Today, we have David Hartmann, Founder and CEO of Helio Additive. Helio Additive has revolutionized 3D printing through the creation of simulation software tools deeply rooted in the principles of physics. These tools have had a profound impact on enhancing production efficiency in the realm of 3D printing. 

During our discussion, we delved into the intricate challenges associated with transitioning a 3D product into full-scale production. These challenges encompass a substantial engineering endeavor, multiple rounds of test runs, and a notable scrap rate. Furthermore, we explored the remarkable opportunities that lie ahead in production optimization by creating models that encapsulate the intricate interplay between machinery, materials, and design files.

Key Discussion Points:

• What factors are hindering additive manufacturing infrastructure for both large-scale and local applications?
• Who is responsible for integrating your software into the system? Is it the OEM or asset owner?
• What are the primary core features of your 3D printing and modeling technology?

Erik: David, thanks for joining me on the podcast today.

David: Yeah, really nice to be here. Thanks, Erik.

Erik: Well, I'm really looking forward to this conversation. Because this is one of the rare opportunities where I get to speak to the founder of a company, where I knew you before you founded the company. I got to be a little bit part or, at least, sitting on the shoulder a little bit watching as you got this going. So really looking forward to this. I think a good place to start then, David, would be to introduce a little bit of your background at Covestro. Because I feel like that's where maybe the seeds of this idea were planted. Then to understand, you were at Covestro for 16 years. You were SVP of Growth Ventures, Co-CEO of Composites. So you had a very successful career there. So what was it that led you to jump into the startup ocean?

David: Well, these are not easy. You started right with the challenging questions, right, Erik? For me, I spent, like you said, 16 years with Covestro. When I joined in 2004, I thought I'd stay for one or two years. Then I ended up staying for one and a half decades, because it was so exciting to be part of what was going on in manufacturing — in Europe, in India, and then particularly in China — and see from our customers how manufacturing was changing and how a lot of manufacturers, especially in the last few years, were going over this chasm of what was the past and what is now the new normal in manufacturing and supply chain, and trying to navigate how they deal with it. That was really, really interesting for me.

In 2014, I took a year of sabbatical effectively to look into that topic in more detail, in particular when it came to digital manufacturing and additive manufacturing. During that year, year and a half, lots of things happen. But I think for me, it triggered an awareness that the manufacturing landscape that we have today is not one that is sustainable and that, somehow. we have to change globally how we're doing manufacturing. I don't want to get too much into the story here. But maybe just to take it that one level further in terms of motivations, in 2017-2018, I went on a self-discovery journey as part of a development program that I was lucky enough to be a part of in Scandinavia called Stifinder. Out of that, I came with a much clearer idea of what I wanted to follow in terms of values and what I wanted to follow in terms of what I wanted to do in my life. One of those things was to see how we could make communities way more resilient given the kind of climate change challenges that they'd face in the future.

I don't know whether you've been following the news in New Zealand the last six months. But Q1 of this year, we had huge storm fronts coming in. Road infrastructure was washed away. And now there's a debate. Can that road infrastructure really be meaningfully rebuilt? Because it's been washed away twice in the last six months. Is there enough money left? Does it even make sense in a changing climate environment to rebuild that infrastructure? The questions for those communities that are isolated are, "Well, how do I get goods coming in? How do I still integrate with a global supply chain?" These were the kinds of things that were motivating me as I started to look around and see, well, what could we do there and how could digital manufacturing help?

Erik: Okay. Interesting. Yeah, I hadn't seen that news about New Zealand. Of course, we're right on the tail end of this flooding in Libya, which is terrible. Something like 10,000 people lost. Certainly, around the world, I think maybe it's, first, the insurance companies, right? Because they actually run the numbers and start looking at this. Then the knowledge is creeping into a lot of other parts of society, that the way we've been operating has to change because of the world changing around us.

I guess, there's maybe two aspects of this. So you mentioned that the way we're doing manufacturing is not sustainable. I think it would be useful to understand the problem before we jump into your solution. So what is it about that? I could imagine they're being two parts. One would be the inefficiencies of the existing processes, waste in that process. Then the second which you just mentioned around New Zealand, which is maybe a bit more of a local problem, is localizing capability. If you don't have a bridge to get logistics into a town, how do you have products in that town? Which is a bit more of a local problem but a problem that a lot of local regions around the world have outside of China. But how do you see that problem space?

