonna sell robots to factories. What a journey. You know, we thought that. We thought that manufacturing was further along and was just ready for someone to sell robots. But the truth is, they didn’t even have outlets on the walls. So we had to go down to the studs and start there. So, you know, I don’t think it’s reinventing, no. And there are many companies doing this right now. I mean, go look at SendCutSend—this amazing company that chose their manufacturing technology.
That’s a good way to think about manufacturing too: at a company level, you go to the processes. You know, manufacturing is down bad as it is in the country—it’s actually still a bigger market cap in America than all of tech combined. And so if you think about it, you can’t compare a FinTech company with a healthcare company—they’re completely different sectors and completely different things. And manufacturing—the process—is the sector.
So there’s this amazing process, sheet metal fabrication, where you’re making stuff like doing origami with really strong paper. There’s this company out in Reno that just figured it out. You know, the founder’s a software engineer that came from doing SaaS work, just wanted to help his country out, and he’s built this amazing shop that’s, yes, software-enabled, but it’s just the modern tools of manufacturing put into a process.
I think a lot of this is going to come. I think this is what’s happened elsewhere. There’s this strategy that China used when Xi Jinping was building—it was an actual initiative you can look up—it’s called the “Made in China” initiative. And they very cunningly required American firms, in order to access the cheap labor and other subsidies associated, to partner with a Chinese factory partner to learn what they were doing.
And of course, we now learn what happened: we taught them all how to do it, and then they kicked us out. Uber—this happened with Uber, this happened with Google—they were operating in that market, they told them how to do everything, and then it was just a one-for-one move over. Chinese factories are probably running the best processes right now in the world.
So I’ve seen a few places where you can flip the script and say, well, there are some great Chinese manufacturing approaches—all the equipment is modern. You go buy the best manufacturing equipment. We didn’t just lose our factories, right? We lost our equipment. There are no factories to sell to college or business.
So, if you want to see what a really good piece of modern industrial equipment looks like—natively digital—it’s got an Ethernet plug, goes straight to the modem and the router (which is why you have to be careful). But just go look at the way those machines are built overseas. So maybe there’s an opportunity there.
But we really don’t have to reinvent the wheel. We just need to get a lot of folks interested in doing this again. I think Aeden is totally right—you set that up by getting the unit economics right, and the unit economics are really, really wrong. Technology won’t do it alone, unfortunately. I wish you could innovate your way out.
But, you know, if you do a little bit of Excel, it’s frustrating. You look at some of these parts—the part costs more to make in America, if you just paid for the electricity and the metal, than it does to have a fully turned part in China. Unless you can develop technology that makes the metal free or the electricity free, or a 99% reduction in labor because you have this perfectly automated factory—or equipment that costs $1 because you figured out how to scale this stuff to the moon—it doesn’t matter. You’re still losing. That’s because of subsidies. That’s because of vertical integration. I mean, it’s been weaponized. We’ve totally, totally, totally come up against weaponized industrialization overseas.
Akhil 35:12
Yeah, Maggie and I come from the other part of California, the Bay Area, where you can’t throw a rock without talking about AI. There are aspects of AI—or let’s just call it software-defined generally—but as you look at the sort of AI space, a lot of it’s digital, but fundamentally a lot of it’s moving to the physical world. What actually makes sense? And I hate using AI as a sort of, you know, “we sprinkle it in our coffee,” but where does it make sense to actually apply either a machine learning or some other artificial intelligence capability when it comes to the physical data?
Aeden 35:44
So, the truth is, it’s much harder to build stuff in the physical domain. It’s much harder to build stuff in the physical domain when you have extremely safety-critical applications depending on it, right? And so it’ll be a while before an AI system can make a decision that is accepted by, you know, one of the, as we talked about, conservative aerospace primes. That said, there are so many parts of operating a factory that are not as critical, where it comes in first in places where then things are going to get verified by a human, eventually verified by a test. And so, you know, we do all kinds of things at every level of this—from extremely simple things like how we inspect parts. There’s a whole bunch of things that are normally done by a human, verifying that a certain number is less than a certain number and greater than a different number, right? And we do a lot of that with software and a lot of that with, you know, I don’t want to call it AI, but more sophisticated software systems, often that are trained on a lot of data. And it’s the same kinds of approaches at the very high level. And we’ve started to toy around with, you know, an LLM that has access to a huge amount of the data in the factory. Is it making all the critical decisions today? No. Will the software get there and the technology get there so that it can sometime in the future? I actually think probably yes.
