BH: Tell me just a little bit about Digital Metal. You started off at Siemens. What led you to the concept?

CK: Yeah, yeah. I worked at Siemens Energy for five years. I started my career out on the gas turbine side. It’s a hot space right now, but back then nobody cared about gas turbines at all. When I told friends from college where I was going after school, they were like, “Why would you want to work in a dying industry?” It definitely wasn’t a popular career path.

I was doing competitive market intelligence and supporting outages and then eventually transitioned into renewables, which was the hotter thing at that time. Some of the fastest growing jobs in the US were there and then that all changed with the most recent administration.

But while I was there, I learned a lot about the energy markets and the types of products we have to build in order to produce energy that you, I and everyone in the US needs. And while I was there, there was a particular problem that stood out to me on the turbine side. With large gas turbines – industrial turbines – you have these things called outages where you take the turbine totally offline for scheduled maintenance. Every couple years, you’ve got to take the whole unit down to do maintenance because these things are continuously running all the time. They run 24/7 non-stop to produce the power that we all use.

At one point, I was working on a team that was supporting an outage. And the thing with outages is that you really don’t want them to run over their scheduled time because you pay the customer back for all the lost revenue missed out on during the turbine's extended downtime. Our outage ran over by two weeks, so Siemens ended up paying some $1 million in liquidated damages back to the customer. And it was all because they were missing a couple of parts, primarily castings. That was my first exposure to the kind of fragility of the supply chain around castings – especially super technical castings and low volume support.

This led me down this long-term rabbit hole to discover why those castings take so long to make. That was the original inspiration for Digital Metal.

Fast forward and I was working on the renewable energy side of Siemens working on wind turbine sales, supporting $150+ million in contracts across some of the largest developers and selling the energy. The exact same problems happened on that side of the business too. Wind turbines are a lot less sophisticated that gas turbines, but part delays for certain geometries – especially castings – could cause a stacking effect that delayed project development and deploying, and ultimately led to these huge cost overruns.

All this experience kept reinforcing the need for faster supply chains around critical component manufacturing, of which castings are a big part.

Then I started working on a side project developing a consumer kitchen appliance. During that build was when I got exposed to these quick turn manufacturers like SendCutSend, OshCut and Protolabs because I was getting all these little parts made to build this product. I’m really grateful that they existed because it allowed me to make a product that otherwise would’ve cost me thousands and thousands of dollars. I could do a single prototype in my apartment.

At one point, I wanted to get the outer shell cast but couldn’t figure out a way to get one made in both low volume and low cost. For a guy who was making <$60,000 a year, this project had to be affordable. That brought me back to Siemens. I was using all these quick turn manufacturing services and knew there was demand for low volume castings because a lack of them was costing companies like Siemens millions of dollars in delays.

I wasn’t planning on leaving Siemens, but I had just moved to San Francisco and was surrounded by startup founders. After a while, I decided to take the leap, quit my job and start exploring a concept like SendCutSend or OshCut for castings.

BH: Talk to me about where castings fit into the broader ecosystem. You have platforms like SendCutSend, OshCut, RMFG. Where does Digital Metal sit in comparison? And when you zoom all the way out, who are you building for? Is it the guy in the garage? Is it Siemens? Is it some mi of both?

CK: Castings are really cool because it’s the most economical way to make hard-to-machine parts, like bulky 3D components, and they’re everywhere. They’re in planes, trains, cars, energy infrastructure, furniture, and a whole lot more. It’s the most economical way to shape metal. But there’s a huge problem in that it’s really hard to make it economical in low quantities. It’s very much intended to be a high volume manufacturing method, like with car parts. In a car >50% of the parts by mass are cast metal. They can’t be sheet metal because of strength and wear resistance reasons.

The market is huge and broad. We want to stay agnostic and be for everyone, which is something we’ve made possible through instant quoting and different online portals.

Through it all, we’ve brought the cost way down for single unit runs. Traditionally, you have to cut a tool (an aluminum block mold that you inject with wax to create a pattern). What we’ve done is come up with a process that completely eliminates fixed tooling, which is a $20,000-$40,000+ startup cost. It allows for both easy access to manufacturing and the ability to do super low volume runs at a super economical level. Those two are where we’ve really focused because it hasn’t been possible before. It’s been possible to go to your local sheet metal ship and get sheet metal parts. Same with machined parts. But never cast parts.