David: I think you've kind of bracketed that really well. There's a whole spectrum in between. There's a team at TU Delft that published a study a few years ago that said if we adopted additive manufacturing in a wide scale, we could save 27% of the world's energy consumption. They had a range of scenarios. So that was probably the most optimistic scenario. That scenario comes from the fact that we're just shipping tremendous amounts of goods from place to place, and we're producing incredible spare parts inventories that may or may not be useful. All of that has embodied energy in some way. Their embodied in the logistics programs or embodied in the spare parts. These are areas where we could work to mitigate our negative impact on climate change.

Then there's all the assimilation stuff. I think Kim Stanley Robinson wrote about this. I think with the heat waves, we're not far away from that. Being in a world where if there was the right combination of weather and humidity and temperature, parts of the world will no longer be inhabitable for weeks at a time, whether that's places in Texas, or in India, or in Bangladesh, or wherever. When that hits, you really want your infrastructure to be working. Because if the air conditioning or the power fails for any number of reasons, and you're missing a spare part, you don't have the luxury to order that spare part on Taobao and wait for it to arrive. You want to have a source for that critical infrastructure locally. That's where local manufacturing is just going to be absolutely critical. I think those are the two book ends with a whole lot of scenarios in between. Those are definitely two areas that we're hoping to address with the widespread rollout of additive manufacturing and digital manufacturing.

Erik: Got you. Okay. And so if I'm thinking through what this might look like in terms of the infrastructure of the future, it sounds like we have to solve two problems. On the one hand, we need a really efficient additive manufacturing infrastructure that can maybe do things at a high volume but still small batch sizes, and with a very high efficiency level. So that would be maybe covering this issue of waste in the manufacturing supply chain. But then on the other hand, we also need to have relatively efficient local solutions that can serve small communities. That maybe won't be the same efficiency level as a large plant, but it can solve that problem of what happens when a component breaks, and you need something ASAP to get it moving.

Why are we not there today? Of course, technology takes time to develop. But it feels like with additive manufacturing, we're just scratching the surface of the potential today. So what is it that's preventing us from getting to the efficiencies where companies are really ramping up adoption of these solutions?

David: In a nutshell, its price performance and the drivers that need to be there to drive price performance up. When I say price performance, I'm talking about everything that goes into the cost per, let's say, part — manufactured part.

Part of that is also support from government and government regulation. In Europe, we're seeing a massive drive to right of repair and giving consumers the ability to not throw away their appliances, for instance, but rather be able to get them repaired. One of the laws that's coming into place in the next — it has already partially been implemented, will be implemented fully in the next years is that appliance companies like Fisher & Paykel, or Miele, or Haier will have to keep spare parts inventories till five to seven years for appliances that are no longer being sold. If you imagine what that would look like in terms of spare parts, it would just be economically unfeasible. So the price performance doesn't actually have to compare to injection molded parts. What the price performance of additive manufacturing in that scenario needs to do is just compare to the cost of holding that spare part and the inventory costs, the supply chain costs, the embodied energy costs, etcetera, etcetera.

In terms of what is needed now to get there, somehow, we have to unlock a massive gap in potential between what today's hardware and materials can potentially do versus what they can actually do today. In our opinion, the only way to do that is to do that with software, to do that with software that really understands the physics behind the manufacturing process, and then is able to optimize all the hundreds of thousands of decisions that need to be made in order to make that part.

Erik: Okay. So if we think about what it takes to do additive manufacturing, you have the hardware. You have the equipment. You have the materials, the polymer, for example. And of course, there's a lot of companies that are continuing to improve those two parts of the equation. But then, you're saying there's this interaction between them which is what the software manages, and that there's a big potential for improvement there. Where's the complexity here? So we have a hardware. If I'm just looking at this from a layman's perspective, I'd say, okay, you got a hardware. You got a software. You put the material in, and you print. You have a CAD software, some kind of specialized CAD that tells you what to print. Where's the complexity? How is the output of that not optimal with the existing solutions?