Cameron 36:55
So AI is only as good as the data that it’s fed, and the amount of data that we have, for instance, in something like self-driving cars, is orders of magnitude above the amount of data that we have on part making. The first problem is, part-making data is in the heads of humans that are retiring right now. Like, this is the first problem—let’s get these into the computers, right? And then after that, we can start to put sensors in place and directly train that system. So I think a good goal is for computers to know just as much about part making in a few years as they do about driving cars.
Akhil 37:38
When you talked about addressing this via technology, processes, and people, I want to actually come to the people part—and we can dive into more of the other two. People—I’d love to get your take on talent, what you just talked about. But actually, first, I want to come back to your story of you both in Berkeley, working on robotics. I remember first meeting you all one summer on Shattuck, at UC Berkeley, in, you know, a small little place. Share a little bit about where you came from. You shared that initial story about, you know, wanting to sell robotics, and then you turned into casting. What was that story like?
Cameron 38:15
Yeah when you met us at Berkeley we, you know, we were in a garage—one of the garage startups. So, in a garage before Berkeley, it was a huge upgrade at this old Sprint store in downtown Berkeley. Border does, because Berkeley is a crazy place to build what we’re building. But yeah, you’re right, we met you there—it was a great day. Yeah, yeah. I mean, we wanted to reindustrialize America, and we thought that the way to do it was by building robots. And I think a lot of folks see this right now, where, you know, we just have to add modern technology. We have to automate. We have to bring robots in and have the robots do the human tasks. We charged you that Rangeview was actually Rangeview Robotics—that was our formal name there.
Akhil 39:01
And that came up because you both were robotics champions in high school, right? Two at one point—is that right?
Cameron 39:07
Yep, yep. We were competitors, and Aeden and I have known each other for a very long time. Yeah, we did competitive robotics against each other — you know, Los Angeles BattleBots, the fighting ones on TV. Even before that, Aeden and I did a lot of high school robotics. And by the way, amazing program — talk about getting started early. I don’t know if I’d be doing this, or if you’d be doing this stuff, if robotics wasn’t a big part of the early education system. And it’s amazing. There are some fantastic people who built that whole system out — FIRST, back to robotics. There’s a lot of great stuff. China’s doing a really good job at robotics education too. DJI actually built probably the best robotics student competition in the world called “Masters,” and it’s incredible. You know, there are roboticists in China who are held to the cultural significance of, like, a quarterback in America.
Akhil 40:04
That’s crazy to me. You’ve got Texas football, and then you have a DJI roboticist.
Cameron 40:09
Yeah, a little bit different in America. You know, it’s hard because Aeden and I were definitely on the nerdier side growing up. So, you know, I wouldn’t say that people — put it this way — people weren’t inviting us to parties because we were good at robots. Maybe one day that’ll change. It depends on how much we really need robots to build. I think we really need them to be built.
So yeah, competitive robotics — that was it. But we looked at it and we said, wow, we need to decouple from labor. There’s no way America, with our way of life, will ever be able to support the hourly rates of overseas labor. And if we’re going to automate manufacturing, this is the big question. Everyone wonders, “Why manufacturing in America? Why don’t we nearshore to another country?” But the truth is, if you decouple labor from the manufacturing cost function, it doesn’t matter where you make it from a cost perspective. And where you make it will bring all the jobs — the people that build the factories, the people that build the robots.
That still means that people are critically important, but their value is being accrued in different or complementary portions. Think of it this way: think about a farmer. A farmer is an amazing roboticist. They’re doing the work of thousands of people with one big tractor. The tractor maker didn’t replace those jobs — the farming industry and agriculture created abundance everywhere. And we can go out and specialize and do all these great things. But what you’re doing is giving a human leverage.
The comparison I made on manufacturing in America is that it’s critically important that that leverage becomes American leverage. So we have to build factories here — not factories for everything. We don’t have to make plastic forks. There’s some stuff that, you know, we should share across the globe. But for what we’re working on, we chose very specifically to work on the things that America deserves to be excellent at.
Akhil 42:25
So you went from Rangeview Robotics to Rangeview.
Cameron 42:30
Yeah, we just lost “Robotics.” It’s like, just Rangeview.
Akhil 42:32
It’s cool, but you haven’t necessarily lost that mission, because a lot of the components you’re now providing will absolutely carry forward, whether it’s robotics or just the autonomous systems wave of the future.
Cameron 42:36
Robotics is in our DNA. Yeah, that Rangeview Robotics is in our DNA.