A cool thing we’ve been seeing is that customers will prototype with our technology because it’s actually a lot cheaper than other manufacturing methods for creating metal 3D parts. Metal casting is one of the only (if not the only) component manufacturing processes that doesn’t scale in cost with complexity. It’s based solely on the amount of material use. You can cast a super simple solid block of steel for exactly the same cost as a super complex steel component as long as it uses the same amount of material. So we see a lot of early teams start with us instead of going direct to machining.

I think this world of specialty components that we fit into is really interesting. The thing I’m thinking a lot about right now is robotics. There’s a ton of investment happening there and a really interesting component development akin to what happened in automotive. If you look at cars, there’s almost no machined parts on them anymore. If you look at robotics, you see a ton of machined parts on that thing. The development path forward for robotics backs the theory that the TAM for robotics is going to be bigger than the TAM of all automotive, which gets floated around a lot. Over time, robotics companies are going to realize the same thing that automotive companies did: that machining parts doesn’t scale to the quantities of production that they need to produce.

We fit in there really nicely then, allowing robotics companies to prototype with what will eventually be their final production components. And you can scale this line of thinking out to any industry that needs to replace machined parts.

BH: Scaling something like this has to be tough. You’re the heavy asset tooling guy in the supply chain.

CK: Our particular manufacturing method is surprisingly really, really difficult. The reason comes down to the massive equipment requirement.

Starting something like a CNC machining service is relatively easy. There are like 10 machine tool vendors that you can go to to buy a machine and they’ll deliver it next week. Buy some raw material and you can start cutting parts. That’s all you need really.

If you want to be a sheet metal manufacturer, you can do something similar. Go out and buy a laser or plasma cutter, cut parts, bend them with a cheap brake press and then you’re off.

For us, we have to buy 12 different pieces of equipment that are either all very unique, built custom for our manufacturing facility or self built. Then you line them up in a very specific process in order to make it all work. I can’t just put a credit card down and buy a couple of machines. But that’s where the opportunity is. It’s so much harder to replicate what we’re doing not just because of the casting technology we’re developing, but because you can’t compete with us right away with a credit card.

BH: As you scale, do you see the company moving into the “enterprise” tier of customers? You talk a lot about driving down the startup costs for early teams, but this has to be applicable for bigger markets and bigger, Siemens-esque customers too.

CK: I’ve always thought about supporting that group from the beginning. It’s where I came from and where the idea for Digital Metal originated. Outside of larger prototype runs, we see ourselves fitting really well into that critical spare parts manufacturing niche.

Say you have a gas turbine down and you're missing two pump valves. You go to the foundry that originally made them. Turns out that foundry closed years ago or the tooling was lost (because that’s happening a lot right now). We can quickly make that spare part for you based on a CAD file if you have it.

This is actually happening most in the military right now. Something like 50% of the US military’s vehicles are currently in need of parts or are waiting for parts because of original documentation loss. It’s a difficult spot to be in because the tooling and supply chains for those parts are really desperate for a solution. That’s a perfect opportunity for us, and it’s generalizable across huge markets like automotive, aerospace and energy.

BH: There’s obviously been a lot of attention around reindustrialization lately. SendCutSend is officially a unicorn now. When you look across everything that’s happening, what excites you the most?

CK: I’m so, so excited to be in this space at this moment. I feel like I got a bit lucky. Looking back, I would never have expected that a company that I was using in my apartment five years ago as a hobbyist was going to become a unicorn. And now all the others are growing a ton too. It’s amazing, and I think more people are now waking up to the success you can have with a business model around better manufacturing interfacing.

Manufacturing is horrible. If you’ve ever sat down and tried to order parts at a large company, you’ll know what it’s like to sit through six different phone calls before you can even exchange funds because there’s so many review and general alignment calls. It’s a waste of everybody’s time. We’re just starting to think about it differently here with new quick turn manufacturers like SendCutSend, but China figured it out a long time ago.

BH: There’s a lot left for us to reach parity with China. What do we need to be building more of inside the supply chain to get there?

CK: A lot of it runs through interfacing with the manufacturer and driving costs down on the factory floor.

Interacting with Chinese factories is interesting. In the US, we’ve started using software as our interface. In China, they just use WhatsApp or WeChat and you’re linked directly to the guy on the factory floor. And honestly, they’re really good at just intuiting the best way to do something rather than jumping on calls to ask questions. In the US, we decided to build software to intuit processes for us. And this comes around to costs too. A lot of what you see in traditional on-shore manufacturing is bloated administrative overhead. We have this horrible habit of building front offices full of sales, operations, finance and HR people who never hit the factory floor when we really just need to be building production facilities.