David: You have three sources of complexity. One is the material. In our case, we're talking largely polymers. Although, we can do metal as well. But polymers are extremely complicated beasts. Thermoplastic polymers are just a real challenge to predict what they are going to do and how they're going to form. We have 60 years of experience doing that with injection molding worldwide or even more, 80 years. And yet, it still continues to surprise us, even injection molding.

3D printing, you have the additional or additive manufacturing. You have the additional challenges of very complex manufacturing process. Whether you're talking about powder bed fusion or some kind of material extrusion-based process, you have effectively a sequential manufacturing process that happens over time that is much, much more complex than injection molding or thermoforming.

Then finally, you have the third source of complexity which is that you can literally have any geometry or any shape. If you somehow restricted the geometry, then what would be the point of additive manufacturing? Once you have that third complexity, you're multiplying three sources of complexity. To make that work requires a deep understanding of the physics behind both the process and the material.

Erik: Got you. Okay. Great. That's a great foundation in terms of the problem space. Let's look at what you and Helio Additive are working on now. So is it my understanding then that you're helping companies to solve the physics challenges of optimizing how these three elements will interact with each other?

David: Yeah, our foundational insight three years ago was that in the additive manufacturing process, you're making hundreds of thousands of implicit decisions — speed, temperature, tool pathing decisions, et cetera, et cetera. Human beings can get part of the way there with trial and error. But it's really not something where trial and error can help you solve get the maximum performance out of it.

When we calculated the price performance gap between what would theoretically be possible with today's materials and hardware versus what is currently possible, it was about 10x. So it is a significant price performance gap. Our hypothesis was, the way to close that gap would be to be able to simulate what was going on. Make some conclusions in terms of what that meant in terms of defects and so on, and then be able to go back and optimize parameters and close that loop so to speak.

Erik: Interesting. So you're saying that today if I want to print a part, basically I have my CAD design, some kind of design. I try to optimize it. But it's kind of humans trying to do something, so it's never going to be optimal. Then I print it, and I see what the faults were. I try to modify that and go through that trial-and-error process until I print it accurately. So the existing software solutions aren't able to tell you that there's going to be a weak point here, or that you're going to have excess material here, or these types of issues.

David: No, nobody is able to tell you that. It's a little bit like you're flying a plane, but you've got no instruments. It's cloudy outside, and you can't you see your horizon. You don't know how fast you're going. You don't know what your goal speed is, et cetera, et cetera. Today's 3D additive manufacturing engineers are really pretty amazing because they're doing all this trial and error. They've got heuristics. They've got experience. They tried to dial in. Of course, they get advice from the companies that make these printers and materials, and then they try to dial in what they think is the best set of parameters. Very often, they make it work. But I compare it to — you've got this amazing sports car, and you're driving through the woods at night. You've got no dashboard. You've got no headlights. You're not going to be driving that sports car very fast, right? You're going to be driving really, really cautiously and slowly. What we're doing is, we're giving the dashboard, the headlights. Eventually, we're delivering an autopilot.

Erik: Okay. Interesting. It seems like software is an excellent fit. Or, let's say, machine learning plus software is an excellent fit for this problem. Because fundamentally, it should be a highly-controlled problem, right? You have a machine that can make very precise decisions or movements, and so it should be something that could be more or less completely automated. But of course, yeah, it sounds like from what you're telling me, we're not quite there today. Let's talk about what you're building then and how you get there. Before we get there, I want to just ask you to help me understand the concept of voxel. Because I've seen this on your website a couple of times. It's understanding 3D printing voxel by voxel. So what are we talking about here?

David: You could divide up just like a pixel. It's a little square. If you get a JPEG or any kind of image file, they're made up of pixels which is a two-dimensional square that represents a small part of that image. A voxel is just the three-dimensional version of that. In additive manufacturing, it's a really important concept because, actually, that's how additive manufacturing works. Everything is built up voxel by voxel, tiny cuboid by tiny cuboid. Whether you're doing material extrusion or powder bed fusion, roughly the same. You could extrapolate in that way. So, yeah, that's a voxel. Imagine a 3D object and then divide it up by tiny, tiny, tiny little cuboids. Each of those cuboids is a voxel.

