Kristian Sons Wins 2025 3DRC Grand Challenge | PixaScale | PI Apparel

Hello everyone and welcome to another Seamless Conversation. Today I'm really excited to be joined by Christian Sons, founder of Pixascale, who led the team behind this year's winning startup submission for the 3DRC Grand Challenge 2025. Their concept on AI-first DAM as an intelligent backbone for agile product development rethinks how we handle digital assets, turning static 3D files into living, AI-aware digital twins that adapt in real time. Their vision is to create a truly connected ecosystem that cuts time to market, automates change, and lays the groundwork for agentic AI that learn and evolve. So with all that said, Christian, welcome and a huge congrats again on your 3DRC Grand Challenge win.

Yeah, thanks a lot, Michael, and thanks for having me.

Absolutely. Really good to see you. Obviously, we saw one another not too long ago on the late of Milan, outside of Milan at Power Power Europe, and it was great to see you there. But obviously a lot has gone on since because you've gone on to win this big challenge. So really, really great to reconnect over this.

Yeah, absolutely. I'm also totally excited and great to have that one. And I think we will talk about this later, right?

Absolutely. We're going to dive right into it all. So the first thing, Christian, that I'd love to ask, obviously I know you, but for those who don't, could you intro yourself briefly and kind of perhaps talk us through how you became involved in the intersection of fashion AI and digital infrastructure.

Yeah, that was actually not a very linear way. So I started my career actually in aerospace and built at Airbus, I built one of the first VR teams, and there we looked into how to visualize product data. It was also already around the digital twin. We didn't call that digital twin back then, we called that digital mockup, but But in the end, it was the same idea from the construction data. How can we create more value than just the production? So, and that kept kind of, that was a little bit the red thread. I moved then from aerospace more to automotive. I visualized nuclear plants before I then went back to academia. I went to the German Research Center for Artificial Intelligence and did my PhD there in the corner between agents and computer graphics and democratizing 3D. Until then, an old colleague called me from Adidas saying, hey, I'm here in fashion. It's, uh, they're 20 years back compared to automotive, but on the other hand, it's a huge amount of products. It's soft tissues. It's, it's, uh, it's an immense challenge. Do you want to look into that? And that was my start then with Adidas. That was 9 years ago where I started then my career, first, um, in, in just the apparel area, but then led a huge sport program with more 120 persons on the tech side to digitize the whole chain from concept down to the digital selling.

Amazing. So you're right, it wasn't a linear journey at all. Amazing that you've kind of been able to bring experience from all of those different, different industries. And I'm sure at Adidas, that was a huge, uh, value add, you know, to cut— to come in with skills and background from industries that, as you say, are much further advanced with these kinds of things than apparel and footwear seem to be at the moment. Amazing. Well, look, obviously, as I said, we're going to dive into Pixascale, but before we do that, what gaps, I guess, or frustrations in the current tech ecosystem, tech ecosystem, really inspired you to build out what you've done at Pixascale?

Yeah, um, actually it was not too much of a frustration. It was more like seeing also what we did at Adidas, what kind of opportunities comes with these technologies. And, um, until then I was not even considering too much AI agents, where the big wave only started end of last year when I decided to, to quit Adidas and found Pixascale then, um, after quitting Adidas then in January. Um, so, so So that AI part was not even fully on the plate. It was more like, hey, how cool can product creation be? What— how many— how fewer meetings you can have if you have a single source of truth, if people talk about the same thing. I've seen so many great things for the users where I thought, hey, this is something that I want to carry generally into the industry. And then of course, um, the last 6 months were— or let's say last 9 months were crazy around Agentic AI. I was sure that AI would be big topic on creating content, uh, with all these cool algorithms that can create images and videos and so on. But I think the even larger impact will be on these agentic AI, um, and, and this is, uh, somehow a coincidence that now I need to build the software completely different than I thought because there will be a new kind of software generation with this, uh, new technology.

Yeah, it's nuts. It just go— goes to show how quick— quickly these things are changing. You know, at Py Apparel earlier this year, it was way more on AI than it had been 12 months before, and I'm sure it's going to continue like that. That's who— it's moving so, so fast. So, Christian, in a nutshell, for anyone that doesn't know Pixascale, what is it all about? What is it that you do?