Aeden 42:41
It’s worth saying that the counterfactual between two people losing their jobs to robotics is not actually “automation or no automation, jobs or no jobs.” It’s actually, “We don’t do manufacturing in the U.S., and no one gets any jobs,” or “We do it here, and it supports people at the labor rates that they want to be paid here.” It’s either good jobs or no jobs. That’s the only counterfactual that there really is.
And so it’s a huge part for the factory to have very few low-wage, low-skill positions, right? This is not what Americans want to do. It’s not what the nice factories are going to be like. And it’s not actually how you make good parts at scale. You don’t want to be reliant on people that aren’t that motivated to be there to make your parts.
Maggie 43:23
So, you know, at least my sense from the outside is that the manufacturing industry — the aerospace and defense industry — is a pretty conservative industry. What does it actually take for you guys to build trust with these customers and for them to entrust you with building some of their most critical components?
Cameron 43:40
Yeah, I think Aeden’s got a really good bit on this, so I’ll give it to him. But these systems already exist. Actually, there’s an entire quality industry whose whole job is to take the liability of ensuring that a part does what you say it will do. And that is something from day one we have designed for, because there have been other innovations that have been hung up on this—saying, “Well, if we just make a cool enough thing, people will want it.” The truth is, you have to be able to prove that that part will behave the way you say it will in order for it to have real value over time. So I’ll kick it off to Aeden here for comments on qualification and what that all means.
Aeden 44:23
Yeah, no, it’s actually a good question, because I think there’s a lot of noise that comes from a lot of new manufacturing technologies about how hard it is to qualify. I’m actually a little bit less sympathetic than you’d think to this. I mostly think this is the case of that technology not actually performing. If you actually have the data to show that your process produces parts that are good enough, you probably don’t have a huge issue qualifying. The issue is when you fail spec and you need someone to change their design or redesign their system—and that’s where big problems show up.
And so for us, a big thing is, you know, we’re not doing level three. We’re not doing some new process that requires totally new design allowables and new designers to be trained up. Because the truth is, those adoption curves are extremely long. I mean, that’s what’s happening in metal 3D printing, right? You’ve just seen these companies that have amazing technology—I mean, I love laser powder bed infusion—but it’s not a technology that’s going to scale immediately to cover the whole industry, and it certainly isn’t going to scale to cover the parts that were designed in the ’70s. And you know, the designers are not only unwilling—they’re actually not even working anymore. They’re retired. You’re not going to change those parts.
And so the way that aerospace works, right, is there are these long, long campaigns to make sure that your part meets specifications. There’s this contract between you and the designer, and you just go through the test. If you pass the test, it’s objective. And so for us, the huge thing is always talking to designers to understand what they need before we’re even making parts, and making sure that we’re building a factory that goes in that direction. But the truth is, if you make a technology that has advantages—if you have cost advantages and can meet all the technical requirements—there aren’t actually so many issues. The issue is when you’re trying to do something new that pushes the designer in the wrong direction.
Akhil 46:00
And it’s not just the cost and technical requirements. You are fundamentally, with the way in which you orchestrated your digital range of your boundaries, able to make novel points.
Aeden 46:11
So that’s the biggest thing. The biggest thing is when someone designs a part, goes through a long simulation and testing campaign, and then says, “Okay, we need to make the first prototype,” and they can’t make it. This is a huge problem. It’s the exact same case as the designer having to redo all the work, change the design, reduce capability.
And so the greatest customers that we work with are the ones that went to somebody else who couldn’t make it, and now only have one way to make it—and we’re enabling something huge for them. Not only are we faster and cheaper, but we’re also enabling the missile to have 10% higher thrust, 10% higher range, and higher reliability. Maybe the cost is lower because they can do one part instead of five in some cases. Those are really the best cases for us. That isn’t everything, but when we can do that, it’s a huge benefit for everyone.
Akhil 46:53
Aeden, are you seeing—you mentioned their scalability, right? And the actual ability—and Cameron were discussing earlier Freedom’s Forge, right? For those, I think we’ve talked about it before. Awesome book, and a good complement, a good sister book, is Arsenal of Democracy. The benefit we had during that period was that there was this two- to three-year, if not longer—really ten-year—period before December 7, 1941, where we in the United States, because of what they were seeing geopolitically, were actually already putting in motion the movement toward scale, right? I mean, you saw Ford already start converting some of its plants toward B-24s; you saw the work on Liberty ships. Obviously, that was a different year, different era. But are we seeing—the question I have is—are we seeing on the end-customer side the actual drive and demand beyond low-rate initial production, beyond just prototyping, to actually drive the rest of the manufacturing enterprise upstream to facilitate that?