Erik: Okay. Thanks. Clear. On your website, what is the value that you bring in to customers? You have it broken down into transparency, flexibility, reliability, and performance. Help us understand. How are you fitting into the tech stack? Maybe first, who's tech stack are you fitting into? Is it going to be the equipment OEM that's going to be integrating your software into theirs, or is it going to be the asset owner who's maybe running a plant who is going to be buying your software off the shelf and then using an API to integrate? How do you fit into that landscape? Then what does your software do within that tech stack?

David: We're an enabling technology. As an enabling technology, you can think of us as an engine that could power many things, kind of like Intel Inside. Those many things can look very differently. There's, in additive manufacturing, a lot of different regimes and a lot of different industries. When I talk about regimes, I'm talking about, for instance, there is desktop industrial where you might have an industrial print farm turning out, applying spare parts. Each of those 3D printers is effectively a small refrigerator-sized cube, and they sit in stacks. People tend to them and get those parts out. Then another regime might be big area additive manufacturing or a large scale granule-based additive manufacturing, where you're manufacturing, let's say, wind blade molds — huge parts out of polymer which are later going to be used for composite molding and composite tooling. So you've got a really big spectrum of technology regimes. Then you've got different applications, whether it's applying spare parts, automotive, consumer goods, tooling for construction or tooling for aerospace. We kind of fit as an enabling technology in all of these spaces.

What we will target, at Formnext this year, we will launch our first product. It's going to be Dragon for large scale granule-based additive manufacturing. For these companies that are operating huge 3D printers, they run it like 30 to 60 kilograms per hour size flow rates of material. They're printing wind blade molds, aerospace molds, construction molds, some furniture, bridges, end-use parts all out of polymers. We're targeting them with a solution that will dramatically reduce the amount of waste and reduce their operating costs. We just got a case study back where we reduced for one large scale granule-based additive manufacturing customer. We reduced for one part, we reduced their engineering time needed by 160 hours. So that was 160 hours of labor from their engineers that they saved and reduced their material use by more than 330 kilos. For one part, we had cost savings of over USD $10,000. And if you think about that adding up on that one printer to 200, 300 parts a year, that's an incredible cost saving for those operators who are depending on efficiency for their business model, right?

Erik: Yeah. Okay, interesting. Why choose this use case? What is it about this particular use case of the larger systems that makes it, let's say, your starting point for launching this product?

David: First of all, the need is big. It's a segment that is growing extremely fast. It is a segment where the scrap rates and the trial-and-error portion is very, very significant. We're talking about a lot of material, a lot of high temperature material at quite high prices. And so this is an area where we can have immediate significant impact. So for us, it's a pretty clear story that we can tell. Plus, these are industrial 3D printers that are in operation today.

We also work, for instance, with brand owners who are developing micro factories and for, let's say, desktop industrial-sized applications. But these are then a little bit further out. And we'll get to those in the future. But this is an immediate, significant need that we can help solve now.

Erik: Got it. It makes sense. Okay. So back to my earlier question, how do you fit into that? Are you being integrated directly by the OEMs, the equipment OEMs, or is it the manufacturer that's purchasing your equipment and then integrate it into their production lines in this case?

David: We do talk and work with printer manufacturers. We also talk to and work with material manufacturers. Both are, after we launch product, direct customers for us. But I think the most exciting opportunities for us are the people who are operating three or four of these large printers. They're quite local, so they're supplying a specific local industry — whether it's construction, or 3D printing bolt holes, or doing tooling for aerospace. We can go directly to them and integrate into their current software tool chain, into how they're doing tool pathing, et cetera, et cetera. Then they directly feel the benefit. I think for printer manufacturers, they also get benefit. But the main benefit is for their customers. So they don't feel the benefit directly themselves. For us, it's more interesting to go directly to the users of the large scale granule-based additive manufacturing, because those are the guys who are feeling the pain today.