Yeah, so we build a platform that allows you to create, manage, and scale. That's also our tagline around the digital twin, right? On the, on the creation side, we help organizing your digital twin more on the really around styles, around products, and so on. On the manage side, it's more around how can an internal team cope with creating so much content, because what we often forget is that the operations inside the brands has changed, right, from some, some people who who delivered prototypes and samples to the sell-in, to people who create content for sell-in and sell-out purposes. So, so in the end, you established a content supply chain in an area that did never deliver any content. And this is where I, I still think a lot of support is lacking to help these people to, to deliver these assets on time and full, um, in particular with fabrics and materials that run in parallel to the products. This is very challenging. It's a huge amount of fabrics, and this is where I still see a lot of, um, brands either cope with this one somehow with a lot of Excel sheets and so on, or even to a point that in, in recent years we have seen that people lowered the engagement in 3D creation even because this is a huge energy that you need to bring into the organization. And there I want to make everything more efficient so that, that people really continue to work in 3D, because I think this will not change, also not with AI.

Yeah.

And then the last one is scale. Yeah, last one is scale. Of course, this is where we created— we rendered in Adidas in the end more than 1 million images per year, right? So this is a huge amount. This is typically at sell-in events. So it's not a flat rendering effort and compute effort over the year. It's something that's very spiked towards sell-in events. And so our platform can basically arbitrarily scale these kinds of compute loads, and you'd never have to wait long for your renderings and automations, tech packs, and so on.

Amazing. I'm sure any brands that are going to read and watch this video are going to— can't wait to hear more about what you guys are doing and where you're gonna go with this. I remember in your submission that you framed DAM not as a static storage, but as an intelligent kind of AI-first infrastructure. So what makes your solution fundamentally different from the other solutions that are currently out there on the market?

Yeah, this was the last part that I was mentioning that I had to transform the whole solution, right? I think a DAM typically holds images, pretty static, let's say stupid images or stupid 3D DAMs, stupid GLB files. So, but with AI, and in AI we need to go a little bit away. I think the last years we looked, when we thought about ChatGPT and so on, we thought about a huge, database that has read a lot of things and so on. But I think we need to think a little bit different. What we have here is now a cognitive engine, a, a piece of technology that can reason, that can, that can understand concepts. And this is where, where I think this is where a DAM now needs to provide the semantics Yeah, it needs you so that the AI can understand this is not just a 3D file, this is a style. It has colorways, it has fabrics, it has this kind of fabric assigned so that it can also use its cognitive engine to do then autonomous things on top of the digital twins. And that doesn't work on just storing a GLB somewhere in a storage vault.

Yeah. Absolutely amazing. And I also remember a part of your submission that was quite compelling was this concept of product awareness. So can you give us like a real-world example of how that would change a workflow or perhaps a decision along the workflow?

Yeah, that's a good question. I really love the idea of getting rid of the boring stupid things, right? One, one typical example is a carryover. Carryover means often I need to take the style from a previous season, I need to remove some colorways that I stopped to offer, I need to add some colorways. Not a really exciting task for neither for a designer nor for a, a developer. So I think that we can come to a point where an AI can do that completely autonomously, just looking into the PLM system, looking what kind of colorways do we offer for that style, then going into the digital twin database, looking up for that style, removing some colorways, adding some colorways. If it adds colorways, it can maybe prepare already assigning some new color. Of course, in the end, a human is in the loop, needs to look at what the AI did. I think this is something that will not go away very soon. That, that of course the human needs to be in control in the end.

Absolutely. And the other thing as well that I remember you mentioning is that Pixascale is kind of responding to a growing pressure for late, late stage agility. How, how are you doing that, whether that's from cost shifts or market feedback or compliance requirements? How are you tackling that?