Cameron 48:01
No, I mean, we gave it all up. It’s not here. When you don’t make consumer goods here because you’ve hollowed out the consumer base, you don’t have that surge capacity anymore, right? Who’s gonna build this stuff? The factories—and the factories for those parts—aren’t here. Even if a Ford is assembled here, they’re getting a Bosch component from a casting house that’s often in Taiwan, or often China, or often Mexico. Once again, a lot of this stuff—they’re screwdriver factories. So when you shut off that spout, like in the case of, you know, if a balloon does go up—I don’t think if a balloon goes up, the faucet’s gonna probably be shut off—and they’re gonna be in the supply chains that exist under the all-American factories that exist today. So, you know, they’re making a car in Detroit, Michigan. That car is not being made with parts that are from Detroit, Michigan. It’s unfortunate, but it’s just the place that we’re in, and so we need to shift that back.
But I don’t blame them, because that part can’t be made on good margins at a good price with the weaponized subsidies that have been leveraged against us—and it really exists in the shadows because we still think that these cars are made in America, for the factories that haven’t moved yet, right? You know, most folks are just looking at that level, saying, “Okay, well, you know what, factories for car making are still in America.” But the ones that are here—Tesla is an outlier, for sure. Tesla makes, I think, something like 70% of their cars with American parts.
But, you know, the hard parts are really—you know, if you just go on a quick deep dive here—neodymium magnet production: almost all neodymium is processed through China, or through some stage of the neodymium process. The furnace step and magnetization step are done almost exclusively in China. There are a few neodymium magnets that are, you know, all through NATO and allied supply chains—and, you know, the prices are more than 15x what you’ll find in America.
But when you talk about these factories and the surge capacity, you aren’t able to say, “Hey, you’re making alternators for a Ford. Now go start making motor components for these new autonomous systems,” or even your old legacy systems. There is no supply chain shift. It just doesn’t exist.
Akhil 50:57
It’s compounding. It’s not only there. The other aspect—I can’t just turn a Model T plant or a Ford car plant into a B-24 production line. It’s just not—if you want American parts, indeed, right? I mean, we’re doing things that are arguably a little bit more complex, but your underlying point is, even regardless of that, we don’t have the capacity. We’ve hollowed out the capacity here to be able to do it all the way upstream.
Aeden 51:23
It’s also worth saying that modern defense systems, I think—the difference between a modern defense system and a car, and a World War II defense system and a car—I think the World War II B-24/Ford maybe a little closer than the modern systems. Like, modern defense is much closer to consumer electronics, right? It’s lots of complex electronics, lots of complex sensors and cameras, navigation, right? This is closer to an iPhone than a car, and we don’t make any iPhones here, right? Maybe we make some cars here, but the consumer electronics supply chain has really been built overseas, and it’s really coming to what the latest defense systems are as well.
Akhil 51:59
Yeah, so if you guys had to—you were czar for a day—and you had two to three things you could do to reshape or reformat the manufacturing base here, what would that sort of look like? And maybe some specific sub-questions there: would it be centralized, decentralized? Would you have edge nodes where it’s closest to the potential user or operational standpoint? Obviously, we’re talking about civilian and military buy capacity—that’s it. It doesn’t matter, whatever incentive structure works.
Cameron 52:34
Buy capacity. Buy as much capacity as possible, and have contractual obligations on that capacity to go make the things that might be needed in the case of, you know, a surge scenario. But there needs to be—like, at the end of the day, someone’s got to make the parts. And the way that parts get made is because you have machines and people that know how to use the machines, and one day they’re cranking out one thing, and the other day they’re cranking out another thing. And that doesn’t exist because of the economics right now. So we have to at least match the subsidies of our adversaries to ensure that we have that when it comes time.
But we have to start doing this now. The lead times on these factories are not a year—they’re not even two years. I mean, in some cases, our foundries are looking at procuring equipment that takes over two years from order to making the first parts. So this is a decision that you have to make even before you want the ability to use it or not. So you have to get on it quick.
But I think the most certain path to maintaining the capability that is required—or building the capability that is required—is to buy capacity. I think the DOD can do this through a series of preexisting mechanisms or through some new mechanisms. DPA Title III is a really great one with good history here, but it needs to be hyper-specific capabilities. So we’re choosing a process, we’re choosing a partner, we’re choosing an application that’s proven to make sure that the execution side of this really works, right?