Erik: It makes sense. In this case, saving something like $200,000 a year per machine. That $200,000 has been saved by the manufacturer, so it would make sense that that's where the business case is. I imagine, they already have whatever design tool that they're using. Is this then somehow integrated into that design tool so that it helps to automate certain processes of that? Or is it a separate user interface that might be doing some of the work that they were using this other tool for previously? What will be the structure of the software from a user perspective?

David: We have custom integration with whatever the kind of workflow that they're currently using is, and that varies from company to company. Some people are using very standard kind of technologies, let's say, from Siemens or from other manufacturers. Some have cobbled together their own solutions using Rhino and Grasshopper and things like this. In any cases, we can connect to that. So everything is fully in the cloud for us anyway. We either provide our own dashboard which they can use. Or if they prefer, they can use the API, and we can help them integrate directly and then have it as a no-touch link between their system and our system. With these kinds of initial customers, support is everything. For us, the focus is on making sure that there's always support there to do integration, to do material calibration, to make sure that the numbers we're giving them back are spot on.

Erik: Got you. Then the core part of your technology would be the physics engine. Basically, you give it a query of how is this going to behave or what's going to be the outcome. Then it gives you a response. Is that the core benefit? What else would be, let's say, the key value-add in your tech stack?

David: In terms of internally what we do, we kind of do three things. One is, we do thermal process simulation, first principle physics base. Of course, we do use big data tools, statistical tools, machine learning tools. But the core of our physics engine is first principles based, which means we don't need to rely on big data sources that may or may not be high quality. We go directly back to what the laws of physics tell us.

Then for the second pillar of what we do, we build our own physics models. Again, first principles based. That's like the Rosetta Stone. So once you have the thermal history or the thermal simulation, the thermal process simulation, that's great. It's got everything that you need in it, but it doesn't mean anything on its own. It needs a key to unlock it. The physics models are the key to unlock that and say, well, what does this thermal data mean? What does it mean in terms of layer bonding? What does this mean in terms of stress relaxation, et cetera, et cetera?

Then finally, we have optimization tools, which then allow us to close the loop and go back to the original parameter set and then say, okay, well, what if we did this? Would we get a better result or not? What if we did this? Would we get a better result or not? What we can do for these large-scale granule-based additive manufacturing customers is give them a really broad scenario set of extrusion parameters and tool pathing parameters where they can then go, okay, well, I'm prioritizing this versus this. And so this set of parameters, this scenario of parameters makes the most sense for me given what I want to achieve. Does that make sense, Erik? Is that pretty clear?

Erik: I'll pretend it's clear. No, it's pretty clear. It sounds like, on the one hand, you're helping people understand how the physics will behave. And on the other hand, you're helping them to decide what they want to optimize. Is it minimizing scrap? Is it reducing engineering hours? Is it maximizing maybe throughput? Or I guess there could be, yeah.

David: Is it speed?

Erik: Okay. If you look at what you're doing today — I mean, you're still a relatively young company. If you look at what you're doing today and what you think you'll be doing in, let's say, three years, as the industry continues to mature and then your company matures along with it, what do you think the product will end up looking? Will you end up, do you think, moving on a road towards being your own specialized CAD type software where somebody's just directly buying your software, doing their design on it from the starting point? Instead of somebody spending a lot of hours designing something, I can almost imagine some kind of GPT engine where somebody communicates into your platform and says, "I want a panel that has these properties and fits into this space," and then your engine would spit something out for them. So if you think forward in several years, what do you imagine the business could look like? I guess there's different iterations there.

David: Yeah, maybe I'll just put a very short little explainer in between answering that question which might be quite useful for dealing with this. So when you're thinking about this process of going from design to manufacturing, you have two elements to it. One is the design of the part, and the other is the design of the process. We are focused on the design of the process. Now, it's perfectly true that when you focus on the design of the process, you directly impact the quality of the part that comes out. If we can optimize layer bonding, then obviously the mechanical strength of the part is going to be better. Then the part is going to be better.

But just to be clear, we're not CAD/CAM tool which would help designers design a part. We are more at the back end behind tool pathing and process optimization and process design that helps them turn that part into a defect-free reality. It's a gray area, but I think it's an important distinction to make — design of the part versus design of the process.