I, I love this topic. Um, I, I think every— or I only have of course, uh, not seen every brand, but a lot of brands that I've seen, they struggle with these many changes. And in particular, late changes, as you say. It can be for markets, it can be from a cost perspective. And this whole topic will increase, right? Because with ESG, um, where, where in the past a lot of these changes were, for instance, cost-driven, where you would meet your target FOB, then the product would come back into a circle to the designer, to the developer, to the supplier to change things to meet the FOB. Now we have ESG regulations where very late in the process, if you only implemented very late in the process, you have even more changes, right? And change, on the other hand, from feedback is good, right? As long as it's not an opinion, as long as it is data-driven and/or comes from a customer or from a demand perspective, it is good. So you don't want to take away these changes. So in the end, a brand needs to cope with all these changes. And I think AI and our systems can help on one side with a single source of truth, on the other hand through this cognitive engine that can tell you what changed, what, what are the differences, so that in the future hopefully not a salesperson is there with a physical sample and an Excel sheet, but what changed since they you produce this physical sample, right? But you have always really the latest presentation of your product available.

Right. And I guess I'd also love to ask, like, why— and you've touched upon this— but why is treating your 3D assets as kind of a living digital twin so critical in your opinion for future operational success?

Yeah, this goes a little bit back to where I said this is the opportunity that I saw. If you make this digital twin really the center of, of your collaboration between design, development, marketing, suppliers? I, I've seen so many wonderful things where the suppliers is interoperating way earlier, which is, by the way, totally typical in the automotive. You include the supply into the development circle very early, and this is now also suddenly possible. Why not send a sketch to a supplier who creates the first digital twin, and then everyone iterates around this one? So, so all these Both things are possible. Our platform is agnostic to allow that. And from that perspective, it would be a pity if you store your digital twin somewhere in the file section, 3 tabs down in some PLM system or in some DAM system. It needs to be the center of everything that you do.

Absolutely. Yeah. I'd love to talk a little bit about the challenges that you've come up against so far, because I'm sure there are some. What have been the hardest technical challenges that you faced or are facing at the moment as you build out an AI-native interoperability? Yeah.

Yeah, I think it's so brand new, no one actually knows how that will work in the end, right? You can— we can prepare the system to, to be ready for that future. But this is also where then in the end, I'm also looking for the partners who want to explore that whole topic because the idea of the agents is that they autonomously, autonomously, um, go to the PLM system, to a marketing system, to a sales system, to our system, pull all this information, come up with proposals. And that means all these pieces need to be ready for AI. This means each and every software component needs to transform within the next month and years to be ready for these kinds of interactions. And I think the wonderful thing is, in the past Brands went nearly into a fatigue around digitization because I hear a lot of users say our systems are not connected enough. And as a technical, technical guy, I was always like, what does the user mean? The systems are connected. We have these integration layers and we built these microservices and whatsoever to connect all these systems. But what our users actually meant is that our systems, they don't interact with each other. They even— the simplest task they cannot do, right? And now you have this cognitive engine. And, and so I see that the whole system landscape will massively transform. Um, and, um, so from that perspective, a super exciting time to build a startup and to find people who want to explore and be in the bus, right? And how you say?

Absolutely. And talking of people on the bus, What does your ideal partner look like? Who are they that you want joining you in the bus?

That's a good question. I mean, to be really honest, the tricky thing is we are a startup now and want to establish a system in big companies, right, in a B2B business, where typically the other systems around PLM, DAM are from huge companies and so on. And of course, we see that brands need a certain risk awareness, or they, they want to really want to go that, that path and that risk. On the other hand, I also don't think that— I think these huge brands need to take also certain risk, because otherwise, typically in these massive transformations, typically you, you, the smaller partners will be more innovative. And so I'm looking for someone where we, uh, for, for partners where we establish a really trustful relationship to also figure out how things will work out in the company.

Yeah, I love that. And obviously, when this goes live, Christian, you will be tagged. And of course, if anyone does read and/or watch this and you want to get involved more, please reach out to Christian and join him on his bus. It sounds like it's an amazing journey you're going on and so many things are changing so fast. So it'd be great for people to partner with you now from the ground up. I remember also from your submission that you described generative AI as like not just a creative tool, but like connective tissue in a company across functions. And what are some use cases that you've seen of AI that you're most excited about in general?