China’s at an unfair advantage because they’re labor rates are so low. They can build out a front office because the economics are so different. If we want to compete in our higher labor cost environment, we have to compress the front office as much as possible as quickly as possible, everywhere. And all of that can be done through software and automation.

My take here is that robotics is actually a red herring. I think we’ll spend more time and money setting up automation programs than we would just hiring two guys to work the floor at $30-$40/hr. If you’re doing the same production run of the same part hundreds and hundreds of times, automate that. But for high-mix quick turn, robots just aren’t there yet. We use them and they’ve just been more a pain than anything. Since parts and trees differ so much, it’s easier to just put a human in the loop instead. It doesn’t really matter if you’re doing enough volume of high mix. I think robots are part of the answer eventually, but the best thing we can do immediately is aggregate scale and compress overhead as much as humanly possible. And no need for a software engineer for that anymore. That role’s shifted to the founder. You can sit in front of Claude Code for 10 hours and probably get a workable solution in a day.

It’s a super exciting time, and I think more manufacturing founders will start doing more of that. They’ll stop relying on the 30-year old veteran for quoting or manufacturing processes and they’ll just start coming up with unique algorithms for it all. And the whole industry is better for it.

BH: There’s something really unique with manufacturing here. It’s not zero sum. In software, it’s all zero sum. You’re competing directly with two or three other platforms for the same group of customers. But for some reason, there’s a massive cultural difference here where people like you are genuinely excited to see more people get into the space and innovate.

CK: Absolutely. I can’t remember the exact statistic, but when a new manufacturing facility opens in the US, it actually has a multiplicative effect on the job and wealth creation in that area. Most businesses when they open will have an additive effect thanks to added product and jobs. But with a manufacturing business, you’re not only adding product and jobs but also creating new demand for every layer that goes into that manufacturing business. You need raw goods, machines, finishing and a whole lot more.

BH: It’s almost like when a manufacturer comes to town, you have to build an entire supply chain around it which sets off more supply chains around those supply chains and opens the ecosystem up to an entirely new group of businesses.

CK: We’re in Austin building our first factory, so we’re excited to see how it unfolds there.

BH: What’s it like building a factory?

CK: I’ll have to write something because I have a ton of thoughts here. *

Building a facility is managing a thousand different things all at the same time. You’re balancing electricians, plumbers, HVAC guys. You’re interfacing with the landlord. You’re working with the city. You’re coordinating with equipment dealers and manufacturers. You’re not just putting a building up. You’re building your supply chain for it at the same time so it can be ready on opening day. Part of it might be a function of our size. We’re just a team of 3, so there’s a ton of work distributed amongst just a few of us.

There are a few things that I would tell a lot of founders to do. You have to gamble a little bit. My investors probably won’t like me saying this, but we ordered equipment before we signed a lease that we originally had no spot for in the factory. We have 16-18 week lead times for some of our equipment, so we had to get those rolling. We did it before we had a lease and before we finished raising money to fund it. And I’m not the first to say this. Jim Belosic at SendCutSend preaches the same thing.

BH: Yeah, the guy plans and sets entire factories into motion before even scouting space. It’s crazy.

CK: The other thing I’d suggest is staying organized and overcommunicating everything. You’re generally dealing with a lot of people tackling different components of the factory build, and a lot of times they’re going off and doing their own thing. It’s really incumbent on you – especially if they’re suppliers or manufacturers – not to constantly ride people. Riding people doesn’t make their lives easier. I’m going to stay on top of people, but I’m also going to look at alternative options at the same time in case we need a backup. I’m constantly calling around to equipment dealers for used equipment from closing foundries knowing that we probably won’t buy it, but because it’s good to have something in your back pocket in case things go sideways.

It’s maybe not great advice, but you need to be an octopus. You need to have your hands around everything and have to go into it knowing that you’re not going to have a social life for 6-8 weeks.

And then it’s funny because once you have a handle of everything on the setup side, then you start having integration problems. We’re going to have process problems and bottlenecks that we’ll need to address, but I look forward to it all because it feels like we’ve worked through a lot of the hardest obstacles already.

* Connor did write an article on factories following our conversation. Read it here.