Erik: Yeah, thanks. That actually is very useful. That helps to design the scope here. So then if I reiterate that question, if you think forward, three, four or five years into the future, is your path going to be continuing to improve the algorithm? I mean, that will certainly be part of it. To what extent do you see other applications growing out of this, or is this evolving into some other type of application as the technology matures?

David: We agreed pretty early on that everything we did would be focused on two areas. One, we focused on revolutionary software tools. And two, that we're based on first principle physics. 10 years from now, that's what Helio is going to be doing: combining these two things in some way in order to drive the adoption of digital manufacturing and by bringing costs down. I think the core engine that we have now is — and maybe that's the interesting thing about deep tech, right? We get taught so often lean methodology, MVP, all of this stuff, which I think is great and has its place. But we're not designing a dating app where you can build a user interface, push it out there, and see what comes back.

When we started this, it was clear that what we were doing required fundamental science work before we could get to engineering. I think when you do deep tech, you don't start with a product in mind. You start with a deep understanding that there is an industry gap that is somehow holding the industry back, and you understand the technological problem which is why that manifest. Then you spend time solving that technological problem. But when you've solved that technological problem, you don't necessarily come out with a product. You come out with a technology that can solve that problem. How that translates into a product is not a one-to-one thing.

I think that's pretty universal. You develop a new RNA vaccine technology. That's not just the COVID vaccine. That's also maybe a vaccine against malaria, or maybe a vaccine against cancer, or maybe part of a gene therapy program. I don't know. So you develop these core technologies. And I believe the core technology that we've developed will be part of a suite of products going forward — some of which, obviously, we can't predict now but generally, which will enable the adoption of digital manufacturing in a more broad way by bringing more and more people into digital manufacturing, bringing the cost down, making it more feasible. I would say, I know that's pretty wishy-washy in general. But I think it's important to talk about those fundamentals behind it as well.

Erik: That makes sense. I think the analogy you meant to Intel Inside suits well here. So Intel's job was to produce great chips. Then there was all sorts of innovation in terms of the devices that could be using chips around that. And to some extent, Intel needed to be aware of those and collaborate with the people that were developing them. But their job was not to develop the hardware or decide what devices should be made. Their job was to make sure that they had chips that could help those devices perform. I think your technology will end up probably serving a similar role in this industry. But then, I'm sure you'll be at times surprised by the application better developed around that and then reacting to those developments.

David: As another example, my old employer, Covestro, just in the polycarbonate space has — I don't even know how many — let's say 10,000 article numbers just for polycarbonate. It's just one company. If you probably go to an ABS manufacturer or to a PDG manufacturer, it's similar. But if you look at additive manufacturing today, you have maybe in the industry a few thousand materials that you can choose from. I'm probably being ambitious with that number. And this gap is enormous.

One of the things we're doing at the moment is working with materials companies to be able to qualify their materials for additive manufacturing digitally. So they don't have to take a team of 5 engineers 18 months to bring a new material to a hardware ecosystem. They can do that all digitally, speed it up, and bring a better product to market faster from their existing portfolio. And if they're specking into an electronics device, electronics has, from design to manufacturing, a cycle of maybe seven, eight months, maybe nine months between when a new, let's say, tablet gets discussed at a brand owner and between then the product launch. If it takes you 18 months to spec in a new flame retardant material, for instance, that's just not feasible in that kind of timeframe. So we're also there broadening the material available in the industry. I think that's another example of just how broadly applicable our technology is as an enabling kind of Intel Inside technology to drive costs down and to bring additive manufacturing forward.

Erik: Yeah, that's interesting. If somebody wants to come out with a new material or a variation on an existing material, would you then be studying the molecular structure of that material? Would you be running a test on it and see how it performs under different temperatures or different environmental conditions? How would you be helping to accelerate that validation process?

David: Effectively, the challenge to spec in a new material in additive manufacturing is, you probably know quite well how that material performs in the real world. But what you don't know is how that material is going to perform in the manufacturing process itself.