This is one of the great things about a factory—you can’t really put makeup on a part. The part passes qualification, and you made it in X amount of time with Y amount of throughput—or not. You can buy a factory that makes 12,000 doohickeys of X type for Y platform. In the case that platform needs to be scaled to 12,000 units a month, you don’t have to make that right now. You can use that capacity to build something else with a similar set of requirements, and that will feed back into the American private economy. But you need to have the option to be able to flip that switch. And today, you don’t have that option because you don’t have that capacity.
Maggie 55:04
So I know you all hosted INDOPACOM Commander Admiral Paparro at the factory a couple of months ago. I’m curious—how is DoD thinking about a capability like Rangeview in a future conflict in the Indo-Pacific in particular? And how will a conflict like that need different kinds of manufacturing capabilities than, say, what’s happening in Ukraine or the Middle East or elsewhere?
Cameron 55:33
The Navy can’t get parts right now. I mean, a sub will come back, a ship will come back, and there will be no way to get a part back on that system except by cannibalizing another system. There are some crazy statistics out there for just how bad it’s gotten, but the easiest way to look at it is: for these new programs, there’s a lot of funding behind them—throwing a lot of money at trying to build submarines. And if you look at the difference between the shipbuilding capability that America has versus our adversaries, it’s astonishing.
And if that’s new stuff—if that’s where everyone’s focusing—imagine just how bad the sustainment and maintenance component is. That’s one of the biggest things that Paparro was interested in when he came by our foundry and saw our technology. I think the flexibility of Rangeview—offering a capability that can one day be making critical components for the private sector, then shift when a submarine comes in and needs a ball valve to be able to make that part on the spot—is a massive capability that no other casting technology currently offers the Navy.
Akhil 56:55
Yep, and there’s a litany of congressional reports, and obviously the last reconciliation bill reflected a lot of the need to sort of recatalyze that. Specifically, Cameron, to what you just talked about—if a commander like the Indo-Pacific Commander, or just the DoD in general, wants this flexibility, does it matter where it’s produced?
And what I’m getting at is: do we need some sort of edge manufacturing capacity? Does it matter where it’s put? And does it actually make sense? If we want to try and build that out—exactly—you’re not going to print drones on the battlefield. It’s just, you know, it’s actually more wasteful to send raw manufacturing goods to those printers. Manufacturing doesn’t have an above-100% yield; it has less than that. So there’s waste generated.
Cameron 57:42
Right. So it is more space-efficient and more material-efficient to actually send preprocessed goods to the end-use point than to ship the manufacturing plus the raw goods to that use point.
And, you know, I’ve heard a few arguments—“Well, what if the need changes? What if one day we need a drone that can fly 500 meters, and another day we need a drone that can fly 1,000 meters?” The work that goes into a drone design to achieve those two capabilities is not work that can be done on the battlefield. That’s work that’s done preprocess. They might load up some instruction file to the system, but the truth is that instruction file—whether it’s the printer’s job or the whole factory’s job—can be programmed somewhere else.
And you can, you know, send 500 drones of one type and 200 drones of another, and the difference in that flexibility is far more beneficial than if you instead had an edge-deployed system that touted bringing that flexibility out of you. I don’t think we’ve really seen it—look at Ukraine. Ukraine is not printing drones on the battlefield. Ukraine is shipping shipping containers filled with thousands of these systems, fairly—at least relatively—close by from where those production facilities might be.
Aeden 59:17
I mean, I think there’s a pretty good razor as to when some sort of edge manufacturing makes sense. And actually, there’s great precedent for this, right? We’ve had machine shops on aircraft carriers since they were new—since World War II.
The case that matters is if you have a very, very big, expensive system that can be taken down by one or two spare parts. Those cases are the rare ones when you want a little bit of edge manufacturing, because it will not be cheaper—it will be much more expensive. It will be much more painful to operate. The logistics of getting materials and maintenance to the actual edge manufacturing site are going to be ten times worse than they’d ever be on the mainland.
But it’s occasionally worth it if there’s a really, really big system that needs to be kept up. And so there are some narrow cases. Most of the time, though, you really want a nice, safe, big factory that doesn’t have to worry about operational security, doesn’t have to worry about being bombed, doesn’t have to worry about, you know, the humidity changing because they’re on a tropical island and there’s a rainstorm. That’s what you don’t want.