Yeah, the tissue, as I explained earlier, is more from the cognitive engine going to different systems, right? And how I see that is What we will see in the future is what we call multi-agent systems. That means each of these systems will have agents that are able to do autonomous tasks. But then there is also a corporate agent who knows the company roles, the company processes, and then can go out to the different systems and orchestrate these tasks, right? And this is the tissue. And so in the future, I also see then something where you can say our data in ecom says we need a floral t-shirt in khaki, and then the agent goes into the digital twin part and looks up previous products. It looks into your PLM or material management system, looks for sustainable materials to hit certain criteria, and then it comes up with 6, 7 proposals. Something where you would as a human already sit and 20 meetings, come up with 6 proposals, and then the humans sit together. I don't want to take away any kind of meetings. Meetings are good, collaboration is good, right? But you can take away the useless, the, the, the teddy ones, right? Instead, you sit them together in front of 6 proposals, you cherry-pick maybe from them, come up immediately with, with a great solution that already hit cost targets, that hit ESG targets. And, and this can be really great, and I'm looking forward to have that.

Yeah, that sounds amazing. I mean, to your point, it would take the mundane actions away from people and allow them to actually focus on the more creative, which at the end of the day is what we all want as human beings, right? That's the holy grail. So yeah, that sounds very cool.

And this is a little bit in contrast to the last 2 years, right, where the AI would take over so many creative tasks. Task or is threatening to take over some of the creative tasks. I think with the agentic AI, we have the opportunity to now take away the boring task and let the human focus on the hard part, the creative part.

100%. And I think, I mean, that's, that's, uh, applicable and relevant to, to any industry and job. My, my job included, you know, it's, it's made a huge, a huge difference. And now, you know, when I'm putting together events and agendas, looking for speakers, you know, it allows me to focus on having deep and meaningful calls and conversations instead, which is why I got into this industry to begin with. So no, it is great.

And that's a great point, Michael.

Sorry, go ahead. Yeah.

This is a great point as well, because opposed to other technologies that we saw coming in the past, I think AI is something where everyone is using that already. Everyone is already interacting with agents. So, so in the end, the whole adoption will be much easier because people will just expect that similar, like you can get a copilot to write mail or to write a piece of software, you also have a copilot to create a nice product, but you are in the driver's seat, right? So the analogy is the same. People are used to it. It just needs to now come also for our fashion products.

Yeah, 100%. I like the barrier to entry is so much lower. I remember, I don't know if you've ever met, uh, a Nicolijn van Enter from Footwearize. I don't know if you know her, but, uh, no, but she was saying that she was teaching— I can't remember now— she was invited to go into a brand and to do a workshop with the design team about how they can use AI to really help their creative workflow and process. And one of the most amazing things that she said she got from it was that there were people there that were more so on the business side, and they kind of suddenly realized that they too could take part in this because they all knew how to use AI too. You know, it wasn't perhaps what they ended up coming out with wasn't as good as a designer, but at the end of the day, it really kind of lowers the playing field— sorry, not lowers, balances the playing field and lets everyone kind of take part and play, which I think is a really, really big opportunity of AI when it's used in a responsible way.

Yeah, absolutely, absolutely.

Um, I do want to talk to you, of course, a little bit about the 3DRC experience itself. So, um, to start with, I guess, why did you decide to take part in the 3DRC Grand Challenge this year, and what do you feel like so far you've taken away from it? I know it's brand new, it's only been a week since it was announced, but Yeah, I'd love to hear.

Yeah, I've been observing the work of the 3DRC team for a while. Also from Adidas perspective, we have been active in some standardization groups like Kronos and contributed also to the GLTF standard and so on. So I think it's super important that the industry is also driving standards and the 3DRC is really doing a great job in maybe not on a standardization level, but at least on a best practice and helping and exchanging I mean, this is super relevant. Um, and so, so I was pretty aware about this group and, and their work. Um, and then I saw the, uh, the challenge. Um, uh, it was in a time where I was really, really busy with an RFP, and I was like, ah, if it was like 2 weeks later, I would absolutely participate. And then a colleague sent me the mail out and said, uh, the deadline has been shifted by 2 weeks. And I said, that's a sign. Now I really need to, to participate, and I'm really glad I did.