Let's say, you start with granules. You turn those granules into a filament. Now when you're making that filament, you've already put that material under stress. So there's already a certain embodied stress there, probably going to heat it and then kneel it and try and get that stress out. Then in the printer, it's again going to go through this extrusion process where it gets heated, pushed through a very narrow tube, and then deposited. Again, there's going to be stress that's going to be built up. If you take an amorphous material, for example, the same crystallization is another interesting problem.

So what we can do is, we can put all that in the digital world so that these companies understand their materials. But now we're giving them the ability to understand that material in the context of a specific printer hardware. We can tell them, "Okay. Well, the profile of the material that you just uploaded, that's not going to play nice with this printer, or this printer, or this printer. Really, it's only going to play nice in this very narrow set, or it's not going to play nice at all." They can do all that digitally. Whereas before, maybe they had 1,000 experiments lined up, now they can reduce that experiment set to 20 or 30. They can save all that wasted material but, more importantly, all that labor and capacity of doing this kind of tests.

Erik: Okay. Very interesting. So you're simulating the production process and then giving them conclusions. Then they'll run some final tests just to validate that your conclusions were correct, and then move forward. Okay. Fascinating.

There's another topic I wanted to touch base with you quickly on, just because you and I are both sitting here in China today. Just as an innovation landscape, China is quite interesting for a number of different reasons right now. I just wanted to get your quick thoughts on, what does it look like as a foreign founder of an incorporated company with a team in China? If you could just quickly share your thoughts on what are the strengths of being in the ecosystem. What are the headaches if there are headaches, and how do you manage those? What are your thoughts on being a global company as a startup that has a strong footprint from its birth basically in China?

David: It's a really hard question to answer. We started here because when we found it in 2020, the pandemic was right at the beginning. I think my original plan was to found in the US, but we were living here. There was no way to get to the US anyway, and so it made sense to start here in Changshu. I think we were extremely lucky that we had really excellent local partners who helped us navigate the government's situation, helped us navigate with hiring. Definitely, I find it was sometimes very, very challenging to understand what was going on in the local government organization, knowing that everybody had good intentions and wanted to support but just not understanding what was the expectation on me and what was needed. I think there's definitely challenges there.

I think the other big challenge has been access to capital. I believe that Chinese investors have great Chinese founders to invest in. I think, understandably, they prefer to invest in a Chinese founder than to invest in a foreign founder. I think that's totally understandable. But it does make access to capital here more challenging for a foreign founder, in my opinion. The advantage, of course, is that things move really fast here. We're embedded in a fantastic hardware and materials ecosystem. Our partners move fast. Everything is available — material testing, new parts for printers, experimental platforms. Effectively, the world's printing hardware is being manufactured here in China. Having that just around the corner is a huge advantage. I would say that kind of summarizes it, Erik. There's both good and bad.

Erik: Yeah, it makes sense. We had a chat, my partner, Michael Maeder, with Daniel Kirchert who is the CEO of Byton — I don't know if you remember that company, but they were one of the competitors of Neal. They raised something like 1.5 billion and then went bankrupt in 2021. They couldn't secure another funding round. One of the things I remember him saying was that he felt they made a mistake that they were incorporated in China. They had about half of their AI, their software engineering team in China and about half in Silicon Valley. The half in Silicon Valley cost three times as much as the one in China. He said, "If I had done it again, I would have done 95% in China. Because in that particular area — maybe it's just not applicable everywhere — there's great talent and also very cost-effective talent in China."

I don't know how you look at this. It's certainly segment by segment. But if you just look at the talent base in China in this particular area of modeling materials, modeling physics, how do you look — I certainly don't want to insult any of your colleagues here. But how do you look at the talent base in China, compared to, let's say, the best-in-class talent bases in the US or in Europe today?

David: It's really hard to generalize, I think. I think what is easy to generalize is the cost base. What we managed to achieve in the last three years in terms of achievement per dollar, there is no way we would have been able to do that in Europe or in the US. I think that's absolutely fair. Not because necessarily people are paid less, but I think just because we were able to just stay very, very lean. We used staff on demand. We used our links with the universities. I didn't take any money for the first year and a half or two, so were really — I think it's stuff that we would not have been able to do in Europe or the US.