In general, you want a big workforce. You want to be able to replace people quickly—if someone gets sick, really obvious things. That’s how you get to efficiency, and that’ll be the bulk of all manufacturing forever. There are special cases, though, when for really big, expensive systems you need to keep them up, you need to turn them around quickly—and there, edge manufacturing is a great solution.
Maggie 1:00:33
What do you think is most misunderstood about manufacturing and reshoring from outsiders?
Aeden 1:00:40
I think there’s one thing to say, which is that people picture manufacturing as, you know, just another thing—like it’s just another industry. And I think when you actually start to get into what manufacturing is, manufacturing is building the physical world, right? That is what manufacturing is. The breadth of different kinds of processes, different kinds of things that are done, and the dependency between those—and how much, you know, having a mine in a particular place can affect the price of valves—is mind-boggling to a lot of people. And I think, you know, people forget, it’s not just one process. It’s not just like fixing magnesium supply in America does not fix the magnesium casting supply, nor does it fix the helicopter gearbox supply problem, nor does it have anything to do with the helicopter gearbox inspection problem, which are all distinct, huge problems with thousands of people that work on them in the U.S. And I think it’s often forgotten, just the scale of these things, because, you know, people don’t see factories all the time. They don’t walk down the street. They don’t have friends that work in factories, right? We don’t have this—especially if you go to the coastal cities—this is nowhere to be seen. And so I think people just lose some of the scale effects of what are all the things that are going on, and how they depend on each other, and how complicated it is to change something in manufacturing, right? If you want to change where something is manufactured, there’s a lot of different pieces to it. I think that’s a pretty big one.
Cameron 1:02:01
For some context, helicopter gearboxes are oftentimes magnesium. They’re a very big supply chain problem. So just to connect that a lot—like they’re all magnesium problems—but, you know, there’s probably a billion-dollar business in multiple of those magnesium problems.
Maggie 1:02:20
What advice do you guys have for other founders looking to build either in this manufacturing or national security, defense industrial base space?
Aeden 1:02:30
Never build a cost model. Just do it on vibes.
Cameron 1:02:40
Build a cost model day one. Update it every time anything changes. We’ve already created so many competitors. Like, literally within two days—but, like, literally, we were ready. There’ve only been like two founders that just 100% exist because we existed, right? So, I mean, even if we say build a cost model, don’t go build robots day one—that’s a good one. We’ve really spent hundreds of thousands of dollars on that lesson, right? Like, that’s actually a pretty big lesson. Purchase Rangeview’s system.
Aeden 1:03:12
Don’t build software. Build an actual company that makes parts. You know, build a factory. Those are up there.
Cameron 1:03:19
Yeah. So for other startups—focus on how many parts you make a day. That’s the metric of success that matters. We need parts, not tools for someone else to maybe use. Go make parts. Make parts yourself. Go make a factory. Don’t build software for a factory that may want to use yours. Factories don’t want to use software. This is why we’re in this space. Go build a factory and make lots of parts, and make your internal success metric proportional to how many good parts you’re shipping a day.
Akhil 1:03:48
Who are you looking forward to come join Rangeview?
Cameron 1:03:52
Oh, anybody who’s tired. Anybody who’s tired of, you know, working on stuff that might not matter as much. And not to say—it’s important that the work you do matters. I think you’ve got to wake up and look around and say, beyond the money, what are you doing? Your career is going to take up the vast majority of your life, right? What are you doing with that career? What are you doing with the world? And if you’re an engineer—a really principled, strong mechanical engineer, software engineer—if you love to learn about castings, please, I didn’t want to plug it here, but our website has a careers link. You click on that—we’re expanding our team like crazy. If you want to build custom vacuum induction melters, if you want to come work on the next generation of 3D printers, you know, if you’re excellent at what you do, I would love to speak with you, and we can build this cyber foundry together.
Akhil 1:04:58
Awesome, Cameron. Yeah, where’s the company going to be in a couple of years? What’s the vision?
Cameron 1:05:05
Yeah, I mean, well, we need another Westinghouse, don’t we? You know, we need someone that’s making the world’s stuff. And I think that’s going to end up with folks that understand the new paradigm. I think that’s folks that understand modern technology as it relates to manufacturing and folks that understand automation. I think we’ve got a pretty amazing team here, and I think everyone here is doing it for the right reason. So I’d hope to be making a lot of parts.
Akhil 1:05:33
Awesome, awesome. Well, Cameron, Aeden, this has been awesome. Thanks so much for taking the time. Excited for what you’ve already built, what you’re unlocking both in what can be built going forward, what is impossible, and what can be scaled going forward. Excited to be part of the journey with you.