Well, us as well. I mean, this year, I'm not just saying this, we had, um, I think a record number of people that took part, but also like the quality of submissions too was super, super high. So I mean, it was amazing to be part, part of it and be a judge too. I'd love to also know, like, as part of this experience, what has been the most, uh, uh, valuable part, part for you? You know, is it the exposure? Is validation of what you're building? What do you hope is going to be your biggest takeaway?

Yeah, I think it's two. On the one hand side, of course, it's the exposure. I think you posted this a few days ago, Michael, and since then, so many brands reached out. Also, partners I want to work with reached out. Even VCs reached out to me today. And so it gives a good exposure. So I can really recommend everyone to participate. And And number 2 is then also, of course, then the feedback that you get from, from this great team. So there was— and, and, and, and the acknowledgment, right? If you do a startup, you go through these ups and downs between, are you crazy to give up a job at Adidas? That's so cool. And to, to heights. And if then experts are kind of validating your ideas, that's of course contributing to your mental health, so to say.

Absolutely, 100% it is. Yeah, to work on something fun like this is an exciting thing. I would love to know a little bit more about what's next. So what's your next milestone, I guess, that you're working towards now, whether that's either on the technical side, commercial side, or strategic side?

So we have our product pretty ready to what you sometimes call the MVP, right? Um, so, so we, um, we are now also ready to go to market, which means this was now of course a wonderful opportunity to now also go public and explain what we are doing. I think this work is not to be underestimated on the technical side. Of course, we have prepared everything for the agentic AI, but we also need now to implement the use cases to show it, to test it, and that will of course be also a big learning curve for us.

Great. And on that note, how do you, I guess, envisage your platform evolving to enable truly agentic AI across the entire product lifecycle?

Yeah, I don't want go too technical. Um, on the one hand side, on the content creation side, of course we, we create some on-avatar experiences. I think this has become pretty much a standard, but here the question is more how do I scale that. I think this is, is a wonderful use case for a platform where everything you throw in the platform you also get an on-avatar experience. And then this is a nice soft start into this whole how do I create marketing content online large scale. That is of course one of the AI topics. The other one, as I said, agentic interface, saying I can talk to the system and do some autonom— the AI can do some autonomous tasks. And then also we expose all the functionality to something that is called MCP, a new protocol that allows you then to use the tools that we offer in the platform from kind of any kind of AI platform, be it now from a cloud desktop or what's your preferred AI, you can— it's exposed as a tool and you can just use it from there.

Amazing. Well, look, I want to end with this, and it's a big one, especially with the rate at which things are changing so fast. What advice do you have for fellow innovators or researchers who are exploring how AI can impact the product development process? Us?

Yeah, um, I think, yeah, I'm, I'm always super cautious with these hypes, right? Um, um, I, um, and I was always very skeptical also about hypes at the point where I was thinking, am I too old? I'm, I'm that guy who said, oh, that's just a hype. But with agentic AI, I think it's really different. I'm, I'm really truly believing it's different. And because it's it has become such a big hype. You see so many things that are now called agentic, right? I think everyone who wants to, and because it has such a big impact on processes, on workflows, I think it's something that everyone should look into and get a basic understanding to become more resilient of all this bullshit that's also around that topic and really to get to the essence and at the core and creating value for your business and get an idea how you can create value for your business. So that would be my recommendation, right? Don't scrap at the surface, really look what does this cognitive engine, which will transform all our lives the next years, it's better to have some basic knowledge about that one.

100%. Well, Christian, thank you so, so much. I want to end by saying once again, a huge congrats on your win. We're all really excited. Excited to see what you do next and how you scale. And once again, for anyone who watches and/or reads this article on Seamless, please reach out to Christian. He's looking for partners now, even if it's just that you want to have a chat and learn more. Christian is available and would love to speak to you. So Christian, thank you very, very much. I'm looking forward to seeing you at a PI Apparel event soon, I hope.

Yeah, absolutely. I'm looking forward to the next PI Apparel and meeting the community again. And thanks a lot, Michael.