When it comes to capability, I think it's tempting to generalize in terms of who's more capable. I think I've heard arguments on both sides. I think at the end of the day, we look for people who are extremely curious and are able to apply deep knowledge in a creative way from many different perspectives. I think, for instance, the polymer physics team we have here. Even though polymer physics is definitely something, for instance, that Germany excels at — because materials is a big topic in Germany — I don't think we could find a better materials people in Germany. I think the approach to solving problems is a little bit different. I think, especially software engineers, Western software engineers versus Chinese software engineers approach topics in a different way is my feeling. I feel sometimes best practice processes. Let's say, how do you deal with GitHub? How do you deal with pushing things to GitHub and having a clear data pipeline and having a clear code pipeline? I think maybe the Western programmers that we had were more experienced with that kind of thing and were more like, "Okay. Well, if we want to do this scalably, then we have to have these certain hygiene things. We can't just jump in and do work." It's very difficult to generalize.

Erik: Yeah, it makes sense. You're dealing with a very niche group of specialists. And so it's very much about finding the right individuals as opposed to necessarily the general base of talent and ecosystem.

David: And we're very mixed group of people. We have local people, Chinese people. We have foreigners based in China and overseas. We have people working staff on demand. They are people who are freelance but totally buy in to our mission. We have people who prefer to be full-time and do eight-to-six, nine-to-seven, something like that, five days a week, and who are tied into the mission but they don't want that risk. They would like to be employees. Then we have people who are coming in from the university system, because the university system is not offering them the excitement or the intellectual challenge that they're looking for. So these are people coming with really different motivations to the same game. I think that's also a source of strength.

Erik: Great perspective on how to build a team with limited resources but builded really an A-class team. David, my last question would be about the future. So you guys are launching a product this year. That's fantastic. If you look out over the next 12 months, what is most exciting for you, both in the company and maybe if you also look a bit more broadly at the ecosystem? What's going to be keeping you most busy and most intellectually occupied over the next 12 months?

David: I think what's going to be most exciting is getting our technology into the hands of those first 100 large scale granule-based additive manufacturing operators, people who are somewhere in the Midwest, in the US, who are making aerospace or construction tooling, or people in Northern Europe who are making wind blade molds, things like that. And really seeing how we can save them the most in operational efficiency, and get feedback from them and understand what we could be doing better, and building the commercial team around that. I think that is going to be super exciting. Then in parallel, continuing to work on. We've already got in the lab the next generation of our core technology. So that will roll out. Seeing how you then bridge those two worlds — one of our now running business versus what's going on with next generation technology — I think that is going to be super exciting. Over the last three years, I haven't really had much of that because it's just been technology.

Then the other thing is geographically building up the Germany presence. We have some really great partnerships in Germany now. We'll be opening our first office in Aachen this year. We already have a small office in Düsseldorf. We're currently hiring for Auckland together with the RWTH, which is like Germany's version of MIT. Getting that set up and getting those first hires, that's going to be super exciting, too.

Erik: Okay. Fantastic. It's going to be a busy year ahead for you. I know you just raised some funds, so you're powered and ready to go. I'm going to quickly share your website. It's Helio Additive. That's helioadditive.com. If there's anybody on our audience who would be interesting for you to talk to, who would that be? I'm just thinking. If it makes sense for anybody to reach out to the team, what type of person or company would you be interested in speaking with right now?

David: I think the two areas or three areas would be partners. Well, let's say, first, potential employees. We're hiring both onsite and remote software senior full-stack software developer engineers, materials, people. If anybody is a material, scientist or materials engineer who wants to get into a super interesting space of working with polymers, come to us. I think that's a priority. And we're looking for that first key commercial leader, an engineer who can speak and sell and can be in Europe and in the US and working with those customers.

On the partner side, anybody in the software space in manufacturing or manufacturing printers or materials. Very, very happy to talk and brainstorm. Lastly, of course, if you are an investor who is looking to invest in US startups around manufacturing software, happy to put someone on our investor newsletter. It goes out every two months. It's not spam. Very short and just keeps you updated as to our progress.

Erik: Awesome. David, thanks so much for taking time to speak with us today.

David: Yeah, really nice to talk to you, Erik. Thank you very much for your time.